The Haiku of Finance for 09/27/17

Building DevOps tech
Loop it back to open source
Monetize each call

LoopBack Builds on OpenAPI

I recently added to my tech knowledge base after watching a programmer run a demo of LoopBack. It is an IBM API Connect project that provides one example of the OpenAPI Initiative. Open-source applications are quite common and offer enterprise developers a straightforward way to build internal apps. They also offer the foundation for a startup business model provided the API supports some kind of freemium monetization strategy.

One amazing thing about LoopBack is how it auto-creates multiple API end points with few code lines. That's a great tool for coding prototype APIs and enabling a rapid test cycle for deployment. The cost savings for enterprise DevOps efforts using LoopBack should be blindingly obvious in Cloudonomics. I am especially intrigued by LoopBack's utility in rule-based model validation. I blogged what I discovered at Decision Camp 2013 about how rule engines automate business processes, and now RESTful APIs running LoopBack can enable rapid validation of the rules' effectiveness.

There's probably more to LoopBack's usefulness than just finessing enterprise rule engines. Google Search results for "OpenAPI monetization" and "OpenAPI pricing strategy" show how programmers can charge for API access if they built something on a platform that processes API calls of open-source data. Google Search results for "LoopBack monetization" and "LoopBack pricing strategy" are less clear on how startups can build a stand-alone business model on this platform. I suspect the most likely path to success for a LoopBack-based startup is an analytics engine that pulls data from open government databases and packages it in a way that will support display ads.

I don't usually get technical since my blog has a finance focus. The whole point in studying the OpenAPI ecosystem is for venture investors to understand where an API-focused startup fits and how they plan to monetize their concept. LoopBack is just the latest way forward.

The Haiku of Finance for 03/31/17

Start some cyber tech
Find cyber channel partner
Cyber-lock it up

Financial Sarcasm Roundup for 01/31/16

Sarcasm on Sunday is way better than attending church. You could listen to some preacher lie about the nature of the universe, or you can listen to me tell the truth about finance.

Alphabet and Apple battle it out for the valuation world heavyweight championship. Both companies are symptoms of the Silicon Valley tech bubble. Investors have chased these stocks because ZIRP made savings accounts look stupid. Apple's (AAPL) P/E is deceptively low at 10 and Alphabet's (GOOG) is really high at 31. Both companies depend very much on smartphone users replacing costly phones more often than necessary in the developed world's saturated markets. Financial advisers who toggle their portfolio optimization searches with "high risk" aren't doing their risk-averse clients any favors by selecting overpriced tech stocks.

The IMF and its lending cartel will review Greece's bailout progress. It's really sick how the world's most important financiers play "extend and pretend" with a country that has no interest in paying its debts. Forcing losses on creditors would clear up credit quality questions a lot faster than extending maturities. I would have hung the foreclosure sign on the Acropolis by now but the IMF never called me to ask for advice. Athens should hire me to fix their problems if they can pay me in something other than gyros (which I really like to eat BTW).

Big SIFIs are cutting the biggest deals with the SEC to settle dark pool allegations. I have always wondered why investors deliberately walk into something they know is dark. The banks telling investors they will get the best execution in dark pools are also the same source of prime brokerage credit for HFT hedge funds trading in those pools. Willfully blind investors, duplicitous banks, and greedy HFTs all jumped into those dark pools to rip each other off. The traditional investors are always the dumbest money in the room, so of course they got taken to the cleaners.

I told you this was better than a church sermon. I'm more entertaining and honest than any religious leader. I should start my own religion so people can properly worship me.

Front Row At AlwaysOn Silicon Valley Innovation Summit 2015

I spent some very valuable time this month attending the AlwaysOn Silicon Valley Innovation Summit (SVIS) 2015 down at the Computer History Museum.  I attended in 2014 for the first time and that's how I got hooked on the AlwaysOn events.  Tony Perkins always rallies a start-studded lineup for these confabs.  I sat front and center for every session.  The stream of consciousness narrative I generate below should capture the event's zeitgeist.

George Gilder displayed his flair for original thinking.  I heard him talk at the MoneyShow San Francisco 2014 and he retains his conviction that gold and bitcoin are future stores of value.  I disagree with him on the usefulness of Bitcoin.  Tech enthusiasts like Gilder love Bitcoin as a currency due to its supposed time-based scarcity.  If something must be scarce to be valuable, the ability to fork a digital coin into another variant (Litecoin, Dogecoin) dilutes its value.  Real currency is recognized as legal tender; forking introduces confusion as to which coin can be recognized in a transaction.  I also don't buy his contention that gold and bitcoin are somehow independent from economic reality.  Those things are very much part of the economy as long as jewelry demand drives the gold price and the demand for stronger blockchain ledgers in transactions drives Bitcoin's development.

I have been a big fan of programmatic RTB advertising placement since I first heard of it in 2013.  RTB requires massively parallel bids that make ad buys more efficient.  The auction method that optimizes a price across several channels can probably apply to other sectors besides online marketing.  Calling Wall Street . . . programmatic bids for securities across several exchanges can eliminate the ability of high-frequency traders to front-run other institutional investors.

The debate between advertising tech and marketing tech will never end.  IMHO marketing tech offers more opportunity than ad tech as Big Data and AI make automated DRM tools more attractive.  Tech advances bring a lower CPC for ad buyers, enabling a wider audience reach and lower CAC.  Standardization will happen as ad networks consolidate.  Only the biggest players have the market power to demand adherence to standards, as Google achieved with AdSense.  The pain point for "fraudulent ad buys" reminds me of payments fraud in finance and it is probably amenable to the same types of fraud detection solutions.  There's a pivot opportunity for fin-tech startups there if they understand ad tech.

It's nice to see enterprise software successfully disrupt the HR function.  Salesforce and other big players will want to buy that stuff someday.  The supply chain is next on the list for innovation.  I have nothing against in-memory computing if it can extend the life of Moore's Law.  Quantum computing already has that law in its sights, so in-memory players may not have five years to mature and cash out.

Big Data and business intelligence (BI) are supposed to work together.  Analysts develop hypotheses from their own experiences, and Big Data is supposed to expand those potential hypotheses beyond analysts' familiar heuristics.  Virtualizing BI empowers non-scientists to run analytics.  Curious types can read the stuff I've been writing about knowledge management and decision rules for several years.

One of the best quotes I heard at SVIS was that "linear solutions don't solve exponential problems."  I'll use that line the next time someone asks me why another European bailout program won't solve Greece's problems.  The human brain is wired for comfort and familiarity; it recognizes unsound patterns.  Automation and Big Data generate statistically sound patterns that will present robust visualizations.

I liked a couple of cool audio / video tools on display.  One automated video editing tool will reduce user friction in content creation.  A smart hearing aid had controls on a smartphone and plays music.  It reminded me of how Adm. James Stockdale turned off his hearing aid during the televised vice presidential candidate debate in the 1992 elections.  The admiral could have saved himself some embarrassment if he had worn this hearing aid.

Lou Kerner mentioned how Roy Amara's Law describes tech investment bubbles.  He concluded his talk by arguing that private market valuations are stretched (echoing VCs' public comments since 2014 about bubble conditions)  but markets are not close to the dot-com era's levels.  He also thinks tech within public markets appears fairly valued, but public markets look overvalued.  I disagree with his use of the NASDAQ's five-year CAGR to show we're not in any tech bubble.  Five years isn't long enough to reach back to the post-dot-com crash bear market, and it still exceeds the NASDAQ's long term average CAGR.  I also think his example of non-VC deal participants ramping dramatically contradicts his data showing VC investing as a percent of GDP not yet reaching 2000's bubble level.  I do like his awesome quote, "Beware of confusing donkeys in party hats with unicorns," originally from Bryce Roberts of OATV.  Good one there.

I loved it when a former management consultant said "people, process, and tech" with no context.  People throw that phrase around a lot in knowledge management with "tools" sometimes replacing "tech."  We'll hear it a lot more in on-demand markets as the demand for micro-task fulfillment continues to disintermediate people from the work they perform.  Laws and regulations are not catching up to the on-demand labor market that TaskRabbit and others fulfill.  Other online markets are making offline experiences in travel, taxis, and real estate more valuable.  It won't make participants smarter if they don't understand market basics.  Investors making "sight unseen" big purchases like real estate can waste money even faster with tech.

Video service entrepreneurs believe content will prompt transactions.  It sounds like a stretch and reminds me of the "push media" fad predating smartphones.  Television ads already have professionally produced content and YouTube enables amateurs to make mashups.  User-generated content now has more distribution options than ever.  I can foresee the existing user-friendly video channels inserting transaction prompts into videos.  The implied corporate development goal will be to acquire startups whose algorithms will recognize user actions as purchase triggers.

Let me revisit the blockchain one more time.  Open source developers need to think hard about how to make a distributed transaction ledger that no one can falsify.  Claims that blockchains are immutable ring hollow.  Ask any merchant who has been the victim of a fraudulent Bitcoin transaction.  If a blockchain is immutable, it cannot reverse a fraudulent transaction.  A falsified blockchain is useless as an auditing record.

I was very impressed with PageCloud's demonstration.  The ease of editing a live website will be a boon to small and medium-sized businesses.  It looked way easier than the tools I've used with other web hosts.  I especially liked how easy it is for a merchant using PageCloud to adjust portal prices.  That feature will come in handy during a hyperinflationary period when shops must update their prices daily.

