There's a moment in every innovation cycle that doesn't get enough attention.

Not the launch. Not the post-launch review. Not even the research phase.

It's the twelve months before the product hits shelf — when a series of commitments are made, one by one, each one making the next harder to reverse.

A production partner is selected. A minimum run is agreed. Packaging is finalised. A retailer conversation begins. A launch budget is signed off. Distribution is committed.

By the time a consumer first encounters the product, the organisation has already decided — repeatedly, and at significant cost — that this is the right thing to launch. The evidence the market will provide arrives later. As a sales velocity report. A ranging review. A market share figure.

At that point, it isn't evidence. It's a verdict.

This is the structural problem at the heart of most innovation failure. Not bad ideas, not bad teams, not bad research. Bad timing. The commercially decisive evidence — what real consumers actually do when confronted with a real purchase decision — arrives at the only moment it can't change anything.

The 80% NPD failure rate is widely cited. What's less often noted is that most of those products passed internal validation before they launched. They had positive concept scores. They had internal alignment. They had a champion who believed in them.

What they didn't have was real behavioural evidence — gathered early enough to act on.

The conventional response to innovation failure is to improve the research. Better surveys. Larger samples. More sophisticated modelling. But the problem isn't the quality of the research instrument. It's the moment at which insight is translated into commitment.

Most organisations treat launch as the test. The market will tell us if this works. And it will — but by then, the investment is in. The irreversible bets have been placed.

The brands that consistently outperform on innovation aren't necessarily the ones with better ideas or bigger budgets. They're the ones that have moved real behavioural evidence earlier in the process — to the point where it can still change something.

That's the shift worth making. Not better research after the commitment. The right evidence before it.