The topic that scares me most as an investor is competitive analysis. Not because competition is fierce (it is), but because it's so easy to get it wrong.

I've looked at a startup and an incumbent operating in the same market, same country, same customer base, and tried to map out how they'd compete.

Then they announced a partnership.

I've watched a founder dodge competitive questions in a pitch and assumed the worst. Only to realize later that the question itself was wrong. The companies I thought were competitors weren't competing at all.

Competition is the thing investors think they understand but rarely do. Including me.

The first trap is the most basic: figuring out who's actually fighting whom. This sounds easy. It isn't.

A recent example: a startup standardizing and cleaning medical data, operating in the same market as a massive incumbent also collecting, cleaning, and monetizing medical data. Same country. Same market. You'd assume they're direct competitors.

They're not. They're working together.

Medical data varies enormously: by type, by where it lives in which provider, by the format in which it's used and for what purpose. Two companies can specialize in different data formats and divide the labor rather than fight for the same territory. Even when data format and location overlap, they may still work together, because their business models and economics are different: "you handle this workflow and I'll handle this other process."

Cybersecurity is another one. You see a startup doing micro-segmentation and think: every incumbent does that too. Game over. But then you drill down. One company focuses on IT environments, the other on OT. Or they both do OT but one specializes in semiconductor manufacturing and the other in medical devices. They're not competing. They're in adjacent verticals that look the same from the outside.

This is why it frustrates me when we default to lazy questions about competition. Not because we're wrong to worry, but because the question is harder than we're treating it.

"How do you compete with Google?" assumes you've already figured out that Google is the competitor. Often, it isn't.

The second trap: even once you've correctly mapped who competes with whom, you still have to assess whether the startup has a chance. The default move is to say the incumbent always wins. Big company, 100x the engineers, distribution, data flywheel. Game over.

If you default to that conclusion, you will miss every single great startup.

This is especially acute in Asia, where investors heavily weight technological superiority and patent defensibility. Any startup that lacks obviously superior technology gets written off as hopeless. And it sometimes means investors back academic geniuses who know nothing about distribution.

That shuts you out of enormous opportunities, especially in software.

The reality is that most software startups, and certainly most generative AI startups today, don't possess durable tech superiority. Look at the big AI labs: What exactly is the durable "tech" moat any of them has over the others? When massive AI labs have competitive advantages that barely last weeks, what makes you think a small startup would have permanent technical moats?

There are ways startups win that have nothing to do with being technologically superior. A few that often get underweighted:

Giants have to cover enormous surface area. They can't execute flawlessly across all of it. A startup that relentlessly focuses on making one thing outstanding, especially if that one thing is a pivotal part of user experience, has a real shot. The incumbent's attention is diluted. The startup's is concentrated. Granola competes against a million other meeting transcription apps, and Google and Zoom offer it to you for free. Yet Granola is still thriving, why? As a user, I'd submit that their user experience and product velocity clearly shows focus + craft. The way the product is designed and communicated to me is deeply considered. From my own portfolio: Rork would seem like any other vibe-coding app. It's not. They're vehemently iOS-only. Everything they build is for users to build and publish better apps to iOS, and I made a bet on their focus, essentially.

If a startup focuses on a segment of the market that's a rounding error in the incumbent's revenue base, the giant has no incentive to fight hard there. The startup out-executes in that sliver, earns customer trust, then expands. Beyond Health is an example: it focuses on combining primary care clinics in the US, and especially 5~10 location groups that bigger players neglect. As they succeed in integrating these and making measurable improvements in revenue and economics, they gain more power and expansion accelerates.

This is one that's dear to me: as you shift from one generation of users to the next, products get abandoned simply because they feel "old." This happens at different speeds across different markets. Social networks are on the faster end.

Remember how quickly Facebook became "the social network my parents use"?

B2C productivity software too: digital natives want the digital environments they work in to feel new. Sometimes even old. Non-technical knowledge workers are choosing the terminal now.

Technology is changing. Business models are changing. Buying patterns are changing. User behavior is changing.

When everything is changing at once, nobody's pattern-matching works. Including mine. The patterns you learned last cycle aren't wrong. They're expired.

The investors who get competition right in this era won't be the ones with the best frameworks from 2015. They'll be the ones willing to admit their frameworks might not apply anymore.