How to follow your heart 💓vs data📈
The art and science of product development
I have worked for six tech companies by now: two in renewable energy (both Sunrun and OPower IPO’d), one in fintech (SigFig), and three in social media (Facebook, Instagram, and Tumblr). Having been in tech for six years (what feels like 20 dog years — trust me — I have aged) working across six companies in three different verticals, I start to pattern match CEOs, PMs, culture, product strategy, and companies. After a while, you realize that all companies are different but also the same. There are universal growing pains and pitfalls everywhere. You start to extract best practices and processes that work universally.
The portfolio of companies I’ve had the honor of working for ranged from Series B (<35 employees) to Series C (scaling it from 100 to 300 employees) to publicly traded big tech company (>12K employees). What unites them all was the gut to data spectrum. They all fall on an axis running from pure intuition to data.
Building lasting products requires both intuition and science. It’s rare to find a PM or company that balances both out well. There is no recipe for success or silver bullet. The secret sauce lies somewhere between gut and data. Going from a pure intuition model is making decisions based on your gut, user research, design-thinking, and product sense. Building products with pure science is like applying the scientific method to rigorous testing and incrementation. Neither is better or worse and they are not mutually exclusive. Where you fall on this spectrum is a function of your PM style, company life stage, and product type.
Exhibit A Company: requires pure intuition
Lifestage: pre-Series B/before product market
Product style: 0 to 1 product development
Pure intuition usually comes at the early stage companies (pre-Series B, sub triple digit headcount). You rely on intuition where there is no clear start or direction. That intuition is usually grounded by what the founder thinks the user needs before there is a product market fit. This is where a “killer product sense” comes in. Your product sense is the internal compass that guides you through choppy waters. You end up dogfooding your own product and start talking to users on the ground. I remember taking phone calls from solar customers who were not saving after going solar and yelling at me “rip this shit off my roof!” I quantified and routed user pain points back to product and engineering and helped make product changes. Every morning, I combed through Zendesk tickets and spoke to SigFig users directly on Twitter and email, troubleshooting their portfolio manager and investigating bugs. Staying close to the ground and listening to your users guides product development at these early stage startups.
Exhibit B Company: when intuition isn’t enough, start swinging towards data
Lifestage: post product market fit, Series C+, has traction and users
Product style: going from v1.0 to v1.1 phase
Once you have product market fit, you need more than just data and a pulse on your users. Once your user base grows beyond a “critical mass point” e.g. 100M, like Dunbar’s number, you can’t keep track of them all. You for sure don’t know what a thirteen year old in India wants on a 2G network. This is where you need to start profiling your users by personas, conducting user research, and market analysis. Otherwise, you won’t know why a sudden drop in DAUs in India was actually from women not feeling safe using Facebook. The numbers start to capture all the nuances of an diverse user base. There is where logging becomes super important. In January 2009, Facebook’s entire growth team just dropped everything they were doing to focus on building and fixing logging. When growth slowed down in March of 2009, they were able to tease apart the nuanced variables at play through reviewing the data they logged.
Exhibit C Company: “sciencing the shit” out of your products
Lifestage: big tech company, close to or post IPO
Product style: iterating from v7.0 to 7.1
As much as I believe that data wins arguments, when you defer to data as the sole source of truth, you start to loose the nuances that comes from intuition. I call this “sciencing the shit out of products”. At Facebook, you moved numbers as a PM. No product would ever ship until you have gone through rigorous A/B testing yielding statistically significant result. You would never release a redesign that tanked metrics. Like chemical titration, you slowly tested and turned the knobs until you perfected the right balance of numbers. This process, albeit effective, stymies a lot of innovation. Just look at the recent strategy of Facebook stories. Do you really need stories in three different places??
I’m wary of “sciencing the shit” out of tech because it removes the humanity. You start to dehumanize whatever you’re building without focusing on the people problems. At the same time, I’m wary of blindly following our intuition too much without questioning our assumptions for what’s “right”. It all comes down to a balance and there is no recipe for success.
“It is in Apple’s DNA that technology alone is not enough — it’s technology married with liberal arts, married with the humanities, that yields us the results that make our heart sing.” — Steve Jobs
If you balance your gut intuition with data, you’ll be able to build products that make your heart sing. Don’t over index on either end of the gut vs data spectrum. Know when to follow your gut and don’t forget to trust and verify with numbers.
Follow your heart 💓 and validate with numbers,