Data First vs Data Second




I’ve talked with many VCs over the past couple years that describe themselves as “data-driven.” That term has always confused me a bit.   I’ve never thought of myself as data-driven, but obviously I look at lots of data about every potential investment we consider – CACs, LTVs, NPS scores, Cohorts, Marketplace Liquidity, etc.  Does that mean I’m data-driven or not?  I’ve now realized there are 2 subtle, but very different ways to approach the data that’s available.  

Data-first vs data-second.  

Another way to frame this is “data-driven” vs “data-informed.”  Data-driven investors ask questions about data first, while data-informed or data-second investors look at the data, but it is second to their intuition and common sense.  Wikipedia has definitions for data-based (or data-driven) vs data-informed decision making as it applies to education.  I believe data-informed to be a much better approach. In doing a of research, I stumbled on this gem – a couple years ago, a test showed that combination AI and human “data-informed” chess players regularly beat the pure data-driven players.  Andrew Chen made a nice post about how data-driven decision making can lead to local maximums and weak decision making.  I choose to modify the labels to data-first vs data-second because the data acts as either a first or second pass filter.  

In the VC world –  data-second investors will always beat data-first investors. 


There have certainly been lots of advancements in data around venture capital over the past several years.

We’ve gone from a world of mystique and opaqueness to a world of greater transparency and advancements in data analysis with companies like CBInsights and Mattermark.  

Tomasz Tunguz publishes data daily about VC metrics and public market comps.  There is frequent talk about data analysts and data-driven decision making as key differentiators for VC firms.   The metrics around SaaS and marketplace businesses are increasingly well understood.  
This wealth of data can be extremely beneficial, but it also causes some investors to get into trouble and buy into hype.  

Let’s examine the behavior of the two types of investors.  In reality, this is more of a spectrum and people tend toward one end or the other, but let’s look at extreme cases: 

The data-first investor: this investor starts by looking at things like CAC and LTV and Sales magic numbers or DAU/ MAU ratios and CPIs.  They keep a log of all the data about hundreds of startups that they’ve seen.  They compare and benchmark any new startup against everything they’ve seen before.  The user, sales, or other traction data is the first-pass filter. 

If the data is not in the top few percentage points of companies that they’ve ever seen, then they will pass on the deal.  Nothing else matters about the vision or the team.  This data first approach leads to data-induced blindness. 

If the data is a positive outlier, then the investor gets incredibly excited.  To this investor, any idea can seem brilliant if the data is strong.  This investor pays a high price when they see the data they’re looking for.  The entire purpose and vision of the company all come second or third.  The evaluation of the team will start to take on a confirmatory-only assessment. This investor buys hype – they buy Zynga and Groupon and Meerkat.  I’m sure all of these companies had phenomenal growth trajectories at the times of financings, but there was underlying softness in the business models that required common sense to recognize. 

The data-first investors have inflated expectations about continuing growth trajectories and forecasts and they don’t pay enough attention to defensibility and sustainability.  They don’t really have a strong thesis about the company or the vision and how it will fit into the world over the next ten years.

These investors take gambles on forecasts and current growth – it could pay off handsomely, but it has an equally likely chance to crash as fast as it grew.  

I don’t like those odds. 

The data-second investor: This investor cares first and foremost about the team and the vision of the company.  What is it you’re building? What problem are you solving and why will your solution inevitably win? This investor asks far more questions about “why?” and far fewer about “how much?”  If the investor falls in love with the vision for the company – then they can proceed to evaluate the traction and metrics of the business.  The data may dissuade the investor from moving forward if it is weak, but even modest traction and growth could support the investor’s thesis.  

The investor needs to dig in and understand why the data is the way it is today.  Why should it continue? Why will it accelerate?  Will this startup be able to continue to thrive for years to come and become a big standalone company? Is the service a passing fad or is it here to stay?

 I’ve had countless founders ask me, “how much revenue do I need to raise a Series A?”  I always tell them that’s the wrong question to ask – the right question is “how scalable and sustainable is the revenue I have today?”  Scalable (able to add money and repeat or improve performance) and sustainable (highly defensible and with strong cohorts) matter much more than current revenue levels.  The current revenue is a helpful benchmark but has very little bearing on where the company will be in a few years.  

The data-second investor uses organized common sense and intuition and uses the data to confirm or disconfirm their intuition.  

At Jackson Square Ventures, we are data-second investors.  Come tell us about your team and vision – we’ll get to the data later.