Women have proven they bring advantages to Silicon Valley.  Marissa Mayer and Sheryl Sandberg know what they're doing.  I would like to believe the conventional wisdom that women project empathy when pitching a believable narrative, but the female supervisors I've had in my career who told me baldfaced lies always left a bad taste in my mouth.  One of the SVIS panelists mentioned a McKinsey study on the elements of female leadership.  A Web search for that report brought up tons of McKinsey work on the subject.  I would like more experts to mention the supplier diversity programs at big companies so that women-owned small businesses are aware of that path to subcontracts.  Closing the knowledge gap in SMB contracting will help women.

The VC outlook for the coming year is usually the highlight of these types of conferences for me.  I try to think like a VC, partly because I always wanted to know why they wouldn't hire me in the early 2000s when I was looking for work.  "Unicorn" is rapidly becoming one of the most overused words in Silicon Valley, just like "disrupt" and "curation."  I just LOL at the late-stage non-VC investors chasing unicorns with very unique preference structures.  Expect plenty of post-IPO writedowns.  I heard one VC endorse wild spending on customer acquisition while cheap capital is widely available.  The VCs who acted that way in the dot-com bubble aren't around for the rebound these days.  The excuse is that free spending is okay if the LTV/CAC comparison makes sense.  My critique of that is a too-generous LTV assumption will tempt inefficient startups to spend like drunken sailors on unattainable metrics.  One laugh line quote from a VC is worth repeating:  "You can walk down the street, shake a tree, and two angel investors will fall out" because risk capital is so easy to get.  I'll try that the next time I'm on Sand Hill Road.  Someone else discovered that a group of unicorns is called a "blessing." The blessing of Uber, Lyft, Airbnb, and others will make a handful of founders rich.

AlwaysOn's latest list of hot private companies shows them staying private longer.  The late-stage capital injections are a big factor in long incubations.  Founders now have every incentive to hold out for larger pre-IPO valuations.  This year's big winner was Docker, a sort of "hamburger helper" for app developers that pivoted during its Series A raise.  Docker's CEO says we should play the hand we're dealt, and I agree.

Analytics for sales data is mostly "undefended territory" because it doesn't integrate with internal knowledge management systems.  No one in Silicon Valley ever discusses KM but I believe that's the ultimate determinant of how analytics delivers value from enterprise systems.

I would like cloud experts to announce which part of their stack is the most expensive to deliver.  The up-front cost of any open source solution is lower than a proprietary solution because customers aren't paying a premium to develop a proprietary code base.  I initially thought vendor lock-in is more likely in the hardware (IaaS) part of the cloud stack.  Now I think SaaS lock-in may be more costly; imagine the difficulty of switching from Oracle to Salesforce.  Consider how Amazon Web Services' price cuts and competition from other data centers makes IaaS a commodity, competing only on price.

Knowing your average revenue per user (ARPU) is a start to building credibility in a startup pitch.  Losing credibility happens when founders don't describe their target market or team background.  They also lose when they assume customers will pay a premium for a commodified service.  They gain credibility when they compare their total cost of ownership (TCO) to their competition, describe entry barriers that will protect their market position, understand their sales cycle, and hack their pedigree for endorsements and tech validations.

I have very little insight into the vice economy.  The Rosewood Hotel on Sand Hill Road is the local epicenter for VC vices.  Knowing that will come in handy at some point.  Apps offer instant gratification and gaming is the only vice that can be done 100% online.  The eventual legalization of online gambling may displace video gaming.  Addictions are more powerful if they generate real financial rewards.

The unicorn roundtable was the single best panel of SVIS.  Truly brilliant people were on hand.  The rise of secondary market liquidity is changing the exit options available to early investors.  I agree with Tony Perkins that companies will find staying private very attractive.  One panelist said that a large stock market decline means mutual funds and hedge funds will be forced to meet redemptions, and will no longer have the late-stage capital for startups.  Another expert said each megatrend has a value pyramid, and each level of the pyramid from base manufacturing to top content adds value.  New tech can invert those pyramids.  Startups must anticipate changes when those value-added activities will shift the pyramid in verticals they understand.  Folks, this was brilliance I never learned in my MBA program.  I attend AlwaysOn events for just this stuff.

Everyone is looking past mobile, just like a previous panel when someone said 5G tech will turn all mobile phone companies into short-sell stocks.  I have not heard anyone say "data sector" or "data supply chain" at recent Silicon Valley events.  I have blogged about those concepts and that's why I have a data-related speaking engagement coming up in September 2015.  It's also why some tech media people have been following my Twitter feed.  AlwaysOn might as well follow me too.  I learn enough from their events to build my own thought leadership.

The Haiku of Finance for 07/17/15

Google stock rockets
It pays to cut risky work
No more non-search stuff

Inspiration At IBM Content 2015 In San Francisco

I scored a seat at last week's IBM Content 2015 conference at San Francisco's Palace Hotel.  I expected heavy product pitches because IBM is one of the sector leaders in enterprise content management.  Knowledge managers are drawn to this stuff because both structured and unstructured records end up in unsearchable archives without some architecture to sort them out.  I had to wonder how my own US Army knowledge management experience (Iraq 2009, been there, got the campaign decoration) would have been different if I had stronger search and archive tools.

Former NBA basketball player Mark Eaton was the morning keynote.  I was impressed at how he translated sports lessons into articulate business lessons.  He forever changed my perception of retired pro athletes.  There's a lesson there for military veterans making transitions to civilian careers.  The dude had four basic principles for functioning on a team:  know your job; do what you're asked to do; make people look good; protect others.  I guess pro athletes have something in common with retired military leaders, because both groups could riff those topics for hours.  What made Mr. Eaton's approach different was his storytelling ability that put each lesson into a narrative.  It was more than a rehash of old sports triumphs that made him likeable.  I detected elements of Joseph Campbell's mythological "hero's journey" in his narrative.  Storytelling through archetypes matters in enterprise sales.  It also matters to knowledge managers who have to construct corporate narratives (you know, like product positioning strategies) out of chaotically arranged enterprise content.

The IBM keynote immediately afterwards was another riff on storytelling as a business function.  I did not know that Kurt Vonnegut had three favorite story arcs:  man in a hole; boy meets girl; from bad to worse.  Searching the web reveals both written and video testimony for Mr. Vonnegut's theses.  Right on, narratives rule.  IBM's favorite story arc is "better and better," something we will all presumably hear customers say if we have the right enterprise content management systems.  I don't need to recap the rest of the IBM product seminars and partner pitches.  My business isn't the right addressable market for them anyway and my own content management needs are easily met with desktop solutions.

The final enterprise content management (ECM) keynote was more pitching, with enough "content" to provoke my thinking.  I learned from experience with MS SharePoint that enterprise search results can be either fast or relevant, but rarely both.  There's usually a tradeoff.  Search systems will improve as machine learning habituates them to the needs of a particular enterprise.  IBM sure is proud of its position in Gartner's Magic Quadrant schema of wonder charts.  The hive mind known as IBM's Watson supercomputer will absorb all of the actions IBM's clients take in their systems.  I have no idea how IBM is going to safeguard all of that proprietary information, unless they're only tracking the metadata from actions.  Knowing how frequently users query, share, store, translate, delete, and do other things with documents is enough for Watson to see without peering into things that would destroy a client's competitive advantage if hacked.  I keep hearing the catch phrase "move document capture to the edge of the enterprise" but it's not a cliche yet.  The payment processing sector has that one covered and now the rest of the economy can learn how to convert dirty data from imperfect image captures.  Any vendor that isn't scanning documents at transaction points right now is way behind their competition.

Chief knowledge officers and their chief information officer friends will either earn their pay or get themselves fired once they commit to branded ECM software.  It's all going to the cloud even if Watson didn't exist.  Building a hybrid cloud and keeping key developmental projects under wraps will be hard for anyone who hasn't learned about Cloudonomics.  The ongoing fight in ECM won't be pretty.

I'll close with one more lesson from Mark Eaton.  He said that if enough people tell you you're suited to do something, you should listen.  That's how he got back into basketball and found his optimal position on the court.  Well, you know something, a whole bunch of people have told me over the years that I have a natural voice for narration, and they ask me if I've worked in radio.  I'll take that as a pretty strong hint that I should add voice work or webcasting to the Alfidi Capital tradition of excellence.

Full disclosure:  No position in IBM at this time.

Getting Finance Right By Automating Lying Humans Out

Financial people do many things wrong.  Sales people are generally the worst and their managers are right behind them.  Saying anything to close a deal is a typical bad habit.  Lying is hard to stop once it's financially rewarded.  The best way to stop it is to never start in the first place.  The next best way to stop it is by automating the functions that lying humans used to perform.

Brokers who lie to clients cause more than missed performance expectations when clients get wise.  They cause legal liabilities that bring lawsuits for misrepresentation.  Compensation on Wall Street has long been a colossal expense for firm that feel driven to pay top dollar for "talent."  The talented liars running brokerages and investment banks are about to meet the new digital world.  Automated sales processes will reduce Wall Street's highly paid sales forces to a handful of domain experts and their DevOps partners.  Everyone else will be obsolete.  The liars' legal problems will be gone when they are gone.

Public companies that send financial statements to the SEC are supposed to make them truthful.  It's hard to discover financial fraud but dedicated forensics examinations can uncover egregious frauds.  Most auditors aren't so dedicated because they just want to make it through the day.  Catching the more commonplace fudging is quite difficult.  Modern accounting rules give CFOs plenty of leeway to report odd things.  Outside auditors that keep teams embedded with large clients are prone to capture.  Enron got away with fraud thanks to its cozy relationship with Arthur Andersen.  Both companies no longer exist except as legal fictions needed to unwind assets on behalf of angry creditors.  Automated auditing is coming next after automated relationship management.  The botnets will share no sentiments for human buddies across the aisle.

