Why Your Website Doesn't Convert
Most websites fail to convert for one of three reasons: the visitor cannot tell within five seconds what is being sold and to whom, the page asks for a commitment larger than the trust earned so far, or the traffic was never qualified to begin with. A redesign fixes none of these — diagnosis has to come first.
The site looks good. The analytics show visitors. The contact form stays quiet. Before you redesign anything, work through these in order — the cheap diagnoses are at the top for a reason.
1. The traffic was never qualified
The most common conversion problem is not a conversion problem. If your visitors arrived from a channel that attracts people at the wrong stage, or the wrong people entirely, no amount of button colour will help.
Check the split between branded and non-branded search, and check what pages your visitors land on. A site getting 80% of its traffic to a blog post about a general topic is not underperforming — it is doing exactly what you built it to do, which was attract readers rather than buyers.
2. The five-second test fails
Show your home page to someone outside your industry for five seconds, then take it away and ask: what do they sell, and who is it for?
Most sites fail this. The headline says something aspirational and abstract, the subhead explains the aspiration, and the actual offer appears below the fold. If a stranger cannot answer both questions, nothing further on the page gets a chance.
The fix is unglamorous: say what you do, for whom, in the first line. Distinctiveness comes from what you say, not from delaying it.
3. The ask is bigger than the trust
Every page makes a request. "Book a 45-minute consultation" from a first-time visitor who has read one paragraph is a large request. "See pricing" is a small one.
Map the request against the evidence you have provided so far. If a visitor has seen no work, no names, no numbers and no faces, the only ask they will accept is one that costs them nothing and commits them to nothing.
The correction is either to lower the ask or to raise the trust before making it — proof, specifics, a case with a number in it, a real person's name attached to a real outcome.
4. There is no single obvious next step
Pages that offer six equally weighted options convert worse than pages that offer one. This is not a design preference, it is how choice works under low motivation.
Count the calls to action on your key page. If there are more than two distinct ones, and they are styled with equal emphasis, the page has no opinion about what should happen next — so the visitor forms none either.
5. The page answers the wrong objection
Every prospect arrives with a specific fear. In services it is usually "will this be worth it and will you disappear halfway". In ecommerce it is usually "will it fit / arrive / be returnable".
Sites tend to answer the objection the founder finds most interesting rather than the one the customer actually has. The remedy is not clever: ask ten recent customers what nearly stopped them, and put the answers on the page.
6. Mobile is an afterthought
Look at your own mobile numbers before arguing with this. Most sites now take the majority of their traffic on phones and convert it at a fraction of the desktop rate. Some of that gap is intent. A lot of it is a form that requires a pinch-zoom, a sticky header eating a third of the screen, and a primary button below three screens of scrolling.
7. Speed, but only past the threshold
If your page takes more than about three seconds to become useful on a mid-range phone on mobile data, fix that before anything else on this list. If it takes 1.5 seconds, stop optimising it and go back to item 2 — you are polishing a variable that is no longer the constraint.
8. The offer is genuinely uncompetitive
The uncomfortable one. Sometimes the page is clear, the trust is earned, the speed is fine, and people still leave — because they compared you and chose otherwise.
No CRO fixes this. It is a positioning problem, and it is solved upstream of the website. The website is only ever an argument; if the argument is weak, better typography makes it a well-set weak argument.
How to actually run this
Do not test all eight. Pick the one your evidence points at, change it alone, and wait for enough data to know. If you do not have the traffic for statistical confidence — and most businesses do not — replace testing with watching: ten session recordings and five conversations with real customers will tell you more in a week than a split test you cannot power.
And before commissioning a rebuild, write down which of these eight it is meant to fix. If you cannot name one, the rebuild will produce a better-looking version of the same number.
Questions people actually ask
- What is a good website conversion rate?
- For B2B service businesses, 1–3% of visitors making contact is typical and 5% is strong. For ecommerce, 1–4% is the usual band. But the benchmark that matters is your own last quarter — averages across industries hide traffic quality differences large enough to make the comparison meaningless.
- Should I redesign my website to improve conversions?
- Usually not first. A redesign changes hundreds of variables at once, so if the number moves you will not know why, and if it drops you cannot revert selectively. Diagnose, fix the specific failure, and redesign when the reason is strategic rather than statistical.
- How much traffic do I need before conversion optimisation is worth doing?
- For A/B testing to be meaningful you generally need a few hundred conversions per variant, which most small sites never reach. Below that, do qualitative work instead: session recordings, five user interviews, and reading your own page as a stranger.
- Does page speed affect conversion?
- Yes, and disproportionately on mobile. But it is a threshold effect, not a gradient — going from four seconds to two matters enormously, and from two to 1.8 usually does not.