1% better every week: treating a site as unfinished
Launching a site feels like an ending. The work is done, the team moves on, and the site is left to fend for itself while everyone gets back to the roadmap. It is an understandable instinct and an expensive one. The sites that pull ahead are rarely the ones that launched best. They are the ones that were treated as unfinished, improved a little every week, until small changes compounded into a lead that is very hard to catch.
What does one percent a week add up to?
The maths is quietly dramatic. Improve something by one percent a week and, with compounding, you are around seventy percent better over a year. No single change feels significant. A clearer headline here, a shorter form there, a section reordered because the data showed people never reached it. On their own, none of these are worth a press release. Together, over months, they are the difference between a site that converts and one that keeps converting.
Why do most sites stop improving after launch?
What makes this hard is not effort. It is attention. Most teams do not have a process that keeps looking at the site after launch, so the site stops being looked at. The data sits in a dashboard nobody quite reads. Pages that should be performing are not, and no one notices, because noticing was somebody's job for the six weeks of the build and nobody's job afterwards. The fix is not heroics. It is a rhythm.
What does a weekly improvement loop look like?
That rhythm is simple to describe. Find where people are dropping off, using the data rather than instinct. Change one thing to address it. Measure whether it moved. Then do it again. The discipline is in the loop, not in any single change, and the loop is what most sites never establish. It is also what makes the improvement honest, because every change has to earn its place by moving a number, not by looking more modern.
When is a full rebuild the right call?
There is a temptation, especially after a raise, to reach for the big dramatic rebuild instead. Sometimes that is genuinely the right call. But often the rebuild is a way of avoiding the less glamorous work of paying attention, and the new site launches to the same neglect the old one suffered. A site treated as unfinished rarely needs the dramatic rebuild, because it never falls far enough behind to require one.
The reframe is small and it changes everything. A website is not a project you complete. It is an asset you compound. The teams that understand this do not win because they launched something better. They win because they never stopped.
FAQs
What is conversion rate optimisation?
What is conversion rate optimisation?
Conversion rate optimisation, or CRO, is the work of increasing the share of visitors who take the action a page exists for, such as booking a call or starting a trial. It runs as a loop: find where people drop off, change one thing, measure whether it moved, and repeat.
How much traffic do you need to run A/B tests?
How much traffic do you need to run A/B tests?
For a classic split test, roughly a few thousand visitors a month to the page being tested, or results take too long to trust. Lower-traffic B2B sites still improve, using recordings, heatmaps, user interviews and before-and-after comparisons on high-intent pages instead of formal tests.
What should we improve first on our website?
What should we improve first on our website?
The page closest to revenue that the most people abandon. For most B2B sites that is pricing, a demo or contact page, or the homepage above the fold. A small lift on a page that already gets high-intent visitors is worth more than a big lift on a page few people see.
Who should own the website after launch?
Who should own the website after launch?
One named person, usually in marketing, with the time and authority to change things every week. Ownership spread across a team tends to mean nobody looks. A specialist partner can run the loop, but someone inside the company still needs to care about the numbers.
How much does ongoing website optimisation cost?
How much does ongoing website optimisation cost?
It depends on how much is tested and built each month. At Shrink, Optimisation starts from £5,000 a month, and Compound, our CRO and AEO retainer, is £10,000 a month + VAT, starting with a free audit. Both run on a set monthly cycle so the work keeps compounding.