Doing the right thing in finance always means reporting results honestly.  Many humans are not up to this task but somehow mask their deficiencies.  Very talented liars can fool a lot of dimwits for a long time.  Hardly anyone can be fooled forever.  Getting finance right will be much easier when automation drives lying humans out of the corner offices.

Alfidi Capital At DeveloperWeek 2015 San Francisco

I attended this year's DeveloperWeek in San Francisco.  The brand new Pier 27 terminal made for excellent views.  The most recent America's Cup made use of some temporary buildings here and it's nice to see The City repurpose the lot into something useful.

The conference started off in a disorganized manner.  The initial Main Stage talk got relocated to Stage 2, or so I was told.  The other stages weren't open on time in the morning.  The opening talk I attended started 20 minutes late.  Sheesh.  I did not have time to stay all day so I had to make the most of the morning.  I guess the trend in tech enterprises towards minimizing meetings means some presenters aren't accustomed to showing up on time in person.

The morning talk on GitHub pull requests convinced me that engineers are susceptible to simple ego strokes like fancy job titles.  Designating someone as "Senior Associate Engineer" makes them feel superior to a mere "Associate Engineer."  I saw this same head game when I worked at an asset management firm whose initials were the same as Baloney Goofball Imbeciles.  The "Senior Associates" at my former employer worked for the same entry-level wage as the new hires even if they had years of experience.  Giving someone a longer title is much cheaper than giving them a raise.

The gist of GitHub is that its suite offers advantages over wikis as a documentation method for engineering processes.  Updating a wiki takes constant curation effort.  I once inherited a US Army unit's wiki that did a poor job of documenting the unit's knowledge management architecture.  I ended up shutting it down and migrating users to a more hierarchical archiving system.  The US Army's new taxonomy for updating its doctrinal publications is actually a very useful governance technique for engineering handbooks.  Base document changes should be much less frequent than minor document changes.  Splitting out the documents into a base series providing broad guidance makes sense if they are not updated as frequently as product-specific technical guides.

I like that GitHub's online process changes contribute to a meeting-averse culture, a good evolution in engineer-based enterprises that value productivity.  Knowledge management systems in larger enterprises should use automated workflows.  Non-engineers seem to have trouble grasping workflows' utility.  GitHub pull requests apparently create an audit trail of changes documenting a process during updates, especially the "why" of an author's update.  That's analogous to clarifying a commander's intent in a military mission planning process.  I need to examine GitHub myself to find more analogies.

The Main Stage talks eventually began where they belonged.  The bottom line in several talks is that many verticals are upgrading their tech stacks and this drives demand for developers in the labor market.  I did not mind the thinly disguised pitches for services like HackerRank because they provide a disruptive benefit to productive enterprises.  I have noticed other such disruptive talent matching services migrate to the finance sector.  Whale Path does for finance professionals what HackerRank does for developers, elevating skill demonstration over a stale resume and obsolete academic credentials.  You're only as good as your last performance.  Having a verifiable track record of accomplishment on HackerRank or Whale Path means job performance has a market value.

Some other talks covered deep learning, a new approach to machine learning.  We're all going to hear more about it as VC money starts pouring into IoT devices that need to operate autonomously.  It was fun to see someone walk through the ease of manipulating images displayed on a Lightbox-enabled website.  No coding knowledge is needed to alter Lightbox parameters.  The clear lesson for non-techies is that object manipulation is an easily mastered skill.  Coding as basic literacy still makes sense but not everyone who touches tech will need to code.

I didn't win any raffles or door prizes at DeveloperWeek, nor did I score dates with the hot tech babes who flirted with me.  I still came away with the perspective I needed on how enterprise tech is changing.  DeveloperWeek matters to people driving tech's migration into ease of use modes that business domain experts can interpret.  Anyone in business who is at least minimally tech literate will be economically viable for years to come.

Alfidi Capital At The Big Talk Summit Silicon Valley 2015

I looked through my pile of notes from recent events and dug out my notes from attending the BIG Talk Summit in Silicon Valley this past January.  It was okay for a one-day conference and a few of the speakers had provocative things to say.  I am not clear about what Baidu is trying to accomplish by hosting such BIG events.  It may be someone's BIG chance to pick the brain of American thought leaders and piggyback their ideas all the way to the backdoor of a Chinese tech lab.  You won't see my name badge in the photo I took below because it wasn't worth capturing.

One Baidu guy and an MIT guy had stuff to say about deep learning.  I would like the next big leap in mobile UI to be a direct digital thought interface rather than a speech interface (listen up, Siri and Droid developers).  Thought transfer UIs would be perfect for wearables as remote control devices once the IoT launches itself into our wired home appliances, personal robots, and private drones.  Baidu's preference for high-performance computing (HPC) in AI over cloud computing is anecdotal.  I need to see independent confirmation of HPC's supposed superiority.

One of the Web's favorite MOOCs got up and talked about disrupting education.  I did that once in a Notre Dame undergraduate philosophy course by insulting several of my stupid classmates.  Content-based models like MOOCs will eventually have problems expanding into markets of authoritarian regimes (China, Iran, etc.) that control mass media, Internet access, and human thought.  Non-ideological course offerings may be universally allowed but menu offerings of the humanities will face political limits in conservative societies.

There's a local "singular" executive education system that charges a lot of money to talk to VIPs appearing on rolling flatscreens.  These people advocate innovative medicine because the cost per genome performance curve now beats Moore's Law in computing.  If you can handle bioprospecting, biosecurity, bioweather, biohacking, mHealth, eFormulations, and other such buzzwords then you have a future in health care.  It also means you'll be overqualified to run the ACA health insurance exchanges.  I could probably sell these people some solutions for biobonanza once I figure out how to grow cell cultures on a dollar bill.  Can biology be securitized?  Maybe genes can be sold like music royalties or IP portfolios.  You heard it here first at Alfidi Capital.

Speaking of finance, another MIT genius wants to disrupt it with Big Data.  Count me in, I'm all over it.  MIT's Media Lab is pushing the needle on social physics.  They should not mistake correlation for causation.  Idea diversity and social tie engagement density emerge in MIT's research as healthy indicators.  In other words, it's easy to achieve high GDP per square kilometer in Silicon Valley.  They also measure the effects of Pigouvian taxes and subsidies on the relationship between idea flow and ROI.  There's a ton of innovation waiting to spring out of all that research.

RAEL Berkeley can't plan smart communities all by themselves.  It takes ECPA and Cool California to move the sustainability bandwagon down the road to riches.  I expect that magical vehicle to get up to full speed once Apple and Google come out with their competing models of automated cars.  Watch one zoom through your ZCTA at highway cruising speed and try to guess which USDOT automation level it uses to avoid hitting your car.  Seriously, I'm being obscure because the obviously proprietary work of the automakers and their Silicon Valley partners is just as obscure.

The combo of robotics and machine learning was the topic that set some of the gurus on fire at this conference.  The breakthrough in neural nets that can execute pattern recognition 30 levels deep due to increases in computing power is begging for commercialization.  We really need to hardwire Asimov's Laws of Robotics into these things before we design them to breed and teach them to adapt to their environments.  Future generations will thank us for not giving them a real-life Terminator nightmare.

Steve Wozniak, aka the Woz, was last marquee guest at the BIG Talk.  He noted that hardware startups grew so well in Silicon Valley specifically because of its ecosystem, while software startups can grow anywhere due to the proliferation of IT.  His observation about software poses a minor challenge to MIT's social physics discoveries about social density and idea diversity being strongly geolocated.  I may have heard him say elsewhere that finding flaws and gaps in dominant tech systems opens paths to entrepreneurial disruption, but repetition is good when the Woz does it.  The dude has an enormous ability to synthesize diverse intellectual concepts.  He should volunteer his thought process to whomever is mapping pattern recognition 30 levels deep.

What's Woz's impression of Silicon Valley?  It is a frenetic work pace and massive environmental stimulation bringing fast progress, bringing a human price of up/down, happy/sad cycles.  Woz's mind must be a model for how rapid testing and experimentation with multiple technical iterations leads to innovative product functions.  His description of Fusion IO's disk drive development condensed a huge sequence of related analytical problems into something simply amazing.  Such creativity!  People asked him about corporate creativity.  IMHO the most creative companies give employees the most diverse stimuli and the most chances to randomly interact.  Look at the communal bulletin boards at the headquarters of Facebook and Google, where people post classes, clubs, movie nights on "campus," and other things that appeal to perpetual students.

I counted the number of times someone said "game changer" at this conference - exactly twice, and Woz was one of them.  Way to go, Woz.  One of Silicon Valley's living icons is sharing buzzwords.  I missed whatever chance I had to tell him so when we left the parking lot of the Computer History Museum at the same time.  I'm sure he'll have another chance to benefit from my wisdom.  The attractive women at this conference definitely benefited from my genius, especially the two who just had to pose with me at the end of the day.  Keep it coming ladies.  I'm getting big exposure at events like the BIG Talk.

The Haiku of Finance for 02/26/15

Python in finance
Calculate all risk metrics
No room for humans

The Choices Between Statistical Programming Languages

I recently attended a talk where an audience member posed a question to an experienced economist about which programming language is best for statistics.  The expert selected Python over MATLAB and Mathematica, and I think I heard the R language as one other choice (it wasn't clear).  Young professionals just entering the workforce with high-end skills take the choices between those languages very seriously.

