The Amnesia Tax: What Forgetting Costs Your Organization

Your organization is paying a tax it never voted for. It shows up every time a team debates a question that was already settled, rebuilds an analysis that already exists, or repeats an experiment that already failed. I call it the amnesia tax, and most organizations pay it every single week without noticing.

Here is the scene. A product team spends forty-five minutes in a roadmap review arguing about whether to build a native mobile app. Good arguments on both sides. Reasonable people. The debate ends without resolution, and someone volunteers to “pull some data.” What nobody in the room knows is that the same organization had this exact debate eighteen months ago. There was research. There was a decision. There was even a document. The people who made that decision have since changed roles, and the document lives in a folder nobody opens. So the team pays for the decision twice. Sometimes three times.

That is the amnesia tax in its simplest form: the cost of re-deciding, re-researching, and re-learning things your organization already knew.

Why the tax keeps going up

Three forces are pushing the rate higher.

First, turnover. When knowledge lives in people’s heads and people leave, the knowledge leaves with them. The median tenure at many tech companies is now shorter than the lifespan of the products those companies build. Your product will outlive the memory of why it was built the way it was.

Second, tool sprawl. Decisions get made in Slack threads, Zoom calls, Figma comments, Jira tickets, and hallway conversations. Each tool remembers its own fragment. No tool remembers the decision. The context is technically “somewhere,” which in practice means nowhere.

Third, and this is the one that surprises people: AI is accelerating the problem before it solves it. Teams can now generate analyses, prototypes, and documents faster than ever. Output is up. But volume without memory means you are forgetting faster, at scale. An organization that produces ten times the artifacts and remembers none of the reasoning behind them has not gotten smarter. It has gotten louder.

The part leaders get wrong

When leaders notice the symptoms, the reflex is almost always the same: mandate more documentation. Write everything down. Create the wiki. Require the decision doc template.

I have watched this fail repeatedly, and it fails for a predictable reason. Documentation is a storage strategy. The amnesia tax is a retrieval problem. Nobody in that roadmap review chose to ignore the old research. They did not know it existed, and no reasonable amount of searching would have surfaced it at the moment it mattered. Writing more documents into a system nobody can query at the point of decision just raises your storage bill without lowering your tax.

The teams that actually reduce the tax do something different. They treat decisions, not documents, as the unit of organizational memory. A decision has a shape: what we chose, what we rejected, what evidence we weighed, who was accountable, and what would make us revisit it. When you capture that shape and make it findable at the moment a similar question comes up, the forty-five minute debate becomes a five-minute review: “We decided this in 2024 under these conditions. Have the conditions changed?” Sometimes they have, and you re-decide with full context. Usually they haven’t, and you just saved a meeting, a research cycle, and a quarter of drift.

How to estimate your own tax bill

You do not need a consulting engagement to size this. Try a simple exercise with your leadership team. List the last ten significant decisions your organization debated. For each one, ask two questions: had we decided something like this before, and could anyone in the room have found that prior decision in under five minutes? In my experience running this exercise, most teams find that a third to a half of their “new” debates were reruns, and almost none of the prior decisions were findable. Multiply that by the loaded cost of the people in those rooms and the delay cost of the decisions themselves. The number is uncomfortable. It is supposed to be.

The organizations that win the next decade will not be the ones that decide fastest. Speed is cheap now; AI made sure of that. The winners will be the ones that never solve the same problem twice.

I am writing a book about this, and I will be sharing more of the thinking here as it develops. In the meantime, run the ten-decision exercise. Then tell me what your tax bill looks like.

Dave Mathias helps organizations improve decision quality and build decision memory. If your teams keep having the same debate, let’s talk.