Website benchmarking before a redesign: what are we moving?

Stage one of eight. The question it answers: what are we trying to move, and where does it stand today?

A manufacturer's brief said the website drives most of its sales opportunities and leadership wants more, especially from an audience the site has under-served. That is a real ambition and it is not a target. A target is a number, baselined today, with a definition everyone agrees on, that will be read again in a year.

Most website redesign goals never get that far. They stay as adjectives, modern, cleaner, better converting, and the redesign ships against adjectives. A year later the debate about whether it worked is a debate about taste, because there is no baseline to hold it to.

Website benchmarking is the discipline of turning the ambition into numbers before any design work starts. It is the first of eight stages because everything downstream is judged against it, and the last stage, proof, is impossible without it.

There are two numbers hiding in the question

When a business says it wants more leads from the website, there are usually two numbers in that sentence, and they have to be baselined separately.

The first is the total: qualified leads the site sends to sales or to locations. The second is the share: how much of that total comes from the audience the business is trying to grow. A manufacturer growing its professional channel needs both, because a redesign can raise the total by winning more of the existing audience while the growth audience's share stays flat, or raise the share while the total dips. If only one number is baselined, a year later nobody can tell a genuine win from a shift in the mix.

The rule generalizes: one number per audience the ranking in stage two names, plus the total, each with its own definition and source. Website KPIs that blend audiences hide the thing the redesign was for.

Website benchmarking template: for each metric, the definition, the source system, the current read, whether the tracking is trusted, and an owner, with audiences baselined separately from the total.

Check whether the tracking can be trusted before writing a number down

A baseline read from untrustworthy tracking is worse than no baseline. It looks like evidence.

On a London university's site, the board wanted a conversion number it could hold the project to. Before offering one, the team checked what the analytics actually recorded, field by field, and found bot traffic inflating the homepage, which made the site look busier at the top of the funnel and leakier in the middle than it was. There was no systematic conversion tracking, so benchmarking was impossible until the gaps were closed. The honest baseline was lower than the reported one and far more useful, because it was real.

The check is simple to describe and rarely done. For each number in the template: where does it come from, is every step it depends on actually recorded, is non-human traffic removed, and would two people reading the same report get the same figure. If any answer is no, the number goes in the template marked untrusted, with an owner for fixing it, and the redesign does not get to claim a result against it until it is.

Separate what the redesign changes from what it does not

Not every metric the business tracks will move because of the redesign, and some will move for reasons that have nothing to do with it. Website benchmarking has to draw the line between the two, or every change after launch gets credited to or blamed on the new site.

The university's undergraduate and postgraduate journeys were baselined separately because they behave differently and respond to different things. On a certification body's program, the conversion baseline was built with separate views per funnel, learning, professional, shop, and travel, so that a structural change to one journey would show up in one place. The general principle: baseline per journey, and record which journeys the redesign will change structurally. Those are the ones stage eight measures. The others belong to the internal team's ongoing program.

Give the people approving the budget scenarios, not a promise

Stage one is also where the redesign's terms with its funders get set, and the temptation is to promise a number. On the university project the investor group behind the institution expected exactly that.

The team declined and offered three scenarios instead, conservative, expected, and strong, each labeled illustrative with its assumptions written down, alongside the day-one targets the launch would be held to and a post-launch program that would do the moving. That is the honest shape of a target when the tracking has gaps: a method the board can hold you to, rather than a guess it will remember when it misses.

The target decides what the strategy is for

A website redesign project plan tells you when things happen. The target tells you what they are for. Everything that follows in the eight stages is judged against it: the audiences are ranked to move it, the journeys are mapped to find where it leaks, the structure and discoverability and conversion work are scoped to move it, and the proof stage reads it again.

Every business measures success differently. For some it is leads and share. For others it is enrolments, memberships, bookings, or a brand outcome that is harder to count and still has to be defined. The point of stage one is not the metric. It is agreeing the measures at the start so they are not discovered, or invented, at the end.

What the target stage hands over

The baseline template: each metric with its definition, source, current read, trust status, and owner.
The two numbers per audience, and the total, baselined separately.
The tracking audit: which steps are recorded, which are not, and the plan to close the gaps.
The journey split: which funnels the redesign will change structurally and which it leaves to the internal program.
The three scenarios with assumptions, and the day-one targets.
The agreed measures, signed by the people who will read them again at the proof stage.

What breaks when the stage is skipped

The redesign ships against adjectives, and whether it worked becomes a matter of opinion.

A blended number hides the mix, and a win for the existing audience is mistaken for growth in the new one.

An untrusted baseline is treated as evidence, and the redesign is judged against a fiction.

The board is promised a number, the number misses, and every future ask is discounted.

Where this sits in the sequence

Stage one is the start of the outer loop. It hands stage two the numbers the audience ranking has to serve, stage six the funnels to instrument, and stage eight the baseline to validate. A validated result at stage eight resets the target, and the sequence runs again from here with a better starting point.

Questions we get asked about website benchmarking

What should you benchmark before a website redesign?

The numbers the redesign exists to move, per audience and in total, each with a definition, a source, and a trust check. Usually that is qualified leads or the equivalent endpoint, the share from the growth audience, and the funnel steps that lead to them. Not everything the analytics can report, only what will be read again.

What are the right website KPIs for a redesign?

Fewer than the dashboard offers. The endpoint the business counts, the share from each audience the strategy names, and the funnel steps that feed them, baselined per journey. Vanity measures such as total sessions are context, not KPIs.

How do you set website redesign goals a board will accept?

As scenarios with written assumptions, alongside the day-one targets the launch will be held to and the post-launch program that will move the number. A single committed figure is a guess. Three labeled scenarios are a method.

What if the current tracking cannot be trusted?

Say so, mark the baseline untrusted, assign an owner to close the gaps, and do not claim a result against it until they are closed. A baseline built on bad tracking is worse than none, because it looks like evidence.

Website Benchmarking Before A Redesign: Decide What You're Trying To Move