The Sunk Cost of Not Documenting Your Work
· 5 min read
There is a type of sunk cost nobody talks about.
Not money. Not time. Not missed opportunities.
The sunk cost of work that happened but left no trace.
The Work That Cannot Be Cited
Between 2012 and 2018, I ran somewhere around 60 industrial projects through Witanabe. Pump installations, fluid system audits, maintenance contracts, diagnostics for factories I still pass on the highway. I know the work was good because clients came back. I know it was varied because the problems ranged from cavitation in cooling systems to corrosion in chemical transfer lines.
What I cannot do is cite most of it.
No case studies. No documented methodology. No photos of the before and after. In some cases, not even a written record of what the problem was and what we did to fix it.
The work is done. The money was collected. The projects are over.
And from a credibility standpoint, those 60 projects might as well not have happened.
What "Sunk Cost" Actually Means
The standard definition of sunk cost is money already spent that cannot be recovered. The reason you shouldn't factor it into future decisions.
But there's a version of this that runs in reverse.
Every project you do creates potential future value, not just in the direct output, but in what you could reference later: the case study, the methodology, the proof. If you don't document it, that potential value is permanently destroyed. Not gradually. Immediately, at the moment the project ends and you move on without writing anything down.
The cost was never money. The cost was attention. You spent attention doing the work. If you don't spend a fraction more attention documenting it, the entire investment loses its compounding potential.
That is the sunk cost nobody talks about.
Compounding Versus Evaporating
Here is what documented work does over time.
A case study written in 2015 becomes a reference you can point to in a 2020 proposal. That 2020 proposal wins a contract that leads to a project you document thoroughly. That 2022 documentation becomes a published methodology. That methodology is cited by an institutional client in 2024.
None of that chain exists if the 2015 case study was never written.
Undocumented work does not stay neutral. It actively degrades your position. Because when you compete for a contract in 2024 and someone asks for evidence of your track record, you have to speak from memory. Memory is not verifiable. Verifiable evidence is what closes enterprise contracts.
The gap between what you have done and what you can prove you have done is exactly the gap between a freelancer and an authority.
Why Practitioners Don't Document
I know why I didn't document systematically in the early years. Several reasons, none of them good in retrospect.
First: it felt like administrative overhead. The work was the work. Writing about the work felt like doing extra work for no immediate payoff.
Second: there was a tacit assumption that clients would remember and recommend. Word of mouth as the documentation substitute. This works until it doesn't, and it stops working the moment you want to move upmarket.
Third: some of the work was under NDA or informal confidentiality. This is real and valid, but it's also often used as a blanket excuse that covers far more than the actual restriction requires. You can document methodology without disclosing client identity. You can describe a problem class without naming the specific company.
Fourth, and most honest: I didn't think anyone would care. The work was visible to clients. That felt like enough.
It was not enough.
What Documenting Actually Requires
This is where I want to be specific, because the prescription "document your work" without parameters is useless.
You don't need a case study per project. You need a record per project.
A record means: what was the problem, what did you do, what was the result. Three to five sentences. A photo if you have one. A date. That's it. You can do this in ten minutes at the end of a project day.
The case study comes later, when you have time or when the project is important enough to write up properly. But the record gives you raw material. Without the record, the case study is impossible.
For Witanabe, I now maintain project logs. For Hibrkraft, EFEO and KPK engagements are documented with enough detail to appear in proposals. For the entity infrastructure work, every methodology step is written down before I run it anywhere.
Not because I have more time now. Because I understand what undocumented work costs.
The Entity Problem
There is a second reason documentation matters that is more specific to the current moment.
AI systems and search engines build understanding of who you are from what they can find and verify. If your work is not documented in findable, structured, verifiable form, you effectively do not exist to those systems.
This is not a future problem. It is the current state.
A consultant with a published methodology, documented client outcomes, and ORCID-linked research exists. A consultant with twenty years of excellent work and no public record does not exist, not in the way that matters for enterprise-level credibility infrastructure.
The 60 projects I cannot fully cite are phantom work. Real to me. Invisible to the systems that clients use to evaluate who to trust.
The Fix Is Boring
Nothing in this essay is original. People have said document your work for decades.
The reason it keeps needing to be said is that the payoff is delayed and the cost of not doing it is invisible until it isn't.
You feel the absence of documentation the first time a potential client asks for a case study and you cannot produce one. Or the first time you are competing against someone with a thinner track record than yours but a better documented one, and they win.
At that point, the work is already done. The sunk cost is already sunk.
The only direction available is forward: start documenting what happens from now on, and spend whatever time you can afford to reconstruct what happened before.
Slow process. No shortcut. But it compounds.
That is the only thing that matters.