MATLAB iterates the manual matrix programming most top MBA programs include in their decision science classes.  It interfaces with both modern languages like C++ and legacy languages like Fortran.  Some programmers, like our expert above, believe Python can replace MATLAB in transforming data into graphical displays.  Maybe it depends on the profession.  I have recently spoken with a hedge fund manager who swears by Python for its ease of use, comparable to MS Excel functions.  That's two anecdotal votes for Python in a week.

Mathematica's Wolfram language is almost as old as MATLAB, and just as heavily used in the sciences and engineering.  They both have open source competition.  R is a more recent innovation with open source origins.  I have heard R mentioned at tech conferences where hackathon alums discuss the techniques they used during competition.  Looking at snippets of R syntax I found on the web reminds me of the very basic DOS programming I did in an undergraduate business class in the early 1990s.  Comparing all of these to Python's supposed symbolic ease of use entices me to learn more about Python first.

Hedge fund quants know financial engineering, where high-level programming languages like Python are necessary.  I have long been skeptical of the value added by hedge funds.  Knowing their preferred lingo should make critiques of their approaches more credible.  I still expect a market cataclysm to shake out almost all hedge funds as redundant.  Any investment firms that survive will attract the best quant talent and programming will still be a valuable skill.  The key for those few quants who keep their jobs will be knowing the limits of financial engineering in marking risk/reward tradeoffs.  Economics are human-caused events, not natural phenomena governed by hard math.  Black Swans can still fly right through any Python script.

I am not yet at a decision point where I am ready to learn a computer programming language, but the changing demands of the finance sector may require such an adaptation.  Those of us in the middle of our careers need to seriously consider learning the basics of some of these languages.  None of the four languages I mentioned were in the curricula of my MBA courses but they are now recognized as de rigeur for enterprise data professionals.  The concept of coding as basic literacy is on many thought leaders' lips.  Young people who can do it are far ahead of middle-aged people in adapting to a high-value workplace.  Coding as literacy leads to a Big Data career.

Alfidi Capital Attends Location and Context World 2014

San Francisco is my usual location, in the context of finding material for some provocative blogging.  That's all the reason I needed to attend Location and Context World this December.  I couldn't go wrong with exposure to three days of expertise that blended real-time data, geographic information systems, and mobile apps.  I expect location based services to emerge as a subset of the data sector.  Pioneering a new sector means launching a new conference series like this one to cover it.

The pre-conference CIO workshop had a bunch of us overachievers generating ideas for retail use cases, enterprise cultural approaches to user privacy, and deployment methods for location tracking.  The 14 of us generated and critiqued 287 ideas in three hours.  That was about 20 ideas per participant, and I'm pretty sure my ideas were the most articulate.  My suggestions:  CX by location; loss prevention risk factors; Big Data analytics linking back-end streams.  I don't know if any of my CIO-related comments will end up in a Gartner or Forrester report, so I'll share them below.

Linking opt-in protocols to customer rewards and loyalty programs is always a compelling incentive to get users past that first data input hurdle.  Enterprises should specify in their privacy policies that they own location data for a user who enters their store.  User data can be abused if taken out of context, and accusations of racial profiling come to mind as a potential liability.  Cross-referenced data for demographic elements other than race can counter those charges.  Opting out in an environment of ubiquitous computing is hard, so the only way of doing so completely may be to just turn off one's smartphone when entering a store.  Opting out of anything after completing a transaction is probably impossible because a user's data history is so easily monetized.  Some CXs can cross the line between convenience and invasiveness.  Only a Chief Privacy Officer can draw that line where outliers of use case data won't cross it.  Life cycle management (LCM) needs more attention; LCM will accelerate, become more complex, and incorporate Cloudonomics.  Finally, I don't know whether retailers have though about how the WiFi spectrum may get crowded indoors with POS terminals, multiple smartphones, and in-store surveillance competing for airwaves.

The first day's analyst roundtable on transforming customer and employee interaction made me wonder whether developers in mobile, enterprise, and Big Data truly work together on integrated products.  The app sector needs consolidation.  I want "data sector" enterprise players to look seriously at vertical integration with data merchants.  Where is the evidence that mobile apps successfully manage workflows?  Mobile displays need simple dashboards, but handling workflows means tapping through several layers of deep dives.  Furthermore, mobile use in field services will need customizable dashboards that make workflows take a back seat to troubleshooting.  I can already see mobile power users getting heavy volumes of push notifications from the enterprise as a workflow method.  The difficulty is that responding to those notices forces mobile field reps to be reactive instead of proactive problem-solvers.  There goes mobile empowerment, thanks to workflow requirements.  Enterprise systems must inevitably be dumbed down for mobile users.  It's sad that field reps won't have much autonomy once some combination of push notices and simple dashboards constrict mobile use.

I suspect the difference between "proximity marketing" and "location marketing" is negligible.  One of those terms will win out when the tech media start hyping one.  Beacons are a hot item in enabling infrastructure and their makers don't care which term wins.  I don't know whether beacons really eliminate the need to spam mobile users by allowing consumers' digital lives to follow them into the physical world.  Many users consider ads on their Facebook news feeds to be spam.  Cueing up those ads based on social media actions means beacons are still in the spam chain by enabling proximity marketing pushes.  Here's my suggested critical path for beacon adoption:  100 CustDev use cases . . . improvements in conversion rate and shopping basket size . . . Cloudonomics ROI of proposed beacon solution.  Inserting hyperlocal data into a customer journey map will define the CustDev segment that makes this adoption work.

Enterprises have discovered telematics and it's not just for automobiles.  Ambient energy sources for IoT devices must have a sufficient power budget for edge computing.  Geospatial precision matters for remote devices and lidar sensors have found a market.  Context doesn't matter so much for remote IoT devices but it is very important in defining a retail UX.  Little Data is really Big Data sliced into location and context.

Jonah Berger's STEPPS concept helps location marketers reverse engineer a successful word-of-mouth campaign template.  Pre-store engagement metrics work backwards from the POS, not just from social media channels.  The "near-store" effort requires geo-targeting.  Proximity triggers (geo-fences, beacons) lead to higher conversion rates.  This quantifies the value of marketing campaigns incorporating location because a customer's actions at each gateway (pre-store, near-store, in-store) are tracked.  Forrester advises us to measure "mobile moments."  The thought leaders at this conference proposed using measuring media actions in pre-store moments, mobile app actions in near-store moments, and beacon cues in the in-store moments.

A couple of panel participants discussed how their enterprises tested in-store apps before a national rollout.  The biggest retail chains devote DevOps to internal labs and make acqui-hires for IT talent.  One IT mercenary claimed that data mining could generate frictionless growth by reducing subscriber churn in membership services.  That is an efficient tactic for SMBs who do not have the funds to deploy beacons and sensor farms.  The SMBs that know the total cost of ownership (TCO) of a plug-and-play location-based solution can easily figure the ROI of a full deployment, just like that mercenary IT dude.

I sat through one startup pitch from a solution in search of a problem.  They had a calendar-based app but were so stupid that they had not thought of a monetization strategy before launching.  Every failed '90s dot-com made that mistake.  Business models like these suffer from the cold start problem because they have not attracted a large enough user base to offer generalizable AI output.  Solving it forces founders to address a chicken or the egg problem with a choice between going for user traction with a basic product or deploying a more advanced product after more AI tweaking.  All of the software thought leaders I've seen in the past three years advocate iterative deployments through constant CustDev and A/B testing to get through these two related problems.  Startups that haven't figured this out are not worth my time.

The second day's analyst roundtable introduced me to augmented reality shopping.  I don't spend enough time in stores to benefit from augmented reality but a couple of years ago I did see some POS concepts enabling consumer choices.  Every time someone in tech makes a bullish prediction on smartwatches, I remind myself of the challenges of the watch display's small form factor.  It can show one or two data points in large type and probably no more than three more in very small type (four are possible if the background color's shading conveys something substantial).  I did learn a couple of cool things.  First, retailers may encrypt beacons to work only with their branded apps.  Second, a closed-loop service app initiates user contact that will lead to a yes/no sales decision, much like a live sales rep's closed-end question driving a conversion.  Thanks, SNL Kagan, for sending a hot babe analyst to explain these things.

Newcomers to location-based services should know that "O2O" means online-to-offline commerce.  Tech media have hyped it for several years as a huge untapped opportunity.  If it were really so huge, there wouldn't be such a large ecosystem of e-commerce apps corralling consumers into staying online all the time.  Think about it.  The whole premise of beacons pushing location-based notices is that consumers are never offline.  A better approach to keeping customers engaged is "locationomics."  Do a Web search for that term.  Informa Telecoms and Media organized this conference and claims to have coined the term.  No one paid me anything to say that.  The Location Forum defines locationomics and LBx Journal covers it regularly.

Tech superstar Robert Scoble came to plug his book The Age of Context.  He revealed that smartphones now have installed beacons, which strikes me as redundant if physical stores have their own beacons to query smartphones.  A smartphone with its own beacon might be useful for querying a purely mobile vendor (like a food truck or a custom delivery service) but that's a very limited . . . context.  Come on, you knew I would deploy that one.

One presenter displayed Beecham Research`s Wearable Technology Application Chart.  I think it's the most mind-blowing corporate thing I have seen all month.  It reminds me of the rock band Journey's album covers from the early 1980s, the ones with the scarab launching into outer space and doing other trippy things.  Hunt down that groovy chart yourself, because I won't repost it here.  It's one of those things destined to live on in Web lore.

