Comparison platforms · 3 min read

Ten Was the Default. Five Was Right.

The structure of a comparison platform travels well. Judgement about who belongs on the list does not. That has to come from the place.

By Jeev Trika

Most of what makes a comparison platform work is portable. Categories, page structure, how a buyer moves from a list to a provider: build it once and it works in the next market. The part that does not travel is knowing which providers in that market deserve to be on the list at all.

I have run that experiment twice, fourteen years apart.

2010: one template, many countries

By late 2010 topseos had country sites: the main .com, and versions for Canada, Australia, the UK, India and the UAE. On 18 November I sent our developer a 24-point spec headed “Search Strategy Points.” Almost all of it was about making one template behave locally.

The homepage of each country site would say what it was in the language of that market’s searches:

Rankings of Best SEO Companies & PPC Companies in [Country]

Addresses on the country sites would use the short words people actually typed, “seo” and “ppc” rather than “search engine optimization” and “pay per click management.” Internal links would use the same keywords as the .com, “but all internal linking back to country specific site.” The “Worldwide” option in the menu would appear only on the five major sites; everywhere else it would show the country you were in. Even the apply page had a local title:

Country - Best SEO Company or PPC Company based out of [Country]? Apply.

The machinery was global. Every page was rewritten to look and read local, because a buyer in Dubai searching for an SEO firm wants firms that work in Dubai, and a page that feels like a corner of an American site does not answer that.

What the spec did not do was decide who was good in Dubai. That was a ranking question, and the template could not answer it.

2024: one city, and a rule I brought with me

Fourteen years later I took the same playbook to a single city. Top of Naples ranks local businesses in Naples, Florida: restaurants, salons, dentists, trades. To launch it, we worked with a local magazine publisher who knew the town far better than we did. He had sent us his picks, mostly five per category. As on our national platforms, positions on the list were sold; which businesses were on it in the first place was where his local knowledge came in.

When the site was set up, many of those categories showed ten. In December 2024, a couple of weeks before launch, his side asked why. My answer was the national rule:

Most rankings lists are top 10, not top 5, top 15, top 20, most are top 10. It drives more CTAs (click throughs).

On a national platform that is right. Lists of ten get clicked. A buyer comparing SEO agencies across the US has hundreds of credible choices, and ten is a reasonable shortlist.

The publisher wrote back within minutes:

When you expand the list to 10., and the choices are not talented, it devalues the rest of the list… please see hair salons

He was right, and the reason was local. In one city there may only be five salons a resident would recommend to a friend. Padding the list to ten does not add choice. It puts names next to the good ones that a local reader knows should not be there, and the reader stops trusting the five. He also flagged a restaurant he had never heard of on the top restaurants list.

Eleven minutes after my first reply, I sent my second:

We will take it back to top 5 for the categories you identified and gave us data for, today.

What travels and what does not

The two episodes split cleanly.

What travelled: the page structure, the categories, the habit of writing every page for the words a local buyer searches, and the business model.

What did not: how long the list should be, and who belongs on it. On a national platform, the number is set by how buyers browse. In one city, it is set by how many providers are genuinely good, and only someone who lives there knows that number.

A national default is a starting point for a new market, not a rule. In Naples it took one local publisher, one example and eleven minutes to show me the difference.