WHEN “PICK AN INDUSTRY” IS BAD ADVICE

Every advisor, agency, and pitch coach you've spoken to has told you the same thing. Pick a lane. Niche down or die. Focus, focus, focus.

The advice is right. But the part a lot of people get wrong is what focus actually means.

When advisors say “pick a lane” they almost always mean “pick an industry”. But industry is just one way to focus. You can focus on a buyer type – the developer, the procurement officer, the formulator. You can focus on a use case – one specific problem you solve. You can focus on a stage of the buyer's journey, on a geography, on a problem category. The question isn't whether to focus. The question is on what.

“Pick an industry” is, for most young companies, sound advice (and I've argued that case elsewhere). But there are four shapes of company where industry isn't your axis. None of them mean do nothing. Each still demands focus. Just on a different dimension.

Discovery stage, before product-market fit.

You haven't found the pull yet. You have customers, but not enough of them in any one place to tell you where to focus. The right move here is to stay loose on your ideal customer profile (ICP) – broad enough that the market can show you who's interested. But loose isn't blurry, and it isn't theoretical. Even at this stage you have to speak to specific buyers about specific problems – what the product gives them, not a catalogue of applications the technology might theoretically serve. The longer the list, the less anyone reads themselves in it.

And you can't stay here forever. Once the right pull arrives – one segment keeps coming back, the others don't – you have to orient toward it. Discovery is a phase, not a position. Stay broad past that point and the exception stops being yours.

Discovery-stage positioning is loose on ICP, sharp on each buyer you're actually talking to.

Infrastructure for builders.

You're horizontal across industries because integration is the hardest problem you solve. Your buyer is a builder. Stripe stayed horizontal across e-commerce, marketplaces, software-as-a-service, and fintech. But it chose web and app developers as the buyer – the people who needed to add payments to whatever they were building – and made the integration take seven lines of code. It led with one application, online payments for startups, before expanding. The narrowing happens on buyer type, not industry. The messaging still leads with one application even when the underlying product is fully horizontal. Infrastructure without a sharp lead application reads as theoretical.

The most common mistake is claiming this shape when the application-specific knowledge doesn't actually transfer. I see this constantly. A platform with three applications, where one needs cosmetic-ingredient compliance and brand partnerships, another needs European novel-food approval, and the third needs US agricultural field trials. None of that knowledge transfers across the others. The founders I speak to usually know which application is closest to revenue. They just haven't said so on the homepage. You're not a platform with three applications. You're three industry-deep companies sharing a lab. The good news: you only have to win one to validate the underlying technology. The bad news: you have to actually win it.

The commodity.

You sell something genuinely industry-agnostic. Grain. Sheet metal. Bulk chemicals. Generic foundry capacity. The product is what it is, and the buyer in food, automotive, and construction all want the same thing. Industry can still shape channels and grading – auto-grade steel is sold differently from construction steel – but the differentiation isn't at the application layer. Real commodities are rare in the kind of companies I work with. Most founders are trying to differentiate, not produce a horizontal product. But the shape exists.

However, even commodities have to prioritise sales. You can't be in every channel, talking to every buyer, networking in every direction. That usually means focusing on the accounts your scale lets you serve, the channels where your certifications open doors, and the geographies where your logistics give you an edge.

If your competition is on price, scale, and certifications rather than capability, industry isn't your axis. The work is choosing which tier of commodity, which geography, which certification, which buyer-size band. Even soil has to be the right soil for something.

The mammoth.

You're already a mammoth. Apple, Amazon, NVIDIA, Salesforce. You've earned breadth by winning a sector first, then earning the next one on the back of that proof, repeatedly, for decades. You don't pick a lane because you're already in all of them. If you're reading this, you're not the mammoth – yet. This case exists mainly to close one escape route: yes, Apple makes phones and computers and watches and services and silicon. Apple earned the right to. You haven't, and quoting them won't get you there.

There's a fifth shape, but it's something you graduate into rather than do now. If your tech genuinely serves commercial worlds that share no buyers, no regulations, and no sales motions, the answer is eventually splitting. Twist Bioscience did this in 2025 – spun out their DNA Data Storage application as Atlas Data Storage and refocused on DNA synthesis for pharma. At Series A scale, splitting is risky: two cap tables, more mouths to feed before either entity has proven itself. The realistic move at €5–20M is pick one, prove it, spin out later. The intellectual-property control question that stops most founders here is real, and worth its own blog piece.

Narrowing feels like a constraint, but staying broad isn't free. Tomasz Tunguz looked at 54 public software-as-a-service companies back in 2015. He found that vertically-focused ones produced $0.97 of revenue for every dollar spent on sales. Horizontal ones produced $0.66. Going broad costs you about 31 cents on the dollar in sales efficiency.

The 2020–22 deep-tech platform IPO cohort tells the harder version. Zymergen pitched a bioproduction platform across consumer materials, industrial chemicals, food, and agriculture. It went bankrupt in 2023, and was fined $30 million by the SEC in 2024 for misleading IPO investors about the market potential of its lead product. Its IPO peer Ginkgo Bioworks pitched over 100 programmes across pharma, agriculture, industrial, food, consumer, and defence. It lost most of its share value, reverse-split, divested its biosecurity arm, and by 2026 had refocused on a single platform – autonomous labs. The platform story is cheap to pitch, but expensive to back.

If you've read this far hoping to find your company in one of these four and you can see you're not, the advice everyone keeps giving you isn't wrong. It's for you. Industry is likely your axis, and staying broad before you've broken through is holding you back.

And if “pick a lane” has never quite fit, the lane isn't always an industry.

Find yours.

-

Want a proper look at your positioning?
Send me your site. I’ll come back within a week with the thing I’d change first.