In 2014 I did two pre-acceleration programmes back to back while running VipKing: Yuzz, Banco Santander's programme for young founders, and NEXT, Powered by Google for Entrepreneurs, in Barcelona.
The company didn't survive. Some of what I learned did — and a decade of watching other people's products has made it fairly clear which parts were real and which were performance.
What held up
1. "What would have to be true?"
The most useful question I've ever been handed. Not is this a good idea — unanswerable, and everyone answers yes about their own. Instead: what would have to be true for this to work, and which of those things is most likely to be false?
It converts an argument about opinions into a list of testable claims, ordered by risk. I use it in nearly every client conversation now, usually within the first ten minutes, and it routinely saves people from building something that had no chance.
2. Your pitch is a diagnostic, not a performance
We spent a lot of time on pitch training and I was cynical about it. I was wrong, but not for the reason they gave.
The value wasn't in pitching better. It was that being forced to explain the business in two minutes to a bored stranger exposes every part you don't actually understand. Every time I got stuck, it was on something I'd never resolved — not something I'd failed to phrase well.
The same thing is true of technical specs. If you can't explain what you're building in two minutes, you don't have a clear enough idea to build it.
3. Talk to the person who signs the cheque
Obvious, universally ignored. We talked to clubbers constantly — they loved it, it was free. We should have been talking to club owners, who had to give up revenue.
In B2B software this shows up as building for the user and selling to the buyer without ever noticing they want different things. It kills more products than any technical decision.
4. Mentors are most useful when they disagree with you
The NEXT cohort had genuinely good mentors — people who'd built real companies in Barcelona. The sessions I remember are the ones where someone told me flatly that a plan wouldn't work. The encouraging ones have entirely evaporated.
This is why I try to give founders the uncomfortable version now, even when the polite version would be an easier sale.
What turned out to be theatre
In fairness to both programmes, some of this is structural rather than anyone's fault.
- The demo day incentive. You optimise for looking impressive to a room, which is a different objective from building something people want. For a few weeks those objectives actively diverge.
- Cohort comparison. Being surrounded by other founders is mostly good, and it also produces a quiet pressure to appear to be progressing. That pressure rewards announcements over evidence.
- The business-plan artefact. We produced elaborate documents that were obsolete on contact with the first real customer. The thinking was valuable; the document was not.
- "Network access" as a headline benefit. Real, but far smaller than advertised. An introduction is not a relationship.
Would I do it again?
Yes — with one adjustment. I'd treat the programme as a forcing function for talking to customers, not as a curriculum to complete. The founders in both cohorts who got the most out of it were the ones who treated the sessions as a deadline to have learned something real by, rather than as the learning itself.
Why I still bring it up
Because it's the reason I can sit on both sides of the table. I've raised money badly, priced wrongly, and built the wrong feature carefully. When a founder tells me the roadmap keeps slipping or the agency quote feels high, I'm not reasoning from theory — I'm reasoning from having made the specific mistake.
I've written up the founder half of that story separately. The short version: the technical skills were the easy part to acquire. This was the expensive part.