Our methodology and update frequency

1. Collection

We collect information from public sources, described in full on the Data sources page: each nation's regulator, publicly available information, and the NHS, plus the practice itself once it has claimed its listing.

2. Matching

We match records from different sources to the same practice using identifiers such as registration details, address and name. Where we are not confident two records describe the same practice, we keep them separate rather than risk merging the wrong information.

3. How we prioritise sources

We use the most authoritative source for each detail, and a more authoritative source wins when two disagree. In general:

If we cannot decide confidently, we prefer to show less rather than risk showing something misleading, and we may flag the detail for review.

4. What we publish, and what we suppress

• We publish only fields we can attribute to a source. Unsourced details are not shown.
• We never generate ratings, claims or availability ourselves.
• We suppress or down-rank details that look stale, internally inconsistent, or that a practice has asked us to correct while we check.
• Marketing-style claims, for example service descriptions, are shown as practice-supplied, not as verified facts.

5. How we rank practices

Every ranked list on this site — a town page, a treatment page, a top-rated leaderboard — is ordered by the same weighted score, and by nothing else. We do not sell position, and a practice paying to manage its listing does not move up.

Ranking on the raw star average alone would put a practice with one five-star review above one with six hundred reviews averaging 4.8. So each practice's own average is pulled towards the average across the whole directory, and how far it is pulled depends on how many reviews it has:

In plain terms: a practice needs enough reviews to be trusted with its own average. At 100 reviews it counts half its own rating and half the directory mean; by 1,000 reviews it is almost entirely its own. C is recomputed on every build rather than fixed, so the benchmark moves with the directory.

Why the constant is 100

It is the median review count across the rated practices we list. A practice is therefore trusted with its own average once it has more reviews than half the directory. We review that number once a year rather than at every data refresh, because tying it to the live median would move a practice's position every quarter even when nothing about that practice had changed.

It was 20 until September 2026. At that setting a practice with thirteen reviews could outrank one with several hundred, and some of our own lists showed exactly that, so we raised it.

Where this formula comes from

It is not ours. It is a Bayesian average, the standard way to rank things that have wildly different numbers of ratings, and the exact form above is the weighted rating IMDb publishes for its Top 250. Trustpilot's TrustScore uses the same shape with a threshold of nine reviews and a starting point of 3.5 stars. Ours differs from theirs in the two numbers, which are set for a directory of dental practices rather than films or retailers, and both are stated here.

We publish the formula and its constants rather than describing the ranking as proprietary. If you believe your practice is ranked wrongly, you can work out its score from the figures on its own page.

6. Our sample, and what it does not cover

The directory holds 12,341 dental practices across 819 towns and cities in all four UK nations. 11,303 carry patient ratings, between them 2,080,111 reviews. We publish four honest limits on that:

• Reviews come from one source. Patient ratings are Google reviews. They are not a clinical measure, they are not audited, and no regulator rates dental practices — a point we make wherever ratings appear.
• Coverage of detail is uneven. We hold usable opening hours for 11,096 practices and a service list for 8,768. A practice missing from a "open on Saturdays" or treatment page may simply be one whose hours or services we do not hold, not one that fails to offer it.
• Thresholds are applied, and stated. A town page needs at least three practices; a treatment page in a town needs eight; a top-rated leaderboard needs ten rated practices. Below those, we would be presenting a list too short to compare, so we do not publish one.
• A directory is a snapshot. Practices open, close, merge and rebrand. Every page carries a last-checked date, and we would rather show a dated fact than an undated guess.

7. Other sources in this sector

We are one source among several, and for some questions we are not the right one. If you are checking a specific claim, go to the body that publishes it:

If you are writing about UK dentistry and want to check anything on this page, or want the underlying figures for a specific town, get in touch.

8. Update frequency

Different information refreshes at different times, summarised below. The authoritative signal is always the last checked date on each listing. Practices with a managed listing update their own details themselves, anytime, rather than wait for our schedule.

9. Corrections and objections

There are two routes, and the difference matters:

• Correcting mistakes in existing content is free. Adding new content needs a plan. Whatever the field, on any listing, whether or not the practice pays. We build listings from public records, so when we get something wrong that is our mistake to put right. Report it through our correction form. There is no "suggest an edit" control on listings themselves, so a correction always comes with a practice attached and a reply address we can confirm to.
• New content is added by the practice, on a plan. That means anything we do not already hold: prices next to your treatments, photos, video, social media links, an online booking link, or more in the sections we only show partially, such as more treatments than we found or more of your team than the two we show. Those are additions, not corrections, because nothing on the page is wrong. A practice adds them with a managed listing, and then keeps them current itself. Reviews, ratings, regulatory status and the regulator's own inspection record stay independent and are not editable on any plan.

Requests to correct or remove personal data are handled free in every case, under the Privacy notice, regardless of plan. Full detail: Corrections and Plans.

10. Internal records

We keep internal records of how each source is used, including collection cadence and how we respond to corrections and objections, so we can stand behind what we publish. We don't publish the operational detail of our collection systems, but our published descriptions are intended to be accurate and not misleading.