One irony of LBS I noticed is that virtual reality (VR) minimizes the need for physical movement.  VR devices that disrupt travel and commuting mean LBS growth won't be nearly as pronounced.  The race between LBS and VR will define which business models dominate the next generation of e-commerce.  Competing standards for M2M and IoT interfaces will hinder this dominance until one breakthrough product builds an ecosystem that forces other aspirants to fall in line behind one standard.

The IT people at this conference must be getting chewed out a lot by their CFOs, because some have finally realized that allocating DevOps budgets between incremental opex and capex affects project ROI.  CIOs who tie their SDLC and LCM to the enterprise's capital budget are winners in the C-suite.  One presenting CTO claimed that incremental changes, personalized workflows, and push notices create context.  Okay dude, that's an extremely insular focus that ignores customers and runs into the problems with workflows I described above.  Shoppers who "showroom" by comparing in-store prices before making online purchases go around such insular CTOs.

Leave it to Google to deliver a talk that makes sense.  The Google rep on hand said cost per visit is a good KPI measuring the cost effectiveness of omni-channel marketing.  I'll take that suggestion further along a roadmap to success.  Find the most-clicked item on a website (i.e., a retail shopping portal).  Maximize that link's visibility in social media and web searches.  Use rich media to enhance its details, especially if it has embedded images or video.   I suspect this tactic is more powerful than linkbait or figuring the number of positive product reviews needed to overcome a negative review.

The venture investors present did more than just talk their books.  They think LBS can multiply the network effect of a first-mover.  I had not thought of crowdsourcing as a way to rapidly populate an original dataset, so perhaps that can overcome the cold start problem I mentioned above.  Strategic acquirers may indeed want LBS in their data streams but that does not mean they will pay for large exit events from VC-backed LBS startups.  Note above how I discussed the tendency of retail chains to develop their own LBS capabilities internally or through acqui-hires.  The VCs who expect their startups to achieve billion-dollar exits need to lower their expectations.  They may find more success in hardware than software given the attractiveness of beacons.  They also need to quit harping on "proxies for growth" like downloads or media coverage.  There is no such thing as a proxy for growth.  Startups either grow their revenue or become worthless.

Location is easier to define than context.  I define context as any monetizable or actionable option nested within multiple data streams tied to identity.  This of course makes anonymity impossible but users won't mind giving up their privacy if they get value in return.  Combining context with location further precludes anonymity.  Go ahead and quote me.  I'm pretty sure I'm ahead of Gartner and Forrester.  My Google search for "digital context" brought up some Forrester commentary from a World Economic Forum panel on digital context; it offered little detail.  Context is a brand new term for thought leaders like yours truly to claim.  I generated more details in this one blog article here than most thought leaders on big-shot panels.

There were some really hot babes staffing this conference, especially the show's CEO.  Maybe next time the unattached babes in attendance can show me some location-based services up in their hotel rooms.  I could fit that into the context of my incredible manliness that women find irresistible.  Alfidi Capital is way ahead of other financial sector players thanks to Location and Context World 2014.  

Alfidi Capital Observes BoxWorks 2014

BoxWorks 2014 was the latest display of how Box builds its ecosystem.  I attended for insights into the secret sauce of how an upstart gets other firms to adopt its collaborative tech.  I have never used Box but I should probably give it a shot.  Find my original genius in bold text, as always.

The kickoff chat between Box CEO Aaron Levie and media mogul Jeffrey Katzenberg was cool.  They must have played some Disney film score when he came out but I didn't recognize it because I don't have kids to raise.  The big lesson was that a strong mission statement, a powerful brand, and great tech make a successful business model.  Okay, Jeff, but what about human talent?  The gene pool of talented writers and animators is only so deep, so the great tech among media giants will have to develop AIs that mimic those human abilities.  Jeff did confirm that the size of the movie going market and the talent pool of animators are limits on the number of movies that studios can produce.  I was stunned to hear Jeff say that the digital volume of a typical DreamWorks movie is so dense that they have to use collaborative software to track the edits.  I look forward to the rich video and user-driven animation that tech is supposed to unleash, but we get what we pay for.  Many of the amateur mashups on YouTube are so derivative and uninspired that they're not worth watching.  Jeff's best lesson from the start of his career is that exceeding expectations in any job or mission assignment lead to winning.  Okay, Jeff, but I tried that in large financial service firms and it only got me fired because no one would tolerate it.  Jeff got lucky and I did not.

Aaron and his top Box people had more announcements to share.  Their new Box.org platform offers content management to non-profits.  That follows the latest trend in Silicon Valley enterprises.  Enterprises want to do well by doing good.  Box has been in mobile file sharing since before smartphones and cloud servers made it easy and cheap.  It's only fair that non-profits now get in on the action.  The big product announcements were Box configurations for individuals using MS Office, cloud multi-users, and an upcoming annotation feature in 2015.  I have seen other purveyors push routinized workflow products and now Box Workflow is coming in 2015 for rule-based routinized operations.  I was quite impressed with the look of these products; the MS Office compatible Box display looked better than SharePoint as a knowledge management solution and the workflow looked like a wise use of BRMS.

The special surprise guest at the keynote was Oscar-winning dude Jared Leto.  I had never heard of the guy.  He brought his Oscar to pass around in the audience, as if a bunch of tech middle managers had something to add to his artistic ability.  Aaron compared the Oscar favorably to Box's Crunchy award and said it must be the height of Jared's career to appear at a software conference.  Jared stayed in character as himself throughout this cameo at BoxWorks, and endorsed Box's ability to share artists' content.  One audience member asked an excellent question about how the cloud can impact older art forms that are not digitized.  Aaron and Jared think it will help older art find new audiences.  I had a mental image of a bunch of artists around the nation collaborating on a sculpture in real time, directing some robotic arm in a studio by uploading diagrams into a Box workflow engine.

I spent some time at the 1st Annual Box Partner Summit that ran concurrently with the first day of the main conference.  The main theme of "commitment" was everywhere in the quotes from senior Box people.  I expect banality at most conferences but I did not know enterprise software managers were deep enough to quote Sartre.  Box leverages its partners' domain knowledge to identify pain points of prospective enterprise clients.  Properly incentivized partners of all sizes are willing to refer enterprise clients to box for SaaS solutions.  Partner rebates are great if they grow earnings first and revenue second.  That may be hard for some growing cloud companies to swallow because they need to impress Wall Street with pro-forma EBITDA if they want to go public.  I wonder how CMOs and CEOs calculate the effectiveness of such incentives.  Good programs should have upsell options with measurable ROIs.  Kudos to Box for positioning its offerings partly as KM solutions for knowledge workers.

I had to explore the finance and legal workshop because I know startups that need to collaborate in those areas.  Permissions management in box sounds a lot like SharePoint, which is really more a policy issue than a tech issue.  I am not clear on how Box is different from SharePoint or Google Drive.  It obviously does document management, file synchronization and sharing, and workflows.  I suspect the Box advantage is its API allowing custom-developed apps that do what competitors cannot do.

The keynote and fireside chat with Jim Collins was phenomenal.  He thinks senior corporate leaders will increasingly come from CIO and IT ranks because enterprise computing has become so important.  I only agree with him if he means the software sector; I'm pretty sure CEOs in manufacturing and energy need to know how to make physical things work.  He organized his talk around ten major questions, which I won't repeat here because the background he presented on each one is in his body of work in Good to Great and the works on his "Tools" page.  His bonus question was inspired by advice Peter Drucker gave him to think about being useful to the world.  Jim's answer was to give moments of kindness and encouragement to others.  Changing others' lives and making people better mattered to Jim.  That may be why the audience gave him a standing ovation.  I have never seen a standing ovation for a guest speaker at a business event.  Wow.

Aaron Levie's chat with Jim explored more of this work.  Jim thinks Information Age management science means different applications of the same principles he studies.  Selecting people matters more and leading a network has a less formal power dynamic.  Jim talked a lot about how enterprises must always adhere to their values.  That's great but I've always known that a company's stated values are less real than the values top executives model in their daily behavior.  Aaron Levie comes across as engaging, irreverent, and intelligent.  If he and his top team exhibit those values when they're not in the public eye then Box is on the right track.

The CIO panel reminded me of the work I avoid by being in finance.  They all thought that business process transformation opens a huge market for unmet needs in enterprise collaboration.  I have never worked on a waterfall chart-driven process but that's okay, because these folks say requirements-driven planning moving from waterfall iterations to collaborative iterations.  I need to see evidence that corporate boards demand more tech-savvy directors who can evaluate enterprise risks.  I think that may be just psychological projection coming from CIOs who aspire to board memberships.  All of the anecdotal reviews I've read on board performance is that they are mostly lap dogs asleep at the wheel, selected specifically because they won't challenge management.

The VC panel on innovations was my next stop.  Professional board membership must be lucrative for VCs and others, but it comes with the caveats I mentioned immediately above.  The VCs mentioned traits like effective communication ability and other successful things that mark good leaders.  They did not mention honesty and personal integrity as desirable executive traits, but they did admit the necessity of firing any key executive who does not embody a firm's core values.  One key insight for tech startups is that lead scoring algorithms now generate significantly higher yields from marketing spending.  I was stunned when someone said tech megacorps (i.e., Google, Facebook) spend $10B on a buyout just to defend their $100B market cap.  That blew my mind.  This amazing insight implies headline deals are driven partly by celebrity billionaires defending their net worth with mergers and acquisitions.  Their parting thought was that investors should favor startups where the cost of capital and other macroeconomic conditions won't affect a business model.

