Who should have final sign-off on our website redesign?

Pick one person to hold final sign-off. Name them before design starts, write down what they can decide alone, and tell everyone else that their input is advice. Not a committee. Not the CEO by default.

Here is why this matters more than it sounds. Design work almost never stalls because someone dislikes the work. It stalls because nobody in the room is allowed to end the disagreement. Two directors want different homepages, both are senior, both are reasonable, and the studio waits.

In our projects that wait runs two to five weeks. On a mid-sized redesign that is a meaningful slice of the schedule, and it comes out of one of two places: the launch date, or the build. Launch dates are usually defended, so it comes out of the build. Compressed build time is where the quality actually goes — the fifth breakpoint that never got checked, the empty states nobody drew, the form error message written in a hurry. Unresolved sign-off does not look like a fight. It looks like a slightly worse website.

Everyone involved in a redesign is doing one of three jobs, and trouble starts when one person is quietly asked to do two.

The Decider is one person. They break ties, they sign off, and they own the consequence if the page underperforms. Advisors bring expertise or constraints — legal, sales, the person who actually answers support emails — and they get consulted properly, but they do not hold a veto. The Informed see the work so it does not surprise them at launch, and they do not shape it.

Merge the Decider with the Advisors and you have a committee, which means the strongest personality decides rather than the accountable one. Merge Advisor with Informed and people stop giving you their constraints, because they no longer believe anyone reads them. Give one person the consequence but not the authority and you get the marketing lead we meet on most projects: responsible for the numbers, unable to approve a headline.

The commonest broken pattern is subtler than any of those. Everyone is treated as a Decider in the review call — all opinions weighed equally, no ties broken — and then treated as Informed in the follow-up email, which simply reports what happened. Nothing was ever actually decided. It was just described afterwards.

Use a test rather than a job title. The Decider is the person who owns the outcome the redesign is meant to move. If the brief says demo requests, it is whoever is measured on demo requests. If it says fewer support tickets about pricing, it is whoever carries the support queue.

That rules out three tempting candidates. Not the most senior person in the building. Not the person with the strongest aesthetic opinions. Not the person who commissioned the project, unless they also carry the number.

The CEO is usually the wrong default, and it is worth saying why plainly. They are the hardest person to schedule, so every decision waits for a diary. And their taste is the least tested against the numbers in the company, because nobody has ever told them their homepage instinct was wrong. A Decider who is unavailable for nine days is worse than a Decider with slightly worse taste.

There is a real exception. In a company under roughly fifteen people, the founder genuinely is the brand. Their judgement about tone is not a preference, it is the product of every sales call they have ever had. Delegating sign-off there is theatre — the work will get reopened anyway, just later and more expensively. If that is your situation, name the founder as Decider and protect their review time instead of pretending otherwise.

Sign-off holds up when it is written down before anyone has an opinion to defend. It does not need to be long. Four lines is enough, agreed in writing, shared with everyone who will see the work.

A remit we would recognise from a recent project looked close to this:

• Decider: Priya, Head of Growth. She signs off page structure, copy, layout and visual direction.
• Needs a second yes: legal on claims and terms, accessibility obligations, use of the brand marks.
• Out of scope for opinion: the three primary audiences, the decision to lead with pricing, the launch date. Settled in the brief.
• Everyone else: advice welcome in review, gathered by Priya, decided by Priya.

They will. A remit does not stop two directors seeing the homepage differently. It just gives you a way through, and the order matters.

First, convert both positions from preference to claim. “I don’t like the hero” becomes “I think the hero doesn’t tell a first-time visitor what we sell.” Preferences cannot be resolved, because there is nothing to be right about. Claims can.

Second, ask which claim costs more if it turns out to be true. That question moves the room off taste and onto risk, and it is usually answered in a minute.

Third, if both claims are cheap to test, stop arguing. Ship the Decider’s choice and put the other on a measurement list with a date. A disagreement parked for six weeks costs nothing. The same disagreement litigated live costs a fortnight.

Fourth, notice when it is not a design question at all. If the two positions differ about who the company sells to, that is positioning, and design review is the most expensive room in which to have that conversation. Take it out, settle it with the people who own it, bring the answer back.

These recur often enough across our projects that we now ask about them at kickoff.

The phantom stakeholder never attends a review, reads the final PDF, and reopens everything. The fix is not more invitations — it is deciding whether they are an Advisor, in which case they attend or forfeit, or Informed, in which case their comments go on the measurement list.

The delegated-then-undelegated decision is the most damaging. A Decider is named, then overruled once by someone above them. After that, nobody believes the role exists, and every future decision quietly goes back to being a committee. One override undoes the whole structure.

The unanimity trap is ours as much as yours. A studio waits for everyone to be happy before moving on. No page in the history of the web has achieved that. Waiting for unanimity is how you turn a decision into a delay.

Reopening should be allowed. It should also be expensive on purpose, because free reopening means nothing was ever signed off.

Our rule: a signed-off page can be reopened by the Decider with a stated reason, and we quote the cost and schedule impact before the work starts rather than after. The quote is the point. It converts “can we just revisit the homepage” into a decision with a price attached, and about half the time that is enough to answer it.

Reasons that justify reopening look like this in practice: new data from the site or from sales, a commercial fact that changed, a genuine mistake we or you made. Reasons that do not: someone senior saw the page for the first time late, or a founder showed it to a friend over the weekend. We have watched a signed-off homepage reopened on a friend’s opinion cost three weeks. The friend was not wrong about much, and it did not matter.

A studio can manufacture this problem. If we present three directions and no recommendation, we have handed you a committee problem we should have solved ourselves. Three unranked options is not choice, it is abdication — we have asked the least informed people in the room to do the hardest part of our job.

A good studio arrives with a recommendation and the reasoning behind it, and defends it. You are entitled to ask for that, and to be suspicious when you do not get it.

We should also say the obvious thing about incentives. Murky sign-off is good for agencies. Change requests are billable, and delay caused by your own indecision is the easiest delay in the world to defend. We are telling you to fix it because a project with a clear Decider is cheaper for you and better for us. We would rather lose the extra invoices than lose the launch date.

The problem you started with was a round of feedback going in circles. It goes in circles because nobody is permitted to stop it. Four things fix that, and all four can be done this week.

Name the Decider using the outcome test. Write the four-line remit and send it. Tell every other person which role they hold, in words, so nobody discovers it mid-review. Agree what reopening a signed-off page costs.

A redesign does not need everyone to agree. It needs one person allowed to decide, and everyone else to know that is what is happening.