Data companies fall in four quadrants: Truth verses Religion and Data verses Application
If you are thinking of starting a data company, you have to make a very important choice: what kind of company will you be? There are four basic types of data companies and all can be very successful … but the biggest mistake data companies make is that they try to do more than one at a time.
First let’s define the x and y axis…
Truth verses Religion
Truth companies are backward looking. They tell you what happened or when something happened or something about a person, product, or thing. The main objective of these companies is to have true data. Good examples of truth companies are a credit bureau (like Experian, Equifax, and Transunion), middleware (like LiveRamp, Segment, Improvado, and mParticle), and financial services data (like large parts of Bloomberg). These companies are usually very long on data engineers.
Religion companies predict the future. They tell you what will happen based on a set of data. The main objective of these companies is to accurately predict the future. Good examples of religion companies are credit scores (like FICO), fraud prevention (like ThreatMetrix), and measurement (like Nielsen, Market Track). These companies are usually long data scientists (and sometimes machine learning engineers).
Religion companies often purchase data from truth companies. For instance, FICO uses the data from the credit agencies as the core ingredient for its credit score.
Data verses Application
Once you have a valuable set of proprietary data, you have to choose if you will be a pure data company or if you will build an application on top of your data.
Data companies just sell data. The best way to know if you are a data company is if you have no UI or a very limited UI. Data companies sometimes sell direct to end buyers but often also sell to applications (which is why it is so important they do not become applications as you do not want to compete with your customers). Good examples of data companies are in financial services (like Yodlee, Vantiv), a pure data co-op (like Clearbit), location (like SafeGraph), wealth predictions (like Windfall Data), and others.
Applications make data sing. To really get benefit out of data, you need an application. These companies will have nice UI and more front-end engineers. Good examples are query-layers (like SecondMeasure), refined datas co-op (like Verisk and Abacus), integration layers (like Vantiv, Plaid), B2B product usage (like G2Crowd) and others.
Winners and lowers and winner-take-most markets
For a “truth” company to dominate its field, it has to be clearly better than everyone else. And “better” means its data needs to be the most true AND the market needs to believe it is the most true. In addition to truth, breadth and price are very important to dominate.
For “religion” companies, the most important factor is brand. When predicting the future, ideally you want to believe that the Nostradamus within the religion company is making accurate predictions. And while some people may dive into the Bayesian logic, most will trust the market perception. That’s why there are so many poor predictive analytics companies, because one can buy brand with money.
Series beats parallel
The biggest mistake data companies make is that they attack more than one quadrant at once. For the first $100 million in revenue, you should be focused on just one type of business.
Note: the original version of this was posted as Truth Vs. Religion: What Kind Of Data Company Are You? in AdExchanger in 2017.
Pingback: Tuesday Links – 03/12/2019 – Rad Kapital
Pingback: Selling Data for Fun and Profit: An Introduction to Data Businesses – Tom Roth