The health care sector still has an innovation curve even after the ACA, according to the next panel.  I do not understand how the ACA's payment reforms will work unless they are intended as a stalking horse for a single-payer system.  I learned that the government pays hospitals to adopt digitized record systems, which is probably more costly than a simple regulatory mandate for any provider who wants to access the Medicare or VA payment networks.  Like I said, I just don't get this kind of reform.  The biggest insurance plan underwriters are able to drive demand for optimal payment networks.  They can identify high quality, low cost providers and eliminate variance from their PPO networks.  Data on these pricing variances in treatments allows buyers to eliminate sub-optimal choices.  The private sector makes that efficiency work because competition for services remains strong in a free market with many choices.  It will be less efficient in a market dominated by government-backed exchanges.  This is one reason why providers can't prove whether they save money by using Medicare's Accountable Care Organizations (ACOs).  This may have been the only panel where none of the panelists mentioned Box, knowledge management, shareable content, or collaboration.  The experts were so contentious because the ACA has politicized health care and skewed the sector's natural tech evolution toward choices that favor top-down intervention.

The retail panel sounded interesting because it's a sector I rarely explore.  The panelists said they want real-time data from all channels on all platforms, and that's a big market opportunity for SaaS.  It was depressing to hear someone admit that most retailers still live in a company-driven data model (internal focus) and not a customer-focused data model optimized for mobile (external focus).  Retail or any sector reacting to real-time data must have a fast DevOps cycle.  Predictive analytics adds value in DevOps by showing where proactive IT fixes should focus.  The weakest link in any retail business model is the high-turnover, low-wage workforce.  The best IT innovations must reduce the cognitive load on those low-skill workers so they make fewer mistakes.  That's why automation will eliminate many fast-food jobs.  Bring on the social CRM and get rid of the minimum-wage workforce before unions can organize them.

My favorite panel was on government innovation, and not just because hot Box.org babe Karen Appleton was the moderator.  The federal government's approach to data management and innovation varies by agency.  The FCC wrestles with over 200 legacy systems while it maps broadband capability for the public.  DOD's high-cost early adoption of tech still makes waves when program expenses draw scrutiny, so other agencies should jump on board that gravy train while it's still on the tracks and ask to share in DOD's tech bounty.  Federal law hobbles innovation by classifying an agency spending money on another agency as a felony if said spending does not benefit the original agency.  Wow, that's depressing.  No wonder every agency has internal counsel.  The FCC panelist was extremely thought-provoking by wondering whether public scrutiny in a democracy combined with social media can harm risk-takers who might otherwise have productive government careers.  He said it's also worth wondering whether authoritarian governments can capitalize on high tech faster than democracies because they are under less scrutiny without checks and balances.  Wow, heady stuff.

I sat front and center for that government panel because I had to ask them my only question of the conference.  I asked the panel for their thoughts on two government programs with the word "innovation" in them:  the NSF's Innovation Corps and the use of the Presidential Innovation Fellows (PIFs) in GSA's 18F incubator.  Aneesh Chopra, the former White House CTO on the panel, loved both programs.  The I-Corps trains people in Steve Blank's methods to commercialize the vast research produced in the Federal Lab Consortium.  The PIFs in 18F have built some interesting tools for other agencies and anticipates the creation of the US Digital Service.  The panel experts had done tons of thinking on driving government innovation; now the world gets to hear my ideas for Uncle Sam to use.  I want to see these innovators push more agencies to advertise their needs on Challenge.gov.  DOD should use it to farm out simpler fighter aircraft concepts that won't violate Augustine's Law #16 on astronomical costs.  It would also be great if FedRAMP used Cloudonomics metrics.  Oh yeah, let's get veteran-owned tech startups into I-Corps' pipeline so they get the inside track on tech transfer to the marketplace.

I caught the tail end of Aaron Levie's chat with Vinod Khosla before everything ended.  Vinod thinks Jim Collins' work is bull-stuff.  Funny!  He advises people not to try to predict the future.  Just go out and build something.  I hope he means "fail fast" and move on.

I was pleased to notice that the Box employees working as show floor guides included some very attractive women in tight jeans and purple Box t-shirts.  I was hoping to see some more of that kind of box but did not get the chance, if you know what I mean.  I also prowled the show floor and asked several booth sponsors whether they thought the iCloud celebrity photo hack was a wake-up call for cloud security.  The collective shoulder shrug I got in response told me that I may have been talking to salespeople instead of DevOps people who fix security holes.

Box did very well with BoxWorks.  I did not have time for the Jimmy Eats World concert but that's okay.  Plenty of youngsters got to have fun.  I'll have more fun at BoxWorks next year.  

The Haiku of Finance for 09/03/14

Where is your data?
Did someone hack into it?
Better check it now

Pain Points in the Semiconductor and Solar Sectors

I mentioned in my report from Intersolar / SEMICON 2014 that the semiconductor and solar sectors have some unmapped pain points.  I took it upon myself to identify those pain points in brief CustDev interviews with vendors on the expo floors.  I was not there to sell them anything, so I had more credibility than a vendor.  That's how I obtained the insights I'll share below.

I'll start with the semiconductor ecosystem.  The SEMICON exhibitors were all kinds of trinket makers that large semiconductor manufacturers would find in their supply chains.  The most frequently mentioned pain point for them was production cost.  Equipment makers don't always know what drives the cost of producing a complex product, which in turn determines price.  They don't even know if the problems reside in their designs or their supply chains, even after analysis.  Longer lead times add costs and customers don't always forecast their needs very well.

Other pain points in semiconductors seem to be hostage to production costs.  Addressing marketing channels is complicated by tight resources (making it difficult to address a changing market segment) and the difficulty of getting product release information into the hands of the appropriate technical audiences with purchasing authority.  One frequently mentioned pain point was the gripe that human resources were a recurring irritant.  Hiring the right people was always difficult and many of the company leaders I interviewed said "people cause problems."  They never said what kind of problems; I suspect those problems are really from lack of knowledge about how to solve marketing and production problems.

I have concluded that the relationship between production costs and marketing costs is a strong source of pain for the upstream parts of the semiconductor sector.  The mismatch between market knowledge and the ability to adjust capacity is costing the sector money.  This is an underexplored area that is fertile ground for solutions in codifying manufacturing knowledge, material costs, and business process maps.  I believe that startups can deliver disruption in this market by deploying enterprise Big Data analytics solutions addressing those knowledge gaps.  There is money to be made in solving manufacturing problems.

I'll continue with the solar sector.  Three very different pain points emerged under the broad topics of regulation, project knowledge, and financing.  Regulation poses unexpected problems for installers unfamiliar with local fire codes or national safety regulations.  This is particularly costly for inverters that must manage fault detections and shutdowns.

Project knowledge is a challenge for solar installers.  Project developers often ignore monitoring until the end of a sales cycle, and this is not weighted in system design until after funds are spent.  Each solar project is unique to a specific architecture and geography.  Developers stumble when they ignore the costs of grid access where transmission lines are inadequate.  I had a hard time believing some solar component manufacturers have difficulty sourcing basic materials like steel and silicon, so perhaps I spoke to at least one operator who was incompetent.

One very knowledgeable solar person made my day by addressing the cost of a typical installation.  Soft costs are a big factor in residential PV.  The expected lifetime of durable modules in commercial PV are still too much of an unknown.  These random costs make the perfect segue from project knowledge to financing, because uncertainty in estimating project cost means financing must be flexible.  The ease of obtaining customer financing would sell more solar PV systems.

I have concluded that the challenges facing the solar sector are more diverse than those facing the semiconductor sector.  Entrepreneurs can't solve the regulatory pain point, but utilities may be an untapped source of support in pushing reform.  There may be a market for apps that help installers navigate local, state, and national regulatory mazes.  DOE has made strides in reducing soft costs and making project planning more transparent.  Solving the project knowledge pain point is an open field for startups developing Hadoop-based knowledge sharing architectures.  Finally, the financing pain point has a plethora of solutions at the federal and state level but consumers may not know about all of the tax incentives they can use.  Yield cos and tax equity are financial solutions for commercial projects but offer little relief for single-unit residential installation.  Once again, apps may help the real estate sector arrange financing for home improvements that incorporate solar, wind, and storage installations as home improvements.

I only had time to hit up a few dozen vendors so my impressions are not as robust as what a startup would need for an actionable marketing plan.  Oh yeah, the funniest part of my CustDev exploration was when I asked one SEMICON guy to name his biggest pain point.  He rolled his eyes and said, "People like YOU!"  I knew when to back off but I was not deterred from gathering anecdotes.  I make it my mission to understand how to make money from disruption.  Someday a startup will catch my eye because it can solve the problems I identified here in bold type.  I'll be ready to commit my knowledge as sweat equity.  

Alfidi Capital at AlwaysOn Silicon Valley Innovation Summit 2014

I attended the AlwaysOn Silicon Valley Innovation Summit (SVIS) last week for the first time.  This conference has a long tradition of sorting through the most promising startups in multiple hot tech sectors.  My first taste of the show was the pre-opening social event, where remote operators entertained us with their mobile screen avatars.  That's what Silicon Valley types do for fun.  You can see one in action below.

