How to measure content design impact when there's no baseline

Measuring content design impact is hardest when you inherit a project mid-flight and there is no baseline. The usual response is to give up on proof and fall back on opinion. That is the wrong move.

You do not need a pre-launch baseline to measure impact. You need one defensible measure you can stand behind, even if it is narrower than the one you wish you had.

Most content designers don't get to start at the start. The site is live, the content is already a mess, and nobody recorded how it performed before it got that way. The absence of a clean "before" is not a reason to stop measuring. It is a reason to measure the diagnostic state precisely, right now.

Why "no baseline" stops projects from proving value

The "no baseline" excuse treats evaluation as a phase that should have happened earlier and now can't. Evaluation is a practice, not a phase. You design measurement intentionally, and you measure something defensible from day one, even if that is day one of your involvement rather than day one of the project.

When a team accepts "no baseline" as a reason not to measure, two things follow. The work goes unproven, so the next budget conversation starts from zero. And the loudest opinion in the room wins, because there is no evidence to argue with it.

A project that cannot prove its value will be asked to justify its existence on someone else's terms. Measurement is how you keep that conversation on yours.

Three things you can measure from day one

You can start measuring the moment you arrive. Choose measures that describe the diagnostic state, the mess as you found it, not a hypothetical before.

Time-to-task on a defined journey. Pick one real user task, such as "find out if I'm eligible and apply." Measure how long it takes to complete now. Time it yourself against the live content, or watch five people try. That number is your starting point, and you can re-run it after the change.

Support volume on a defined topic. Count the calls, emails, or webchat sessions about one specific thing over a fixed period. Narrow it hard: not "enquiries," but "questions about whether a concession applies." Support volume is the clearest signal that content is failing to do its job.

Content density. Measure words per page, or words per task across a journey. A task that takes a user through 4,000 words across nine pages has a number you can cut and report against. Density is easy to capture and easy to explain to an executive.

Which proxies are defensible, and which aren't

A defensible measure is one you can explain in a sentence, repeat the same way twice, and connect to a real user outcome. If it fails any of those three tests, it will not survive scrutiny.

Defensible: support volume on a named topic, time-to-task on a named journey, the number of pages a person must read to finish one task. These describe whether someone can act.

Not defensible: page views, because traffic is not comprehension. Bounce rate without context, because you cannot tell a satisfied leaver from a frustrated one. A reading ease score standing in for plain language, because the Style Manual is about whether a reader can find, understand, and use information, not about syllable counts.

The before-and-after that doesn't need a before

Here is the part teams miss. You don't need an old baseline if you have a precise present one.

Record the diagnostic state with a defensible measure, change the content, then measure again the same way. Your first measurement is the baseline. It is honest, it is dated, and it belongs to the work you actually did rather than to a measurement window that closed before you arrived.

This also protects you. A baseline you established during diagnosis is one you can describe accurately and defend in a review. A reconstructed "before" you never actually recorded is not, and someone will eventually ask.

The Digital Service Standard expects teams to measure performance and keep improving. It does not ask you to have measured before you started. It asks you to start.

So the question is not "what was the baseline?" It is "what is the most defensible thing I can measure today, before I change anything?" Answer that, write the number down, and you have proof on the way.

Photo by Milad Fakurian on Unsplash.