Christian Renaud's Blog

Trendspotting: Something I noticed during the last two market cycles was when you started to see startups that were lifestyle-pandering to the VCs. When all the VCs are driving new Bentleys (like in 99), then there were a glut of 'Bentley waxing' Startups (only partially kidding).

Some friends who just got back from an incubator program on the coast told me there are a glut of 'foodie' startups bubbling up now, much to the pleasure of the locavore/foodie angel funders out there.

This seems a bit premature, given the stage of the current venture bubble, but keep an eye on it nonetheless.

Futurecasting

[For those of you who are new readers to this blog, you probably don't know that I periodically go off on some sort of futurist bender from time to time. This is completely unstructured and half-baked, so feel free to skip this post entirely.]

I think we can all agree that society has changed considerably in the last ten years, driven primarily by access to information and the widespread deployment of the Internet.  I recall speculating with friends in the 90s about what a world with ubiquitous Internet access would be like, and in the 00's, it was 'what would a world with ubiquitous broadband be like?'.  Now that broadband penetration rates are where they are in the developed world, and with even broader penetration of mobile phones as information portals, we are at an unprecedented point in human history where access to information is nearly universal, inexpensive, and instantaneous. 

This access is the highway, and it is the uses of that highway where the actual value is delivered.  We have been so preoccupied by getting access everywhere, that we are just now beginning to focus on what we can do with that access. What we have done for Internet applications so far is training wheels.

At the same time, we are in the midst of digitizing all of the world's information at an unprecedented rate.  Take genealogy as an example.  Get on Ancestry.com and search for any of your ancestors and you have reasonable odds of finding scans of ship's manifests, Ellis Island logs, marriage and baptism certificates, and so on.  We are gobbling up the paper data produced by past generations and feeding it into the global repository, making it nearly eternal.

And it isn't just historical information we are digitizing.  We are instrumenting, capturing and indexing the world around us as well.  Shopping patterns, location checkins, photographs of your favorite pizza parlor.  I was in a city council meeting yesterday where a representative from the police department was requesting funds for license plate cameras.  These cameras sit in police and traffic cars and automatically scan 360 degrees for anything that looks like a license plate and stores that data.  If it comes up with a 'hit', it flags it to the driver so they can apprehend the wanted felon/parking scofflaw.  In addition, should you ever become, say, a 'person of interest', they can mine the recorded data to determine if there are any behavioral patterns that you adhere to "John typically parks his car on 5th street between 8am and 9am on Tuesdays" so they can conveniently find you later.  Think about this paired with facial recognition on all of the street surveillance and traffic cameras in place and you have an exhaustively instrumented environment.

There are a few other little things that will be powerful value enablers over the next ten years, that we've seen prior attempts at, so lets cover those briefly in this blogpost and some of their initial value, and I'll expand on them in subsequent blogposts.

1) Identity

This seems like a no-brainer, but it is a bitch to crack because everyone wants to own your identity for you.  We all have seen the well crafted phishing emails claiming to be from your favorite financial services organization requiring your credentials.  To be honest, when the Internet was being built out with all of it's glorious protocols, we (I say 'we' because I've been building Internet infrastructure technologies for ~20 years now) were more preoccupied with making it actually work at all than in two-factor authentication to determine if it was actually you on the other side of the transaction.

Now that we've moved into Internet delivery of darn near everything, being able to trust that your communication or transaction is with whom it claims to be, be it for individual communication or for a stock trade, is paramount.  Without known identity, you can't trust the exchange.  Without trust, the entire structure breaks down.

So, assume we will see widespread development on strong identity (and I dont mean OAuth for Facebook or Google accounts to confederated websites), followed by reputational systems on top of that.  Think about how much you could accomplish if you actually KNEW that the communication was legitimate.  When Jim Fenton was developing DKIM with Yahoo, he and I had many a communication about different treatment of trusted vs. untrusted senders, now that you could actually verify the sender.  That would save countless hours of compute cycles and terabytes of storage when you currently have to put every single email, known or not, though anti-spam and anti-virus filters. 

2) Location

My single favorite feature in IOS5?  Location-based reminders.  My phone now reminds me when I am driving by the grocery store of the list of things I wanted to pick up.  If I have identity and location to a relatively fine resolution, I can begin to infer context.  So you are parked in front of Java Joes every morning at 0830?  I can infer your context, and if you are available or not.  Once upon a time, there was a grad student at the MIT media lab that wrote some software for mobile phones that offered to watch your behavior for one day and then customize your menus based on your typical use-case.  If you were a soccer mom, your menus would be different than if you were a full-time college student, based on just watching you for a 24 hour period.  What if you had days or weeks of observational data?  How much more fine-grain could you craft a behavioral profile, and build context from that?

3) Context

Context is a holy-grail for technology companies.  If you are driving by a shopping mall and there's a sale at your favorite store, you wouldn't mind receiving a notification.  It wouldn't be spam at all, it'd be a service that you'd probably opt-in and even possibly pay for.

It would also inform all of the background systems if you were available or not for an interruption, such as a phone call or ad-hoc meeting request, based on intelligent policies you trained it for over time.  The aforementioned student's first project that I saw was a matchmaking service for Cambridge bars that would text you when someone who met your desired profile was nearby, by using bluetooth IDs.  if it got a match from both of you, it'd send an intro text to let you know and ask if you wanted to be connected with the other person.  Location-based/identity-based and [albeit manual] context-based matchmaking. 

So location and context are two nascent areas where we are just now scratching the surface.  The current over-population of crowdsourced rating systems (yelp, amazon reviews, rotten tomatoes) will eventually consolidate, and finally re-home to be more about you than about each silo'd service. 

Until these foundational building blocks are complete, expect a number of low-value/high-marketing services to deploy around making something optional (like posting a picture with a dining review) really convenient.  I read the blogs and newsletters from the top VC firms touting their newest investments and how great they will be, but seldom are they gamechangers.  They are the 'better than facebook' or 'better than twitter', but not 'we finally solved internet identity, you're welcome', because that is a hard road to hoe and no VC has the patience for the return on that.  Like Hollywood, uniqueness is rare, sequels are overdone, and people who want to do something that fundamentally changes the genre, be it movies or technology, will be vilified for years before being grudgingly accepted.

That pretty much sums it up.