A few months ago, I snuck into a talk at Arizona State University given by former Google/Alphabet Chairman Eric Schmidt. When the Q/A started, I tossed out a bold question.
“It seems like we have an unequal relationship with Alphabet. You collect our data, then sell it back to developers who can afford it. I’m working on a parking startup, and it seems unfair that our well-funded competitors can just buy Google’s trove of impressive parking data. Isn’t Google just concentrating access to data among the biggest companies?”
Eric Schmidt responded that who Google chose to share data with was a matter of user trust, after all, the data they collect benefits us all, right? Google has data, thus they can built better search. Their search becomes the best, and now we get a killer search product. The side effect? No one can compete with Google on search. But does this mean startups can’t compete because Google has collected the best of the data?
Andreessen Horowitz partner Benedict Evans pointed out that it was unhelpful to say “Google has all the data”, comparing data to relational databases and making a claim data is a technology layer that can flow through many businesses.
The only problem with this analogy is, well, everything. Database tech is more comparable to the combustion engine. Every car could take advantage of the tech. Databases, from the beginning, were an open source technology for any business to build on. Data? Well, it’s complicated.
Instead of using analogies, let’s use Elon Musk’s preferred technique: First Principles. Data has a few interesting properties. Data is
- a replicable asset — in theory Google, Facebook and Your Startup could all know that Scott likes Soylent.
- only valuable if actionable — a piece of data needs to lead to a desired action like clicking an ad or making a sale
- niche and specific — for example the number of pets you own may influence Amazon recommendations but isn’t valuable to Spotify
That last point is key. The reason startups need not fear the Big 5 is that a startup can apply data in niches that big companies can’t.
Consider how much trendy upstart Warby Parker knows about your sunglass purchasing habits. They know your face size, frame preferences, prescription, tinting needs, and budget. Amazon, while having a wholistic picture of your purchasing habits, just can’t match the trove of industry specific sunglass data Warby Parker has amassed.
Yes — Google has a lot of data. And they use it to maintain a near-monopoly on search. And yes — they are using it to expand horizontally. But when it comes to training their self-driving cars, they basically start at zero, which is why the autonomous future is anyone’s game. Each industry has different data, and data is not a faucet Google can stop at the source, it’s a diverse landscape and fertile landscape that can help businesses grow in the right conditions.
Google does not have a data monopoly, but data has enabled Google to build a monopoly. And data is really, really, good at helping companies build monopolies. That should both scare startups and give them hope. An attentive company can use data from a small set of users to build a better product, then attract more users, then get more data, and build an event better product. Repeat the virtuous cycle.
Perhaps the future isn’t Amazon’s vision of The Everything Store, it’s the Perfect for You store, where an ecosystem of companies specialize in being really good at one thing.