Why Most Comparison Sites Fail

I have built comparison platforms since 2002, about 250 of them at the peak. Some became real businesses. Some never got going. Some did well for years and are now semi-dormant; in 2025 I told a former partner as much, in so many words: “Yes I have other platforms as well that are semi-dormant.”

The failures had very little to do with the listings. Listings are the easy part. These are the places it actually goes wrong, each with the article on this site where I have told that part in detail.

1. It is a directory, not a decision

A list of companies is inventory. A buyer does not need inventory; they need help choosing. The platforms that worked did something with the list: ranked it against stated criteria, matched buyers with providers, explained the differences. In 2010 topseos had a leads service that matched buyers with firms, and every ranked firm got tools to use its ranking in its own selling.

The provider side has the same problem. In 2009, when an agency asked what our traffic was worth, my answer was that the listing should not be valued on traffic alone, but on what the firm did with the ranking (Before Anyone Is Looking). A directory gives a provider a line of text. A platform gives it something to use.

2. Nobody comes first

Buyers will not come until the providers are there. Providers will not pay until the buyers are there. Most new platforms die waiting for the other side. topseos got through it by ranking firms before they had signed up, giving them seals and press release templates so they would tell their own audiences, and finding buyers one at a time. See Before Anyone Is Looking.

3. The categories are wrong

If the platform is organised by the industry’s jargon instead of the words buyers search, nobody finds it and nobody understands it. Not every category deserves a platform either. In 2014 I let nine domains go that had been registered for categories and spin-offs that never became platforms. See Name the Category the Way the Buyer Searches.

4. It is national in a local market

A country page that is a copy of the national page with the flag changed does not help a local buyer. In 2018 an agency in Singapore, buying placement in several countries’ lists, put the problem in one sentence: “It will be much tougher for us to convert any leads compared to the local vendors.” Local buyers wanted local providers, and the local list had to be built for them. See Ten Was the Default. Five Was Right. and A Country Page Is Not a Local Site.

5. The money buys the wrong thing

Providers pay; that is the model almost everywhere. The trouble starts when the buyer cannot tell what was paid for, or when the order of the list is for sale rather than the chance to be considered. My own platforms sold positions at monthly prices. See Charging the Companies You Compare.

6. It depends on one road in

Most comparison platforms get their buyers from search. That makes the search engine the landlord. In 2014 Google removed every one of our websites from its results; we sued and lost. A decade later the results page itself changed, and answers started appearing above the links. A platform that has only one source of buyers is one policy change from a bad year. See Rented Land and When I Took Google to Court.

7. It was built for a results page that no longer exists

A platform built to rank as one of ten links has its assumptions in its code: how pages are generated, what is on them, how they link. When search moved to answers, some of those assumptions could not be fixed with a content edit. In 2025, asked to change one key page for AI results, our answer was that it “would change all the programming/coding for the whole site.” See Built for Google, Rebuilt for Answers.

8. Nobody keeps working on it

The last one is the plainest. A comparison platform is not a website you finish. The rankings need re-evaluating, the categories need re-cutting, the buyers need finding again every year. When the work stops, the platform does not die immediately. It goes quiet. Of one platform, in 2025, I put it simply: “At this time, not a lot is going into organic optimizing / monetizing from the platform.”

What they have in common

None of these failures is about the number of listings. Every one is about whether a buyer can make a better decision with the platform than without it, and whether someone is still doing the work to keep that true.

I built about 250 of these and watched which ones held: the portfolio is here, and building one properly is work I still take on.