Impact-Driven Data Ownership — The Generalist Approach
What happens when you erase the lines between specialists to focus your attention on impact?
Traditionally, data teams are comprised of analysts, engineers, and data scientists, each specializing in their distinct profession. While this can generate very strong results in large companies, when you’re a small startup trying to stay ahead of the curve, this can actually impede your growth.
Small teams with a north star
One of the core values that drive the culture at Simply is maximizing Impact Velocity. By Impact Velocity we mean doing the things that will drive the most growth to our business while wasting as little time as possible. This is a task that’s easier said than done, as the startup life provides many distractions and potential dead-ends that you might run into.
So what’s the secret ingredient to staying on the right track to Impact Velocity? Focus! It sounds simple — trivial, even — but it’s actually very hard to acquire.
We’ve built our company structure as an extreme version of Spotify’s Squad method, splitting the company into multidisciplinary teams called pods. Each pod focuses on a problem or takes advantage of a growth opportunity and is measured by a single high-level KPI. The sum of these opportunities is what we believe will ultimately bring us the fastest and highest possible growth. Four years of triple-digit growth supports this decision. The KPI guides the decision-making process by serving as a north star and keeps us focused. This methodology has pros and cons which were discussed by our CEO Yuval Kaminka, but at its core, it works well for maximizing Impact Velocity.
Being a small multidisciplinary team that effectively runs as an independent startup has certain requirements from its team members. Mentally speaking, it requires you to consider yourself a co-founder, who is fully invested and accountable for the success — or failure — of the pod. Attitude-wise, it requires being a generalist and having a knack for problem-solving. It also requires being comfortable with re-evaluating priorities which can change as fast as the business does.
The data generalist
I’ve talked a lot about Simply so far, but not about what it means to be a data professional at the company. This is to stress that our profession is only a part of how we can contribute to the success of our pod and the company — what’s written above is true for any profession in the company.
So what does being a generalist mean for data people? For one thing, we don’t have data engineers, analysts, and data scientists as separate entities. A pod’s data person is first and foremost a highly skilled analyst, which is the core skill needed to be able to do discovery work. They are also the pod’s data engineers, taking care of any infrastructure or visualization tool requirements that will support data democratization. They need a solid understanding of statistics to support the pod’s quantitative and qualitative needs. Last but not least, we like our data people to have Machine Learning chops in their toolboxes, and the experience to know when to use them, and when not to.
Being able to perform cross-discipline tasks is the key to focusing on getting things done, rather than getting your part done.
Making an impact
In a typical pod setting, our data people focus on 3 main tasks that are crucial for pods to be successful:
- Data Democratization
As a consumer company with millions of end-users, we are blind without data.
‘What’s the current status of our pod’s KPI?‘
‘What’s going on with that test we’ve been running for the past few weeks?‘
‘How many users are still using iOS 9 — we’d like to drop support because of the costs associated with its maintenance‘
The answers to these, and many other similar questions, are the basis of day-to-day decision making in a pod. Our data people are entrusted with making (reliable) data easily accessible to pod members, to support their velocity of impact and decision-making processes. This is the critical path to making not just fast, but smart decisions.
2. Data Discovery
This is the familiar process of digging through massive amounts of data, survey results, and user interviews to uncover opportunities that would dramatically improve the pod’s KPIs. However, our data people not only push the product roadmap ahead with their findings but also participate in framing problem statements and coming up with hypotheses for solutions. They assess the impact of these solutions, supporting the prioritization of the product roadmap in the process.
3. Hypothesis testing
All of this comes into fruition by designing, monitoring, analyzing, and eventually drawing conclusions about the results of the many tests that we run to validate our hypotheses and to measure our progress. Having said that, it’s also important to manage risks and to know when not to run a test. That is an art-form in and of itself.
A wide range of skills
These tasks require a significantly large tool-box of skills; I like using the metaphor of an arch, where analysis skills are the center-piece that holds everything together, while the rest are the building blocks that comprise the arch.
Obviously, most people aren’t that “well rounded” when they join us. As people, we constantly grow and learn, and strengthening our weak-spots by learning from others is a great way to support meeting our day to day deliverables and individual growth.
Final thoughts
Though it’s definitely not for everyone, we’ve found this approach to be very effective for us at Simply, allowing us to massively scale as a company without compromising our velocity of impact. Personally, I find it’s an exciting way to work as it offers endless opportunities to grow as a professional and bring your strengths to the table. There’s nothing quite like the sense of accomplishment when you get things done — or as the great John “Hannibal” Smith often said:
Stay tuned for additional posts that will explore what the day-to-day of a data person at Simply looks like and check out our open data roles.
