What to Do With a Claim You Can't Back Up

You have a line on your resume that you believe is true and cannot demonstrate. Maybe the numbers lived in a system you no longer have access to, maybe the result was collective and your share of it is unclear, maybe you simply cannot remember well enough to defend the detail. There are three honest responses: narrow the claim to the part you can support, replace it with a different claim about the same work, or cut it. Which one applies depends on why the evidence is missing.

First, work out which kind of gap this is

The word “unverifiable” covers several different situations that call for different answers.

You know it happened but cannot produce a record. You shipped the feature; the internal wiki is gone. The claim is sound and the proof is inaccessible. This is the mildest case, and narrowing usually fixes it.

You know something happened but not the specifics. The process got faster; you have no idea by how much. Here the general claim is safe and every specific attached to it is guesswork.

You are not sure of your own role. The project succeeded and several people contributed. Your memory says you were central; you have not tested that memory against anyone else’s. This is the case that most often needs a different claim rather than a narrower one.

You are repeating something a tool wrote. The bullet came out of an AI draft and you accepted it because it sounded right. This is not a claim you cannot back up so much as a claim you never made. Delete it and start from what you remember, since generated drafts drift in predictable directions and the drift is what you are looking at.

Option one: narrow it

Narrowing means keeping the subject and cutting every element that outruns your evidence. It is usually the best outcome because you keep the substance.

Take “reduced onboarding drop-off by a fifth in the first quarter.” If you cannot source the figure, the narrowed version is about the change itself: what you altered, why, and what it was designed to do. “Rebuilt the three-step onboarding flow into a single screen to reduce abandonment” makes no claim about outcome and is entirely defensible, because the artefact exists and you built it.

Notice what survives. The reader still learns the domain you worked in, the kind of problem you take on, and the scale of the change. What they lose is a result figure, and a result figure that you would have to admit you cannot substantiate was doing you no favours.

Option two: swap the claim for a nearby one you can prove

Sometimes the interesting thing about a piece of work is not the thing the draft chose to highlight. If you cannot evidence the outcome but you can evidence the responsibility — you were the only person on it, you presented it to a named committee, it became the standard approach afterwards — write that instead.

This is not a downgrade. Responsibility claims are often more informative than outcome claims, because outcomes are shaped by things outside anyone’s control while responsibility describes what you were trusted with. A bullet saying you were the sole maintainer of a system for two years tells a reader something durable.

Option three: cut it

Cut when the claim’s only content is the unverifiable part. If a bullet exists to carry a percentage and the percentage is going, there is nothing left. Cut also when you find yourself constructing an argument for keeping it — the effort of the argument is the signal.

The reflex against cutting is that the resume gets shorter. In practice a page with fewer, firmer lines reads as more confident than a page padded with claims of uncertain provenance, and it gives you space to develop the work you can actually discuss. Deciding what to cut first is a normal part of the process, not an admission of a thin career.

What not to do

Do not hedge into meaninglessness. “Helped contribute to significant improvements in key metrics” is not more honest than a specific claim, it is less useful and equally unverifiable. Vagueness is not a form of accuracy.

Do not move the claim somewhere less visible. Burying a figure you cannot support in a summary line or a skills section does not change its status. If it would fail a question in the interview, it fails wherever it sits.

Do not attribute it to someone else’s estimate. “Estimated to have saved considerable time” invites the obvious follow-up of who estimated it, and if the answer is nobody, you have added a second problem to the first.

Do not assume nobody will ask. Whether a given claim gets tested depends entirely on who reads the resume and how the process is run, and you have no way to know that in advance. Write as though the person who cares most about detail is the one reading.

Recording the decision

When you narrow or cut a claim, note in your own facts file what the original claim was and why you changed it. This sounds fussy but it prevents the same sentence reappearing in a later draft, and it prevents you talking yourself back into it in six months when the memory has faded further and the phrasing sounds harmless again.

It also means that if you later regain access to the evidence — a former colleague confirms the figure, an old report turns up — you know exactly which line to restore and what it should say. Sometimes the proof does resurface, and sometimes it never will, and keeping track of which is which is part of maintaining a resume you can defend for years rather than for one application.