I spent the next two days at the Computer History Museum receiving the wisdom of the Valley's all-stars.  Tony Perkins has been tracking tech trends since the 1990s dot-com boom when he launched Upside and Red Herring.  I still have a couple of hard copies of those mags in my archives.  I also have hard proof that I attended the show, in my typical name badge photo.  The bolded comments are my original action items, for any of you enterprise folks who are need a clue.

Satjiv Chahil let us all know that the world is not flat after all, as Thomas Friedman once wanted us to believe.  His work with the American India Foundation develops the digital and physical infrastructure of rural India.  The foundation uses street theater to advertise their developments in India in a native twist on Silicon Valley's guerrilla marketing styles.  Satjiv advised innovators to think of the cost structures in the rest of the world.  I think I know how that applies . . . cheap labor is everywhere but immature infrastructure complicates product delivery.

The discussion of new opportunities in data analytics covered the the changing real-time information that data fusion culls from formerly static sources.  I agree that data users are starved for real-time info and I believe the proliferation of APIs allowing data manipulation is just getting started.  I was not surprised to hear them say "Chief Data Officer" and "Chief Analytics Officer" are emerging enterprise titles.  I suspect that the CDO/CAO should work with the CKO under the COO, not the CIO, which would cement BI as a KM function.  Follow the acronyms to victory.  These guys got me thinking about KPIs, which tends to happen when I hear something thought-provoking.  IMHO those verticals that are the fastest adopters of real-time KPIs will be those that must integrate the most data streams.  The sheer number of discrete data streams flowing into an enterprise will drive analytics solutions, rather than gross data volume.

The venture capital outlook for 2014 implied that current tech funding is a boom and not a bubble.  I was stunned that most of the audience (82%) in the live poll thought startup valuations were not in a bubble.  Come on, people, the Fed's ZIRP stimulus has made everything a bubble by driving normally risk-averse investors to seek yield in riskier assets.  I would hate to be in some VC's shoes in a couple of years when they have to explain their cratering venture funds to institutional investors.  Anyway, I appreciated the panel's assessment that the amount of capital needed to start a tech company is at a record low, but scaling it up requires a record high amount.  High-valuation exits in some verticals like social media are easier because they're cheaper to scale.  I did not know that the M&A market in tech is twice as big as the IPO market, mainly because big tech companies like to grow with acquisitions.  These VCs are among the few I've heard who admit that the crowdfunding movement is on fire.  I can't wait to see the crowdfunding portals with the largest numbers of accredited investors get acquired by brokerages.

Online education is at some inflection point, according to the next set of experts.  MOOCs that assemble distributed content in playlists will resemble online music stores.  The challenge to universities is clear, but I believe the panelists' expected shakeout of failing colleges will be delayed as long as low interest rates keep student loans widely available.  Bad models can survive longer than they should even with poor content, but I agree with the panelists who expect distributed content to reign.  They introduced the term "knowledge portfolios" defining the marketable skill sets people will port from more scalable education models.  I see a role for the US Department of Education in helping to accredit MOOCs and develop universal standards that make their content portable, once said Department loses its rationale for funding soon-to-be-gone campuses.  I see their optimism for course offerings favoring entry-level job skills with an immediate ROI, and I'll raise them the next step in educational innovation:  disrupting state-run K-12 schools.

The "top company presentations" were actually pretty compelling at SVIS.  A lot of such pitches at conferences leave me with little hope that entrepreneurs have a clue.  That was not the case at SVIS.  I'll offer up a few random observations as blind items, since much of what the presenters offered was proprietary.  APIs allowing data collection and aggregation should be monetizable, especially since they show promise in replacing advertising servers and accelerating lead generation.  Data-source agnostic computing is great but it must allow for virtualization, and virtualization in turn can solve a lot of cloud collaboration problems.  Exchanging consumer rewards points on exchanges makes me think arbitrage will be possible.  One app doing both visual messaging and photo sharing is not an elegant solution because those are two very different functions.  I am very impressed with automated analytics that deploy results using visualization tools, because it's perfect for Peter Principle middle managers who would otherwise have to hire expensive outside consultants just to understand their own data.

Tony Perkins probed the CircleUp guy for crowdfunding insights.  They discussed the higher IRR for consumer retail startups, which IMHO is counterintuitive given the earlier panel's observation that scaling up is costlier than ever.  I would like to see consumer product startups address obvious gaps like urban "food deserts" underserved by grocery stores.  I get CircleUp's model of charging startups a commission and taking part in their warrants, and I think other portals can tweak this by adding revenue streams from investors who participate in a raise.

Jay Samit gave a brilliant lunchtime keynote on decoding disruption.  His mantra for cannibalizing one's business in favor of disruptive product development would be lost on Microsoft and other big firms notorious for product managers defending their turf.  I was one of the audience members who chuckled when he sketched out the post-literate world's penchant for communicating with images.  Wait until the images have embedded geodata, Jay, when they'll be more data-dense than ever.  Jay thinks smart shoes are on the horizon because walking recharges batteries, and Apple does have a patent for shoe sensors.  I will spend the next year or so examining tech startups through his social / location / mobile lens, which will probably bring me back to the geodata hint I dropped above.

Tony Perkins followed Jay's act with his own keynote.  Alvin Toffler's "infobesity" means saving steps in what people already do will build billion-dollar companies.  Got it, Tony, and lots of people working at Facebook agree with you.  He pitched the AlwaysOn model connecting its subscribers to emerging trends, and it sure worked because, hey, I attended, and I'm "always on" hot trends.  I like their crowdfunding platform concept . . . but I'm pretty sure they'll have to register it as a broker dealer given emerging SEC and FINRA rules.  This will challenge their ownership structure and I'm not sufficiently knowledgeable of securities law to say whether this portal should be a separate entity.  The business profiles they plan to publish for clients strongly resemble IPO "tombstone" ads, which are traditionally in the SEC's lane.  I'm pretty sure they can iron out the legalese to fulfill Tony's vision of disintermediating venture investing from the human gatekeepers on Sand Hill Road.

The Big Data panel enlightened me on how in-memory computing is a huge leap ahead in database order processing by creating a new layer in the software stack.  They admit that selling enterprise-level Big Data deployments is a challenge.  I knew that from several previous conferences.  It's a challenge because the solutions are pretty much the same price regardless of seat counts, which means the price doesn't scale along with the tech.  The panel thought that opportunities also come from layering apps and and blending data science with business domain knowledge.  I already knew that too, thanks to my attendance at Decision CAMP 2013, and you'll know it too once you read my blog article on that event.  I say the KM / DM / BRMS confluence is still the ultimate Big Data opportunity.  Cloudonomics calculations will validate the ROI of cost savings from business process improvements.  The panel wants business analytics to replicate VCs' pattern recognition abilities to help identify winning business models.  There's already plenty of historic data on startup failure for them to aggregate from the NVCA, National Science Foundation, and other sources.  I must disagree with one panelist who said privacy concerns will be a future trend.  I believe that people don't care as much about privacy as they say they do.  The whole concept is an anomaly of the first two Industrial Revolutions and it will not survive much longer as Big Data erases anonymity.

The social enterprise panel reveled the audience's displeasure with the amount of time spent communications tech.  One consensus point was that enterprise power users are in non-techie business domain functions who prefer simple tools.  I'm pretty sure the growth curve in social enterprise solutions will slow if business domains can't prove these things generate results.  Corporate treasuries aren't unlimited, people.  Where are the CKOs and COOs in all this?  Are they ever going to figure out that crafting enterprise KPIs is their job?  I know it's hard in fragmented organizations but it's worth the effort to grab project leaders and shake out their team-level KPIs.

Bill Gurley demonstrated his genius in his fireside chat on the instant economy with Tony Perkins.  His take down of the dot-com stock analysis algorithim was brilliant.  No one should assume an implied CAGR over 100% for several years running.  He also thinks it's a mistake to assume future bubble peaks will asymptotically match some previous bubble peak.  Yeah, this bubble peak has already surpassed the previous several.  The good news is that corporate VCs are the dumb money in the room, timing their investments poorly and ramping their buys prior to market highs.  I'll be sure and tell any startups I meet to pitch to corporate VCs in the best of times.  I learned from Bill that long-term, high square foot leases were a major reason for startup bankruptcies and recapitalizations in the 1990s dot-com bubble!  Wow.  That means the commercial real estate sector's health in major tech corridors is a pretty good leading indicator for tech bubbles.  I now feel sorry for anyone working at a VC-backed tech startup with its own gourmet food court; those luxury amenities add unneeded costs to their leases.  Here's a classic Gurley line:  Discounting long term risk makes us comfortable, "like a boiling frog."  Awesome!  Imagine it spoken with the remnant of Bill's original Texas accent.  Bill is amazed that privately held startups can raise $1B in capital without profits.  I'm amazed that the Valley can see the same phenomenon and still think venture funding isn't causing another bubble.  Bill also sees that startups launching variants of success stories is a sign of a market top, overfunding, and too much comfort with risk.  He nailed the missing link in Big Data analytics:  Pareto optimization.

I have to disagree with Bill Gurley on two points before I move on.  I don't share his thoughts on Uber or the effects of market corrections.  First, he said Uber's ease of use redefines its local driving use cases beyond the TAMs of the taxi and limo market historical sizes.  Okay, but I still say Uber's TAM is still limited urban, upper middle class transport.  The outsize use of Uber in San Francisco is probably a phenomenon of tech workers and their funded ecosystems (consultants, media people, nannies, gourmet cooks) living inside the new tech bubble!  Secondly, Bill thinks the 2009 market correction wasn't as painful for startups and VCs as the 2001 correction.  Hey Bill, the Fed lowered rates so quickly after the 2008 crisis that risky things looked a lot safer than in 2001.  Hardly anyone in the Bay Area can see the Fed's bubble now . . . except me.

