Where startup opportunities actually come from

Ideation is the least respected stage of company creation and the one that quietly determines the ceiling of everything that follows. A venture built on a weak opportunity can be executed brilliantly and still be worth very little. A venture built on a real structural opportunity can survive a great deal of imperfect execution. Within the GCV Labs Venture Builder Process, Ideate exists to produce a shortlist of opportunities worth the cost of proper evaluation — nothing more, and nothing less.

The common failure is treating ideation as a creativity exercise. Teams gather, generate a hundred ideas, vote, and leave with a favourite. What they rarely leave with is an explanation of why that opportunity exists now, why it has not already been taken, and who would pay for it being solved.

The four sources of credible opportunities

Across the ventures we have created and co-founded, the opportunities that survived evaluation came from a small number of recognisable places. We use them as a checklist during Ideate rather than as a theory.

Two of these — regulatory change and technology shift — explain why an opportunity exists now. Two of them — structural inefficiency and domain insight — explain why it has not already been taken. Strong opportunities usually have at least one from each pair. Intelligence Fusion, which we co-founded and built through to exit, sat at the intersection of a domain insight about how organisations consume threat information and a technology shift that made continuous, structured intelligence deliverable as software.

Idea, opportunity, business

It helps to be strict about vocabulary, because the three words are used interchangeably and they are not the same thing. Being precise about which one you are holding tells you what to do next.

Most things presented as businesses are ideas. That is not a criticism — it is the normal state at the start. The mistake is skipping the two steps in between, which is exactly what the Evaluate stage exists to prevent.

How to run ideation so it produces useful output

We generate ideas using structured brainstorming and mind-mapping, focused on sectors where we have real domain experience and where technology or regulatory change is creating movement. The structure matters more than the volume. A hundred ideas from a sector nobody in the room understands is worse than five from a sector three people have worked in for a decade.

1. 01Choose the sectors first, not the ideas. Constrain to areas where the team has genuine domain depth.
2. 02Ask what has changed in the last twenty-four months — in regulation, in technology, in cost, in behaviour.
3. 03For each candidate, write one sentence on who has the problem and one on why now.
4. 04Reject anything where nobody can name the budget holder.
5. 05Shortlist three to five for evaluation, and write down what would have to be true for each to work.

Why ideation belongs to a process, not a moment

The advantage a venture builder has at this stage is not better ideas. It is that ideation is a repeatable activity connected to everything downstream. Because the same team runs Evaluate, Create and Incubate, the standard applied during ideation reflects what it actually costs to build and prove a company. Ideas that would sound exciting in isolation get held to the standard of the people who will have to deliver them.

That is one of the practical differences between a venture builder and a capital-first model, which we cover in more depth in venture builder vs venture capital and in what is a venture builder?.

The output of Ideate should be uncomfortable in a specific way: a small number of opportunities that look genuinely promising and that nobody in the room can yet prove. That is the correct handover into Evaluate, where the job stops being generation and starts being disproof.

Testing an idea before you spend anything

The source an idea came from tells you the cheapest way to test it, which is why we insist on naming the source before anything else. A regulatory opportunity is tested by reading the instrument and speaking to the people who will be held accountable under it — the deadline is public, and so is the obligation. A technology shift is tested by building the smallest working thing that demonstrates the new workflow, because the argument is almost never accepted in the abstract. A structural inefficiency is tested by measuring what the current workaround costs the buyer in staff time. A domain insight is tested by finding out whether people inside the industry agree with it, and paying attention to the ones who do not.

None of these tests requires a product, a company or capital. They require a fortnight and a willingness to hear an unwelcome answer. Ideas that survive them arrive in Evaluate with something better than enthusiasm attached: a first piece of evidence and a named person who confirmed it.