Netflix: Content → Customer matching

Uber: Need ride → Get ride

Amazon: Order placed → Package delivered

What's your leverage point?

Real data strategy asks: "Where can data make our competition irrelevant?"

Most data strategies ask: "What should we measure next?"

Spot the difference?

Tesla doesn't just track car performance.

They own the transition from "car needs maintenance" to "maintenance completed."

If your "data strategy" can be copied by buying the same tools, it's not a strategy.

Before: "We need better dashboards for every department"

After: "Which business transition can we dominate with data?" The transformation was immediate….

No of course it wasn’t! No one got it at first.

Persistence was needed, but luckily, great managers know when to push and when to fold.

They pushed.

Everything changed.

But while Blockbuster obsessed over store locations, late fees, and inventory turnover, Netflix asked something completely different: “Where can we use data to compete in a game our competitors can’t play?”

They didn’t try to optimize everything. They focused entirely on one transition: content-to-customer matching. That’s it. They bet that if they nailed that, everything else—retention, growth, customer happiness—would follow.

And they were right.

This is what separates companies who “have dashboards” from companies who use data to dominate: the first tries to measure all parts of the machine, the second picks one part—and reengineers it with data.

Whenever I say “data strategy is business strategy,” the room splits in two.

One half leans in. They start talking about LTV:CAC, dynamic pricing, churn signals, decisions. They see data as the raw material of better business design.

The other half smiles politely and says, “yes, we’ve got dashboards in place.”

I did my PhD thesis on mathematical singularity theory, and yet whenever someone asks me how that helped me become better at my job, I feel awkward.

Of course doing a PhD helps you to solve super complex problems and learn complex topics fast, but if you think about it, people who do PhDs often already are inclined to solving hard problems and learning fast (like I am).

So does a PhD really help you move forward in life and your career? I don’t know, at least I’ve learned to become sceptical.

Startups spiral when every meeting ends with “why don’t we also try…?”
Product teams stall when they chase every shiny use case.

Even AI gets better with constraints—tell it which “bowl” to search in, and it stops chasing ghosts.

Prioritization starts before the choice. It starts with cutting down what’s on the menu.

Since time moves only forward and moves constantly, there is no such thing as the same event. No single event ever happens twice; no two events are ever alike.

I feel like there’s a lot of wisdom buried in that simple concept. We love to think that events are “alike” or that they happen “again and again throughout history” or that we can deduce decision-making rules to apply to “similar events.”

But for physics, there really is no such thing as similar events. A lot of time, I think, we’re placing way too much on the idea of the “same” event.

The fix might be expensive.

It might be ugly.

It might be hacky.

It might not work.

It might involve handy manual work.

It might break constraints.

It might be only temporary.

But usually, it is indeed quick, and it exists.

However, hedges are not limited to the investment world. In fact, I cannot think of a single discipline that cannot utilize the concept of a hedge.

It’s a simple idea but a strong one: Next time you make a decision and identify a set of risks, instead of thinking whether you want to take those risks, why don’t you start to think about spending money and time to simply remove those risks?

by Henrik Kniberg

(FWIW, The Economist reports the number went up from 15% to 25% over the past six decades, making this a strong trend)

But today, stepping back, i saw it wasn’t about the initiative. it was about direction.

Like laying bricks for a wall, only to notice the foundation is off.

No amount of perfect stacking will fix that.

The failure didn’t waste time. it cleared space.

The hard part is deciding what to build.

And then why it matters.

And how people should use it.

In that sense, AI is like a super fast car on a road without signs. You’ll go faster, sure. But where are you going?

The new leverage isn't about output. It's about upstream clarity.

Product thinking, user empathy, real problem discovery.

That’s the new hard part.

The belief has to come first. not because mindset magically solves things, but because it’s the only way the hard things even get attempted.

Founders don’t build great products because they know how. they build because they decide it’s possible to figure it out.

Probably, but that’s an assumption.

Write it down, think about what numbers you need to collect/analyze to validate or disprove your assumption, and go for it.

I know it’s no fun, but there’s a reason it’s called work.

However, as I was going through the data, I realized I could make my life easier by creating new data! At first I didn’t want to stop my analysis, but then I spent an hour refining the data we collect.

And suddenly, I started to see patterns in the new data that I would’ve missed otherwise.

It pays off to spend time with making data, not just with analyzing it.