Competitor website analysis: what their page says that yours does not
Three buyer surveys agree the short list is set before anyone talks to you. So read the three sites you are judged against, then say the thing none of them says.
"A tool cannot know my market, so whatever it tells me will be generic." Fair instinct, wrong conclusion. Competitor website analysis changes your conversion rate not by telling you what to copy, but by telling you which two or three sites your page is being judged against, and what those sites already promise.
Which is where the search results for that exact phrase send you wrong.
What does competitor website analysis actually tell you?
Competitor website analysis tells you the claims, offers and friction sitting on your rivals' public pages. It does not tell you their traffic, their ad spend or their sales. The useful output is the comparison your buyer is already running in another browser tab, written down, so your page can answer it instead of ignoring it.
Search that phrase and you get tool lists: keywords, backlinks, estimated visits. Real questions, different questions. None of them tell you why someone who looked at two other sites this morning picked one of them and not you.
How do you read a competitor's page in ten minutes?
Open the three sites a buyer in your category would shortlist, write each headline promise in one line, note whether price is shown or hidden, count the steps to the first commitment, then put your own page in the same list and read all four cold. The gap you find is the opening.
- Pick the three. Not your favourite rivals. The three a buyer would actually find in search.
- One line each. Their headline promise, in your words, under twelve words.
- The offer. Price shown or hidden, trial or demo, guarantee or nothing.
- The friction. Form fields, steps, and whether a card is needed to start.
- Add yours, last. Read all four in one sitting. The line that reads interchangeable is your problem.
Most founders doing this find all four lines say roughly the same thing. That is the finding, worth more than any backlink count.
Short on time? Revslip reads the comparison for you.
Why does their messaging change your conversion rate?
Because nobody evaluates every option. Buyers build a short list, rank it, and judge each site against the others on it. John Hauser and Birger Wernerfelt named the reason in the Journal of Consumer Research in 1990: evaluation has a cost, so people stop looking early.
Their paper collected consideration set sizes across 23 product categories, from studies with 300 to 600 people each. Shoppers facing between 6 and 47 available brands carried a median of four shampoos. Not forty. Four.
Thirty-six years later the same shape shows up in software buying, only tighter.
What do three independent studies agree on?
That the list is short, and set early. Three surveys run by three organisations with different interests, different countries and different sample sizes all land in the same place: the vendor that won was already on the buyer's list before anyone spoke to them.
- 90% of buyers choose a vendor that was on the short list at the start, from a Bain and Google survey of 1,208 US buyers
- 95% of winning vendors were on the Day One short list, from 6sense's survey of close to 4,000 buyers
- 83% of buyers shortlisted three products or fewer, from TrustRadius, 1,862 buyers, January 2026
The two shortlist figures disagree, and the gap is the interesting part. Bain and Google asked US buyers of software, hardware, telecoms, logistics and industrial equipment. 6sense asked across three regions, where the median purchase sat between $200,000 and $300,000, and its own number moved from 85% to 95% in a single year. Buyers did not change that fast. The definition of a list did.
6sense also found the first vendor contacted won around 80% of the time, which reads like a prize for answering the phone quickly. Then look at the subgroup. Among the 6% of buyers who had not ranked their list in advance, the first vendor they spoke to won only 57%. Speed is not winning anything. Preference is, and the preference was built on the pages.
What happens when every competitor says the same thing?
You get the one move this whole exercise exists for, which is saying something else. Talia Wolf of GetUplift tested it on a large US moving company whose rivals all led on price. The variation led on a protected, premium move instead, and lifted conversions 14%.
Handle that 14% carefully. GetUplift sells conversion work, the client is unnamed, and the write-up gives no baseline rate, no sample size and no significance test, so it is a documented decision rather than proof. The direction is what earns it a place here: the win came from reading every competitor and then refusing to repeat them.
That is the opposite of what the guides ranking for this query recommend. They tell you to find best practice and match it. Match three sites that already sound alike and you have built a fourth.
Want the comparison priced? Run a free audit and the findings arrive ranked.
What moves if the new message is right?
Watch the page carrying the new claim, not the whole site. Bounce on that page, scroll depth and clicks to the next step move first, usually inside a week or two. Sales move last, because a short list takes a buying cycle to refresh.
So give it one full cycle before you judge it. In consultative B2B the chain is visitor to lead, times qualified rate, times close rate, times deal value, and the first link can move weeks before the last one does. On low traffic, judge the first link and wait on the rest.
When is competitor analysis the wrong instrument?
When your question is about their numbers rather than their words. If you want a rival's traffic sources, keyword rankings or estimated visits, buy Semrush, Ahrefs or Similarweb and ignore everything above. Those tools do that job properly. Revslip does not do it at all.
The three numbers of theirs we cannot see Revslip reads rendered public pages, so it can report what a competitor claims, offers and makes difficult. Their conversion rate, their ad spend and their revenue stay invisible to us and to every other page reader, so discount any tool that quotes those as modelled rather than measured, and treat the claims we do read as a hypothesis to check against your own buyers.
Our own position, said plainly: the research that reads competitors' claims, offers and friction is a paid feature here, not part of the free snapshot, which covers roughly 200 conversion signals on your own site. The pattern in this post comes from 134 or more businesses audited, and that sample selected itself. These were people who already suspected their site was the problem, which skews it toward sites with something wrong.
Questions founders ask about reading rivals
How many competitors should I read?
Three. Both buyer surveys put the real short list at three or fewer, so reading ten teaches you about a market rather than about the comparison you keep losing.
Does this still work if my competitors are much bigger?
Usually better. A bigger rival has more to defend and so says less. The sentence they cannot say about themselves is the one available to you.
What if I genuinely have no direct competitors?
Then your buyer's alternative is a spreadsheet, an agency or doing nothing, and those are what your page is compared against. Read the three nearest substitutes the same way.
Related reading: why a headline says what it is instead of what it does, whether your conversion rate is actually bad, what a model can and cannot judge about your copy, and whether to show pricing on a B2B site.