Abtin Aghagolian, co-founder and chief technology officer of Pikd, a London-based spatial computing company. Photo credits: Abtin Aghagolian

Earlier this summer, a panel of expert evaluators assessed my company's application to a major European innovation fund. Most of them rated the innovation itself very good. The application was declined anyway, and the reasons had almost nothing to do with the technology, but with the evidence which, ironically, the grant itself would have paid us to produce.

I am not writing to complain about the decision. The assessors applied their criteria correctly, and some of what they asked for was fair to ask. I am writing because of what happened next, which is that my small team began preparing a resubmission, and I watched us divert some of the scarcest resource we have, senior attention, away from building the thing the grant exists to fund and toward producing documentation about it. 

We had done this before, more than once, and from the inside it never felt like a choice.

This is the treadmill, and I want to describe its mechanics precisely, because it is usually discussed as if the problem were bureaucracy. It is not paperwork volume that does the damage. It is three structural features that compound.

Asking for evidence before there is any

The first part of the treadmill is that the application demands evidence that the grant is meant to produce. An early-stage deep technology company is asked to demonstrate commercial traction, a complete team and validated performance, which are approximately the things the grant would pay to produce. 

The companies that can already show them need the money the least. This is not a paradox the evaluators fail to notice; it is a rational response to their own accountability, since funding a company with evidence is defensible and funding a promise is not. 

But the aggregate effect is that public money designed to absorb risk migrates toward applications where the risk has already been absorbed by someone else.

A no is an invite to resubmit

The second part is that having a bid declined does not release you. Many European instruments allow resubmission, and some cap the number of attempts before excluding an applicant for good. A rational founder therefore treats a refusal with a strong innovation score as an instruction to try again with better packaging, and the system treats persistence as signal. 

Each cycle consumes months of preparation for a binary outcome, and each cycle is individually justifiable, which is exactly what makes the treadmill hard to step off. The sunk costs are not financial, they are narrative: with every iteration the company gets better at describing itself and no better at anything else.

Encouraging professional applicants

The third feature is the quietest. Grant competence compounds. A company that has been through three application cycles has learned the vocabulary, the evaluation psychology and the portfolio logic of the funder, and that knowledge transfers to the next application in a way that engineering progress does not. 

Over time, the population of successful applicants is enriched not with the best technology but with the best-adapted applicants. Anyone who has sat on both sides of these processes knows the type: the consortium that exists because the call exists.

What can funders do?  

Third, and this one costs nothing: publish, per instrument, how many awardees had professional grant-writing support. Funders have this information or could ask for it. If the fraction is high, that is not evidence of a healthy competition for the best ideas. It is evidence of a market in access, paid for out of exactly the early-stage budgets the grants exist to supplement.

Despite all this, we will probably resubmit. That is the honest disclosure at the bottom of this argument, and it is also the strongest evidence for it. I can see the treadmill clearly, I can name every one of its mechanisms, and the rational move for my company is still to get back on it. A system that produces that sentence from its participants is not selecting for what it thinks it is selecting.

Abtin Aghagolian is co-founder and chief technology officer of Pikd, a London-based spatial computing company.

Editor's note: The author, who is not a native English speaker, used an AI assistant to aid his writing. Science|Business editors then edited the text for length and clarity.