Rob Olson

Partner

Growth companies need a culture of learning, the directive to test, and a willingness to fail.

Trajectory
  • DigitalOcean (Director & Team Founder, Data & Analytics)
  • MDC Partners (Director, Analytics & Insights, Performance Marketing Services Group)
  • Crossix Solutions (D2C Analytics Lead: Pfizer, Sanofi, Amgen)
  • Princeton University (B.A., Economics)
  • Top 100 Young Entrepreneurs — National Federation of Independent Business
Perspective

I’ve always gravitated toward opportunities to work with teams who are looking at data for the first time — not always startups.  During my time at Crossix, I sat down with consumer marketing teams at Pfizer, Sanofi and Amgen to understand insights generated by a new data set they had been operating without for decades.  With advances in data privacy technology, Crossix offered them the ability to link their marketing efforts to actual behavior at the pharmacy, at the patient level. Overnight, these seasoned marketers were able to precisely measure the efficacy of various acquisition and retention tactics, in near real-time, enabling quantitatively-driven optimization decisions.  For these century-old corporate giants, this technology drove a dramatic (and sometimes contentious) shift toward a culture of fast and frequent decision-making, supported by a mathematically rigorous test-and-learn philosophy.

When I came to DigitalOcean, the company was growing so quickly and generating so much data, that it was both a challenge and a necessity to focus on only the actionable data that mattered.  Data paralysis is a growth company’s worst enemy — time is almost always against you, and it’s far too easy to feel like you have “too much data.”  When you decide to “dive in,” it’s imperative to have a compass directing you to a narrow subset of data. More importantly, you must force yourself to lead with action-oriented analysis rather than curiosity-driven calculations.  “If I find the answer to my question, what will it allow me to do differently?”

In the era of “big data,” Excel remains indispensable. Decision-makers in new businesses can easily over-complicate the way they think about using data. Being data-driven doesn’t necessitate having a PhD on your staff to programmatically mine your data lake for insights. Many highly impactful “data-driven decisions” have been made using data that can fit in Excel and math that can be done on the back of a napkin.  Where winners shine is in acknowledging that their “lake” of incomplete and imperfect data can be used to directionally guide their decisions — these teams prioritize decision-making velocity over academic precision, and in so doing gain a huge advantage over the competition.

Trust (your gut), but verify. As a founder or growth company leader, operate first with your gut and let data directionally guide your decisions when you reach a fork in the road. Acknowledge that there are some decisions that simply cannot be made with quantitative support, either because there is no available data, or because the decision is impractically difficult or impossible to measure. Move with high velocity, and use data where you can to continually verify your assumptions and to evaluate your overall progress.

“I haven’t failed, I’ve just found 10,000 ways that won’t work.”  Growth companies need a culture of learning, the directive to test, and a willingness to fail.  When you prioritize actions and hypothesis testing, every outcome is a gift and insight into either something to do more of or something to do less of.  Both are valuable insights that allow you to re-assess, re-prioritize, and move forward.

Step #1: Collect and store.  Before you even know what you may need the data for, find ways to allow your customers (or potential customers) to volunteer information.  Ask them questions, and ask them about their needs, intents, and demographics. Track observable data like website behavior, product interactions, and transactions in a way that is attributable to a single customer.  Get in the weeds and stitch together a customer’s story across datasets, so that you understand where there may be gaps in information. Start storing all of this data now, and then when you have questions to answer, you likely won’t need to wait as long (or at all) to generate enough of a sample to give you the insight you need to move forward.