Why we run one site per trade, not one per neighbourhood

The question this answers

If you have looked at local service directories for long, you have seen the other model: a network of near-identical sites, one per neighbourhood, each listing more or less the same companies under a different headline.

We do not build that. Our independent Singapore service directories run one site per trade, and since the area-network model is the standard way to do this, it is worth saying why. The short version is that Singapore does not have the geography that model assumes.

Singapore is one market

The per-neighbourhood model makes sense in a sprawling city where travel time is the constraint. In a large metropolitan area, a plumber in one suburb genuinely will not take a job an hour away, so a directory per suburb lists a genuinely different set of businesses.

Singapore does not work that way. The island is small, the trades are mobile, and the same companies serve Tampines and Jurong in the same week. A locksmith who declines a call because it is across the island is unusual, not typical.

So a site per planning area would list the same businesses over and over. Twenty sites, forty companies, repeated twenty times. Each page would look local while telling a reader nothing they could not learn from one page, and the reader would have to work out that the “Bedok” and “Clementi” lists are the same list.

What we do instead

One site per trade, covering the whole island, with area coverage handled inside it.

Each listing carries area tags: the planning areas a business says it covers. Inside a directory such as the Singapore Plumber Guide, those tags drive planning-area pages that filter the island-wide set, so if you want plumbers who cover Woodlands you can see them, drawn from one maintained set of businesses with one set of scores.

The advantage is not aesthetic. There is one place where a business’s details live, one place where its score is calculated, and one place where a correction has to be applied. When a company closes, it disappears once, at the monthly re-check, rather than persisting on eleven sites nobody remembered to update. Accuracy is a maintenance problem, and this structure has far less of it.

It also keeps the trades separate, which is the division that actually matters. A locksmith and a plumber are judged on different things, as set out in how we rank businesses, and separating by trade means each directory can weight what matters for its own trade rather than averaging across all of them.

How to tell this apart from a doorway network

A reader assessing whether a site is a real publication has a few reliable tests, and we would rather hand them over than ask to be trusted.

Does each page exist for a reader or for a search engine? A page that says nothing a neighbouring page does not say was built for a crawler.

Is there one maintained data set, or many stale copies? Ask what happens when a business closes. If the answer involves updating dozens of sites by hand, it will not happen.

Is the method published? A doorway network has no method to publish, because its ordering exists to sell placement.

Is there a named person accountable? Ours is on the about page, with the role described plainly.

Can a listed business get out? A network built to monetise listings makes removal difficult. Ours is one email, honoured, free, and described at corrections and removals.

Why trade, and not something else

If not geography, why split by trade at all? Why not one directory of everything, or one per job type?

One directory of everything fails because the scoring would have to be generic. The whole argument of the method is that different trades fail in different ways, and that a locksmith should be judged on response and call-out honesty while an aircon company is judged across years of repeat visits. Averaged into a single home-services score, both weightings disappear and you are left with a star rating that says nothing useful.

Splitting more finely than trade fails for the opposite reason. A directory of chemical wash and a separate one of aircon repair would list mostly the same companies, which is the neighbourhood problem again in a different costume. Trade is the level where the businesses genuinely differ and where the criteria genuinely differ, which makes it the right seam.

The trade-off we accept

This model has a cost, and we would rather name it than pretend otherwise: a page for every trade-and-area combination would probably capture more search traffic than one directory per trade does. That is exactly why the other model is so common.

We think a reader is better served by one directory per trade that is accurate, maintained, and honest about its method, than by twenty thin pages that reshuffle the same names. If a directory is going to argue that its rankings are checkable, it cannot be built on a structure whose main purpose is to occupy search results.