“If you don’t have any facts, we’ll just use my opinion.” – Jim Barksdale (former president and CEO of Netscape).
Startups are the sum of the decisions made by the people who run them. Should you raise money? Who should you raise money from? What should be your marketing strategy? What are the next features you should build? Who should you hire? Ok.. you get the point.
If decisions are so important then it might be worthwhile to think about how to make them better. A lot of research has been done on this subject and you can literally spend years going through the books, papers and the various theories and schools of thought in decision-making. Needless to say, that will probably be a bad decision by itself. Instead, it is more important to understand why data driven decisions work and to instill such a culture in your company.
Why Data Driven Decisions Work
From all the papers written on the subject, there was one paper that left a huge impression on me. In 1979 the late psychologist Robyn Dawes wrote a paper titled “The robust beauty of improper linear models in decision making”. The idea is simple. Prior research discovered that a decision made using a good professional objective decision-making model will always trump expert intuition. The breakthrough in Dawes’ work was the finding that even improper, simple, “stupid” models are better than intuition.
Let’s put that into context and talk about hiring. Say you want to hire for a new position. Building a proper model would require you to collect and analyze all data about people you hired so far and come up with a statistical model to evaluate and identify the top candidates based on various parameters. This is what Google does. However, Google has a LOT of data since it hired tens of thousands of people so far. As a startup, you don’t.
Turns out that you can build a simple model based on what you think are the important parameters of a candidate and then evaluate candidates based on that model. More importantly, it will force you to think analytically about what are the important parameters to consider for that specific role. This by itself will lead to a better decision and will remove subjective biases. This is why data driven decisions are so powerful.
The idea from Dawes’ paper can be extended to other areas in your startup. Almost everything that you do can be boiled down into a formula. Doing that will force you and your team to think about what are the critical parameters for success in any given project or initiative and then better optimize for that success.
As mentioned before, as a startup you don’t have a lot of data. There are many resources online that you can use on day one. For instance, if you are building a marketing model and are wondering what is an average conversion rate for the freemium model, you can find an answer. However, it is critical that you start collecting data from day one so you can make more accurate estimations and as a result, better decisions.
Jack Dorsey and The Inference Team
I am not a fan of the “before and after” marketing tactic but it works well because it focuses on results. In the case of the data driven startup, one of the most interesting cases is the case of Jack Dorsey who has seen the power of data driven decisions and made it his job to put together an “inference team” at Square. This team focuses on collecting and analyzing data to make better decisions. In his words: “For the first two years of Twitter’s life, we were flying blind… we’re basing everything on intuition instead of having a good balance between intuition and data… so the first thing I wrote for Square is an admin dashboard. We have a very strong discipline to log everything and measure everything.” Amen.
The earlier you start collecting, analyzing and inferring from data, the better your decisions will become. A year from now you may wish you had started today.
If you are interested in decision-making theory, then you should really spend time reading about Prospect Theory – the Nobel Prize winning work of Daniel Kahneman and Amos Tversky. It will give you a good grip about how we make decisions and what are the common biases we fall into.