Lenny Rachitsky’s Post

I’m a third way type of person. In a lot of cases, you do need metrics to track latency, first contentful paint, engagement scoring etc. I think the problem is when so much effort goes into trying to track the before and after in a dynamic environment. Who knows why a metric really changed? It’s a deep analytical exercise which is rarely done. And then you get into Campbell’s Law: make the metric important enough and people will find a way to game it, rendering it meaningless. I’d much rather have a human looking at qualitative and quantitative metrics as well as their judgement on how well I am doing given the context. Sounds like this is close to what they are doing, keeping the goal in front of the metrics. Great stuff.

I’ll have to listen to the rest for the whole context, but this feels like a function of luxury given the founder, leadership, and culture at shopify. It’s really tough for everyday product leaders to shift to anything like this without some sea change in leadership and the overall mindset for how a team or company crafts product.

I feel like someone s going to have a hard conversation with stakeholders at some point 😅 A company can t be properly run without numbers to understand if we re going in the right direction or not . Sounds like a “let s just do what we want and burn cash” pitch from the era when investors money was available at any corner with a good sales pitch. In 2024, I wouldn’t encourage PM to be unconscious and unaccountable for profitability and viability. And yes, accountability comes with numbers :-) Still, wishing the very best to every PM :-)

It seems there’s always a datapoint to rely on to support nearly any position. OKRs are great, OKRs are bad, lean is great, lean is bad, etc. maybe this points to the importance of “PM taste“ (someone give me a better name) as distinct from product taste. Knowing when to use certain frameworks being a more valuable skill set than mastering any particular framework.

The problem always is the definition of "right" and "indicators of progress"! Always a debatable topic!

Wow I love this - I’ve ignored my intuition in the past because of ‘the data’ and as a result, it led to the delay of a super quick feature we could have implemented which would have doubled our sales, instead of waiting eight months for it to be developed. I find that the culture of big data, data science, and most MBA programs ignore the value of anything that cannot be quantified. And that is a serious loss, because so much of what’s outside our collective knowledge about how the universe or how people work, hasn’t yet been figured out and can’t be quantified yet.

Metrics are a lagging indicator of great intuition. They can’t tell you why something happened or what to do next — but a little intuition and experimentation will get you close. Like the old adage: A horse-cart driver wouldn’t ask for a car. Users and the data they generate won’t predict the future for you. You’ve gotta get inspired, invent the automobile, and see if that changes the country.

I've been noticing a pattern where the best teams are given, and can deliver on, fairly abstract goals. The better the team, the more abstract or "first principles" their goals can be. It is then critical for leadership to not overly constrain the problems or possible solutions, which OKRs can certainly lead to.

Few questions about this model.. 1) how do you create a through line from sr leadership objectives >> lead-level >> IC-level execution to eliminate strategic drift? Aka how are we staying on track 2) at what frequency is strategy revisited? 3) how do adjacent teams get a quick snapshot of what your team is working on & what their guiding objectives are? Archie Abrams