Here comes the wearables boom, according to the next crop of luminaries on stage.  The panelists think software is the key to viable wearables, but frankly the ones I've seen at tech fests have no shortage of analytical horsepower.  These things need to have physical and visual appeal just like jewelry, because ordinary folks will be proud to show off a status symbol.  The coolest thing  learned from these experts was about regulatory loopholes applying to consumers who take control of their own data.  Medical privacy focuses on transmission confidentiality.  Voluntary patient use of medical wearables IMHO opens up a an indemnification can of worms for health care providers even with the devices' obvious benefits.

The team from 451 Research crunched the numbers for a look back at last year's top 250 startup picks.  The AlwaysOn SVIS show is one of the few conferences I've attended that holds itself accountable this way.  Several of last year's top picks went to IPO at valuations over $1B.  Apps, e-commerce, and mobility were the most active acquisition segments.  Analytics was a big common denominator.  Strategic buyers were the vast majority of acquirers.  The most important conclusion is that AlwaysOn's companies have exit valuations at premiums (measured by multiples of trailing revenues, both average and median) to most startup exits.

Cloudera got a special fireside chat as the summit's company of the year.  Their talent bench is pretty deep with people who worked at Google, Yahoo, Facebook, and Oracle.  I agree with their CEO that a salesperson with a technical background can easily be seen as a client's trusted advisor.  He's in the school of thought that startups raising lots of capital gain a flexible IPO date and create a secondary market for any early investors seeking an early exit.  His observation that most popular management books merely restate common sense reminds me of my undergraduate days when "reengineering" and the "learning organization" were the hot ways for consultants to make a buck.

Tony Perkins kicked off the second day of SVIS with a segue from app and smartphone dominance into a keynote from Smaato.  These guys are understandably fond of the freemium model for apps because it allows advertising penetration.  Mobile ads have scaled rapidly, certainly beyond my own expectations.  The latest hot term for big sites like Facebook is data management platform (DMP), which encompasses both users data and ad network data.  The Smaato dude must be reading my blog or something, because he noticed that the smartphone is becoming a hub for other wearables.  I've been blogging about that since late 2013.  Wait until the smartphone becomes the power management hub as well as the data warehouse and connectivity platform for wearables.  Here's another game-changer I'll throw at the world.  New programmatic ad pattern buys can address bids for similar demographics in parallel.  This means ad buyers pay more for precise demographic targeting, and users still get free content.  Smaato is catching on to this potential by doing both RTB and SSP.

The next panel addressed the multi-billion dollar mobile ad opportunity.  I get the difficulty of measuring how ad spending drives revenue, even for SMBs running location-based ads, and the integration of RTBs and SSPs into DMPs is supposed to make that seamless.  I wanted to hear more from the panel about how shortening a sales cycle makes the ROI of ad spending instantaneous but they moved on pretty quickly.  IMHO ad buyers are slowly recognizing the power of mobile ad spending combined with geolocated user data to drive campaigns that quickly clear out unwanted inventory.  The panel thinks about 50% of brand ad spending is wasted but marketing analysts don't know how to identify that waste, thus they pay twice what they should to build brands in target markets.  It wasn't always obvious (at least to me) that mobile ad gurus would figure out how to translate ad form factors from print to the various mobile platforms, but they have certainly succeeded.

Will cybersecurity really get worse before it gets better?  The next set of experts on stage at AlwaysOn said the mobile threat landscape is small but persistent.  I do not think making mobile OSs more secure will solve security problems at the user level as long as app developers are not building security into their apps.  The three legs of security, convenience, and performance in mobile computing mean one factor must diminish if another is to be enhanced; so far this is a zero-sum game.  Enterprises have methods to manage the risks of tradeoffs between these areas, and not just from a Cloudonomics perspective.  I've addressed enterprise risk before and my readers know what I would do.  Constructing a 2x2 matrix of severity versus probability allows plots of security incidents in each quadrant, and the risk profile for each quadrant should include Six Sigma estimates of their impact on operations.  I can lead horses to water on this blog but I can't make them drink.  Security is always an afterthought for product designers because they build for functionality first.  Performance and convenience are the factors that meet market opportunities because "security" is only a pain point for a savvy subset of each market segment.  The IT/OT convergence means security must catch up.

The "last-second economy" panel said Uber's success leveraged Bill Gurley's network multiplier effect, but I'm not sure the sharing economy qualifies as a back-end process.  That philosophy is not some internal part of the enterprise, folks.  It's common knowledge that Millennials are driving early adoption of instant economy concepts.  We have not heard the last of sharing commodified services.

Bill Cleary was up next to address Apple's market dominance.  He sounded like a pretty cool dude.  I never thought of Apple's top brand as a compelling narrative on its own until Bill positioned it that way.  The cool, hip ad images Apple uses are part of that narrative.  Apple's genius is seeding its brand into pop culture channels so that it's not seen as solely a techie image.  It was totally awesome to hear bill ridicule Google Glass, Amazon drone delivery, and driverless cars.  Someone finally has some great insights into unworkable tech that gets undeserving praise.  I offered my only question of the conference when I asked Bill whether Apple would make a 3D printer for the desktop manufacturing revolution.  Bill said not yet because it's too early; Apple needs to focus on other things first.

Other panelists spoke on how growth companies can leverage the cloud.  Customer demand for speed and performance means startups will have to go to the cloud to scale up.  Delayed page loads cause incremental revenue losses and caching content in other time zones is not a universal solution.  IaaS is a good solution for pre-populated caches.  The next amazing developments we can expect in the cloud are hybrid cloud brokering, data security partitions by nation, and ubiquitous cloud services as reliable utility-type background activities.  I repeat all of this cloud thinking uncritically because I want to see how it shakes out.

The Expensify dude easily had the best keynote of SVIS, with spontaneous applause.  He ripped apart the financial metrics that so many tech followers throw around as conventional wisdom.  If valuation is more than revenue and a multiple derived from a "vroom model," then ROI must be more than LTV divided by CAC.  This made his point about the difficulty of attributing sources of both costs and revenues.  Nouveau enterprise valuations are not quantifiable by traditional metrics.  We need new economic concepts to replace unquantifiable estimates.

The video boom panel noted how online video is now a completely personalized experience.  Every enterprise wants to create video because it is the most emotionally manipulative medium, but creation means nothing without distribution.  I just don't see how video startups can succeed by incentivizing viewers to watch videos simultaneously with their networks of contacts, given the medium's personalization.  No one wants to see their friends making faces on a screen while they're watching a movie.  Viral videos are sharable because people watch them individually at their convenience, no communally.  I'm pretty sure a whole bunch of venture investment money is about to be wasted on sharable video features that no one will use.  Multi-channel networks (MCNs) are figuring out how to use YouTube to distribute content and capture shared ad revenue.  The panelists made some bold predictions about how Google could enable video distribution in Gmail and how pervasive video will become in our lives.

The final panel covered one topic that provides me with no end of blog fodder . . . Bitcoin.  Yes, the cryptocurrency I love to ridicule has a whole slew of experts ready to sing its praises and raise venture money.  Listening to them gave me several OMG moments.  So many things they claimed about Bitcoin were nonsense.  Transactions are not at all frictionless given the market's illiquidity and the unreliability of Bitcoin exchanges.  Bitcoin does not ignore national borders because some nations do treat it differently; look at China's outright bans and the US's IRS tax treatment rulings.  Come on, people.  I almost went slack-jawed when one guy said he was proud to accept Dogecoin with a straight face.  Wow.  The dude also praised Silk Road as a "great use case" for Bitcoin's privacy.  Wow.  In case you missed it (ICYMI), Silk Road was a online market for illegal drugs that the FBI shut down, and Bitcoin did not shield any of its particpants' privacy because they caught the "Dread Pirate Roberts" dude who owned the site.  None of these panelists had any user advantages in mind.  None could explain why Bitcoin transactions are a better value than national currencies.  All of these thought leaders resorted to easily disproven claims and cargo cult wishful thinking about a "knowledge economy."  I'll give kudos to one guy who debunked the power of the "network effect" to speed adoption.

Any of you who are unfamiliar with my thinking on Bitcoin must read my extensive blog history on the subject.  Claims to marry cryptocurrency and trusted computing will always ring hollow.  Device hardware securing a private key does nothing to restore a key if the hardware is stolen.  Losing the hardware storing a private key means losing the Bitcoin if the private key isn't backed up.  One dude told me during one of the breaks that he thought Bitcoin was the greatest thing he's seen in forty years of trading currencies.  His rationale was that machines and bots could use it to trade with each other.  Ooookaaaay.  Folks, automated trading already happens in the hedge fund world with conventional currencies, and it adds zero liquidity to the global economy.  People pushing for the Internet to have its own currency don't understand that the Internet is not a sovereign government.  Using Bitcoin as a remittance mechanism for overseas expats makes no sense at all because Google Wallet already allows currency transfer as an email attachment.  A lot of people pushing cryptocurrencies will learn some very hard lessons about the real world.

Those of you who need more than my very faithful paraphrasing of what these speakers said can check out the SVIS 2014 archived videos on the AlwaysOn network.  This conference was a very valuable use of my time.  They had real honest-to-goodness expert speakers who have actually accomplished important things in the tech sector.  The lists of top 250 startups they keep generating are terrific coincident indicators of tech funding trends.  I'll see them next year for the next 250 great things.