Landing Page Conversion: The Fixes That Move the Number
Button colour is not your problem. Here is the order to fix things in, and how to know if a change actually worked.
CTFM Team
Most conversion advice is a list of tweaks. Change the button colour, add urgency, use a video, reduce the form to three fields.
Those are the last 10% of the problem. The first 90% is whether the page makes sense to someone who arrived from a specific place with a specific expectation. Here is the order that actually matters.
Unbounce's Conversion Benchmark Report, based on more than 41,000 landing pages and 57 million conversions, puts the median landing page conversion rate at 6.6% across industries, ranging from about 3.8% for software up to about 12.3% for events and entertainment.
Do not manage to this number
Three caveats, all of which matter. It comes from pages built on one vendor's platform, which is not a random sample of the web. A "conversion" spans everything from an email signup to a purchase, and those are not comparable events. And the industry spread is large enough that the all-industry median describes almost nobody precisely.
Use it for one thing only: a sense of whether you are in a normal range or wildly outside it. If you convert at 0.4%, something is broken. If you convert at 7%, you are ordinary, and the only benchmark worth chasing after that is your own last month.
Does the page say what the thing that sent them there said? An ad promising a free template that lands on a generic homepage loses most people in two seconds, and no amount of design fixes it. Read your ad, then read your headline. If a stranger would not connect them, stop and fix this first.
Clarity about what this is
Someone landing cold should know what you sell, who it is for and what happens next, without scrolling. Most pages fail this because they open with a benefit statement instead of a description.
The single action
One page, one thing to do. Navigation links, footer menus and secondary offers all give people somewhere else to go, and some of them will take it.
Objections, handled where they arise
Price, contract length, data, whether it works with what they already use, how to cancel. Answer these on the page, next to the point where the doubt occurs, not in a FAQ nobody reaches.
The form
Every field costs you completions. Ask for what you need to take the next step, and get the rest later. This is where most people start, and it belongs fifth.
Speed and mobile
A page that takes five seconds on a phone has already lost people before any of the above applies. Test it on an actual phone on mobile data, not on your desktop.
Copy that describes your product's features rather than the visitor's situation. The test: could a competitor put their name on this page and have it still make sense? If yes, it says nothing.
Too many asks
Newsletter signup, demo request, chat widget, download and a discount popup on one page. Every extra option reduces the chance of the one that matters.
Proof that proves nothing
"Trusted by thousands." Unattributed quotes. Logos of companies that trialled once. Vague social proof reads as filler and can cost more trust than it earns.
Hiding the price
Sometimes justified, usually not. If people cannot work out roughly what this costs, a large share will leave to find out and not come back. At minimum, give the shape of the pricing.
The five-second test, done properly
Show your page to someone outside your company for five seconds, then take it away and ask three questions: what does this company do, who is it for, and what did they want you to do? Do this with five people. It costs nothing, takes twenty minutes, and finds more real problems than a month of A/B tests. If they cannot answer, no tweak to the button will save the page.
This is where most conversion work quietly falls apart. A change is made, the number goes up that week, everyone declares victory, and nobody notices the number was going up anyway.
The honest constraints:
You probably do not have enough traffic for reliable A/B testing. Detecting a small improvement takes far more visitors than most pages get, and calling a test early is the same as guessing.
A test that ends the moment it looks good is not a test. Decide the sample size and the duration before you start, and then wait.
Small samples produce large fake effects. Two hundred visitors and a 30% "lift" is noise wearing a suit.
Seasonality and traffic mix will fool you. If your traffic source changed during the test, you measured the traffic, not the page.
If you have modest traffic, do not test. Make the obviously-right changes from the ordered list above, measure over months rather than days, and spend the saved effort on getting the message right. That is not a compromise; for most pages it produces a bigger improvement than testing would.
Traffic level
What to do
Low
Fix message match and clarity. Watch session recordings. Ask five real people
Medium
Change one significant thing at a time. Compare over full weeks
If the page is clear and the number is still poor, the problem is usually upstream.
Wrong traffic. Cheap clicks from people who were never going to buy. Check conversion by source separately; one bad source can drag your whole average down.
The offer. No page fixes an offer nobody wants. If your best traffic still does not convert, this is the answer nobody wants to hear.
The next step is too big. Asking for a demo when people wanted a price. Offer the smaller step and let them escalate.
Data is drawn from pages built on one platform, and "conversion" spans different actions across industries, so the figures indicate a range rather than a target.
