You are here because you are about to depend on somebody else's data, and you want to know whether ours will hold up. Fair.
Our core data is our own, sourced from public information, and every hour we ship an update file with the changes we recorded in that hour. Whatever you are building, it will be wrong the moment your data is. That is the problem we have spent most of our company’s existence on, and the only one.
In 1999, I heard that Google's mission was to collect and organize the world's information. That struck me as one of the most valuable things a company could do.
In 2011, I started MixRank with the same mission and a different business model, on a server under my desk that I bought with money saved from tutoring math.
Instead of funding data infrastructure with advertising, I wanted to charge for it directly, so our interests and our customers' interests pointed the same way from the start.
For the first stretch we were a sales intelligence tool.
We worked out which technologies a company was running on its website, and for everything else, the people and the company records, we bought from other vendors.
There were 15 or 20 of them back then and the supply seemed endless.
Then it stopped. Within a couple of years most of those vendors had either shut down or stopped selling. We had a product built on data we could no longer buy.
So we had a decision to make. Give up on that data, or go and collect it ourselves. We chose the second, and it changed what the company is.
We run our own infrastructure, four data centers, on hardware we own, which is why we can check records this often without the cost getting silly.
The part that told us we had made the right call came later, when one of our competitors stopped competing and asked to license the data from us instead. Everybody wanted this data at scale. Almost nobody wanted to do the unglamorous work of collecting it properly.
That is still the job; everything else we sell is delivery.
Scott Milliken, Founder and CEO
Vendors ship a file and call it data. From the moment it is written, it starts being wrong.
Headcounts move, people leave, apps swap vendors, companies change what they run. Refreshing monthly or quarterly means shipping something you already know is out of date. We ship a refreshed file every sixty minutes.
There is a lot of modeled data in this market presented as fact. We report what is publicly observable.
On mobile apps that means we do not track daily or monthly active users or session length, and we will say so on the first call rather than three weeks into an evaluation.
For most of 15 years we built the data instead of talking about it, which is why some very large companies run on MixRank and most people have never heard of us.
We have changed that, but we would rather be the company that publishes what it finds than the one with the loudest category page.
“It wasn't clear how powerful the tool was until we had a demo and learned more about the full functionality.”
“Not having MixRank would be a lot like walking with your eyes closed.”
“MixRank is great at being flexible to suit different customer needs and being able to fill a large gap for research generation.”
“The only dataset with the depth and breadth for powering quantitative analysis at USC Marshall.”
Assembled his first computer at eleven, skipped high school, enrolled in college at fourteen, and finished a math degree at Berkeley in four semesters while running businesses to pay for it. Has written code used by tens of millions of people and code that has processed hundreds of petabytes.
Runs revenue, and is usually the person answering the technical question on your first call. He will tell you what a dataset does not cover before you ask, which is a slightly unusual trait in someone whose job is selling it.