How to read ONS business data without getting it wrong

There are two official counts of UK businesses and they disagree by nearly three million. Neither is wrong. Here is what each one counts, which to quote, and the four mistakes that turn free public data into a bad decision.

Watch: Two answers, both right

Most of what you need to size a UK market is published, free, and checkable. Very little of it is quoted correctly.

This post is the short version of how to use it: what the two official counts actually measure, when to use a count and when to use a rate, and the four mistakes that turn good free data into a confident wrong answer.

The two counts, and why they disagree

There are two official answers to "how many UK businesses are there", and they differ by nearly three million.

• About 2.86 million, from the ONS. This counts businesses registered for VAT or PAYE.
• About 5.7 million, from the Department for Business and Trade. This counts all private sector businesses, including the very small ones under both thresholds.

Neither is wrong. They answer different questions. The gap is mostly one-person businesses with no staff and modest turnover: about 4.3 million of that 5.7 million employ nobody but the owner.

Which one to use depends on what you are asking:

• Sizing a market you can sell to? The ONS number is usually the honest one. A business too small to register for VAT is often too small to buy.
• Talking about the economy or self-employment? The wider number is the right one.

Quoting 5.7 million and then applying ONS growth rates to it is the single most common error in this area. The rates belong to the smaller population.

Count or rate: pick deliberately

A count tells you size. A rate tells you intensity. Confusing them makes big places look successful and small places look asleep.

London creates more businesses than Northern Ireland every year, in raw numbers, because London is far bigger. That fact is nearly useless. The rate answers the better question: of the businesses already there, how many are new? On that measure London still leads at 12.7%, but the ranking underneath it changes completely.

The rule: use a rate whenever you are comparing places or trades of different sizes. Use a count only when the absolute number is the point, such as how many firms you could actually contact.

The lag, and what to do about it

The annual business demography bulletin describing 2024 was published in November 2025. It is roughly eleven months behind the year it covers.

That is not a flaw, it is what it takes to produce accurate annual statistics. But it means two things:

• Always state which year a figure describes, not the year you read it.
• Do not use the annual data to answer "what is happening right now". Use the quarterly series for that, and read at least four quarters of it.

The quarterly series is fresher and much noisier. Business creations fell 8.0% in the first quarter of 2026 against the same quarter a year before, then rose 2.2% in the second. Nothing about the economy changed that fast. What changed was the quarter being compared against.

Four mistakes worth avoiding

Reading openings without closings. The one that costs the most. Transport and storage had the highest opening rate of any main trade in 2024 at 15.6%, which makes it look like the best market in the country. Its closing rate was 16.5%. More firms left than joined. The full worked example is in the fastest growing UK industries.

Treating one quarter as a trend. See above. Four quarters minimum, and check the annual figures underneath them.

Using a national average for a local decision. The UK opened 11.1% and closed 9.8% of its businesses in 2024. If you sell inside one region, that describes a country you are not selling in. Regional survival ranges from 43.5% in the South West down to 30.6% in the West Midlands, which is a much bigger spread than most people assume. More in where new UK businesses are being created.

Assuming "closed" means "failed". The ONS counts a business as gone when it stops being registered for VAT or PAYE. Some of those failed. Others were sold, merged, wound up deliberately, or simply shrank below the registration threshold. The data cannot tell them apart, so neither should you.

Where to look

Three sources cover almost everything:

• Business demography, UK for annual openings, closings and survival, split by trade and by area.
• Business demography, quarterly for the freshest openings and closings, with the caveats above.
• Business population estimates for the wider business population by size band.

Each publishes the underlying tables alongside the summary. Read the tables. The summary is written for a headline, and the interesting splits are almost always one level down.

Contains public sector information licensed under the Open Government Licence v3.0.

When the research is done, you still need names

Statistics describe a market. They cannot tell you who is in it.

That part is public too. Every UK company appears on the public record with a name, an address, a trade code and a start date. Leadistry reads that record, then finds a working website and a verified business email for each firm, so your research turns into a list you can act on.

Counting is free. Start free with 25 leads and see the real size of the slice you have been reading about.