Signs you need conversion help, not a new website

If your site converts poorly but people find it, read it and get most of the way through checkout, you have a conversion problem and a rebuild will not fix it. Rebuilds solve structural and technical failures. They do not solve funnel leaks, and they routinely reintroduce leaks the old site had already solved.

The five signs it is conversion, not construction

People reach the cart and stop. Cart abandonment averages 70.22% across Baymard’s meta-analysis of fifty studies. If your drop-off is concentrated between cart and payment, that is a checkout problem with documented causes, not a design problem.

Mobile is dramatically worse than desktop. Across 486 million sessions and 117 brands, DRIP found desktop converting 1.56 times higher than mobile while mobile carried 78% of traffic. A larger gap than that is a mobile experience problem, and a rebuild will only fix it if mobile is the explicit brief, which it usually is not.

Traffic is fine and revenue is not. If sessions are stable or rising and revenue per session is falling, nothing about the site’s structure changed. Something about the offer, the mix or the funnel did.

One template underperforms the rest. A single product template or collection page converting well below its peers is a page-level problem. Rebuilding forty pages to fix one is expensive arithmetic.

People read and leave. Session recordings showing careful reading followed by exit is a messaging or offer problem. A new visual system will not change the answer to a question the page never addressed.

The signs it genuinely is construction

Be fair to the rebuild case, because it exists.

Platform limits that block every test. Technical debt severe enough that a simple change takes three weeks. Performance so poor that incremental fixes cannot close the gap, which is not rare given only around 42% of mobile sites pass all three Core Web Vitals. Or a brand repositioning, where the visual system has to change for reasons that have nothing to do with conversion and should be budgeted as a brand cost rather than expected to pay for itself.

Why rebuilds so often lose

They change hundreds of variables at once, so a revenue decline cannot be diagnosed and a revenue rise cannot be attributed. With average win rates near 12% across 127,000 experiments, most individual changes teams believe are improvements do nothing or hurt. A rebuild bundles hundreds of those bets and ships them together.

They also produce no learning. A test that loses tells you something true about your customers that stays useful for years. A rebuild that loses tells you only that the rebuild lost.

If you are rebuilding anyway

Three things make it survivable. Test the big structural bets on the current site first, so evidence carries into the new design. Launch to a traffic split rather than to everyone, so you can reverse a mistake before it costs a quarter. And plan the first three tests on the new site before launch, because the six-month post-launch freeze is where most rebuild value is lost.

What a rebuild costs that nobody prices

The freeze. Most sites stop changing for three to six months before launch and another two or three afterwards while bugs clear and the team recovers. That is most of a year with no testing, no learning and no compounding, and it appears on no quote.

Price that gap against the rebuild’s expected return before deciding. For a site that already converts adequately, the freeze frequently costs more than the redesign gains.

The question to ask before signing off a rebuild

What specifically is wrong, and how do we know? If the answer is a list of evidenced problems, ask whether each could be fixed and tested individually. Usually most can, for a fraction of the cost and with a result you can actually read.

If the answer is that it feels dated, make that decision as a brand decision and budget it as one.

You can see how the diagnosis works before committing to either path.