Email deliverability: limits of ONS regional data

What to take away

What ONS data can and cannot say about email engagement

The Office for National Statistics publishes some of the most useful context available to a UK email team. It counts households, people and businesses across the United Kingdom regions. It describes how they connect and what they do online.

What it does not do is measure your campaign. There is no ONS figure for your open rate in Leeds or your click rate in Cardiff. Nobody outside your platform holds that number.

That gap matters because email deliverability advice often leans on regional averages that were never measured as email engagement in the first place.

The ONS home internet and social media usage statistics cover access, connection type and a set of online activities by nation and region of the UK. They are the closest thing to a regional denominator you can cite.

ONS digital economy data sits alongside that, describing internet and email use by UK businesses rather than households. Read the two together and you have a rough picture of who is reachable and who is at work.

Neither release knows whether your message landed in the inbox or the spam folder. Deliverability is a mailbox provider decision made per recipient, per sending domain, per reputation history.

So treat ONS figures as the population side of a fraction. Your platform supplies the other side. A rate without both is a number with no meaning.

This is the same problem that appears when a benchmark is not comparable to your list, your sector or your sending setup, however respectable the source looks.

Using ONS regional internet and household data

The ONS publishes internet access and usage at a level of detail that most commercial benchmark reports cannot match. Its breakdowns for the United Kingdom regions are consistent, documented and free.

Household internet access in the UK has been near universal for years, so the interesting variation is no longer whether a household is online. It is how that household connects, how many adults live there and what they do once connected.

Age structure drives much of the difference. Regions with a higher share of older residents show different patterns of online activity from regions with a younger, more mobile population.

Business use matters too. Regions with a large professional services base send and receive far more business email per head than regions dominated by sectors where email is a small part of the working day.

ONS people, population and community data gives you the demographic layer underneath all of this: age, household composition, employment and migration by region.

Put the two together and you can describe an audience before you mail it. That description is what makes a regional claim defensible.

One caution. ONS regional figures are estimates from a sample, published with confidence intervals. A one or two point difference between two regions is often not a real difference at all.

If you are briefing colleagues on who receives your mail, it also helps to split buyers, operators and recipients before you start quoting regional statistics at them.

Comparing London, the North West, Scotland and Wales

London, the North West, Scotland and Wales provide useful examples of regional audience context. The points below describe population, housing and employment, not email engagement.

London has the youngest population of the four and the highest share of private renting. Addresses change often, which means list decay runs faster and hard bounces accumulate quietly.

London also concentrates professional services, finance and technology employment. Business email volume per working adult is high, and so is competition for attention in the inbox.

The North West combines a large urban core around Manchester with wide commuter and rural districts. Household composition varies sharply between them, so a single North West segment hides a lot.

Scotland has a distinctive settlement pattern, with a large share of the population in a small number of cities and a long tail of sparsely populated areas. Broadband availability and speed vary accordingly.

Wales has an older population profile than London and a higher share of residents in smaller settlements. Public sector and manufacturing employment carry more weight than finance.

Read the table as context, not as a forecast. It tells you why two United Kingdom regions with identical campaign settings might still produce different results.

It also tells you why a national average applied to all four is a poor planning tool. The average describes none of them well.

Reading open and click rates against regional denominators

An open rate is a fraction. The numerator is opens recorded by your platform. The denominator is delivered messages, or accepted messages, depending on how your platform counts.

Change the denominator and the rate moves without anything changing in the inbox. That is the single most common error in regional reporting.

Apple's Mail Privacy Protection inflates opens for recipients who use Apple Mail, and its effect is unevenly distributed across audiences. A region with a higher share of Apple Mail users will look better on opens for reasons unrelated to your subject lines.

Clicks are harder to fake and generally more stable, but they inherit the same denominator problem. A click rate per delivered message and a click rate per recipient are different numbers.

Worked example. A retailer mails 40,000 recipients in the North West and 12,000 in Wales with the same campaign. The North West records a 21 percent open rate and Wales 26 percent.

Before concluding that Wales engages better, check three things. Which mail clients dominate each list. How many addresses in each were recently re-engaged. Whether the Welsh list is smaller and therefore skewed by a handful of heavy openers.

A small list produces a volatile rate. Twelve thousand recipients is workable, but a segment of eight hundred is not a basis for a regional claim.

Build a dashboard that shows its denominators so the delivered, accepted and recipient counts sit next to every rate you publish.

Segmentation by region without inventing benchmarks

Regional segmentation is a targeting decision, not a benchmarking exercise. You are deciding who receives which message, not proving that one nation opens more than another.

Start from what the ONS can support and build outward. The steps below keep the ONS layer and your platform layer separate.

Where regional differences survive those five steps, they are worth acting on. Where they do not, you have saved yourself a segmentation project.

Keep the ONS layer visible in your reporting. A regional segment justified by household composition and employment mix is far easier to defend than one justified by a single campaign result.

If your measurement chain runs from opens through to revenue, separate the links carefully. Moving from mixed denominators to attributed revenue without stating each step is how regional claims become overstated.

ONS release calendar and how to time a regional data story

The ONS publishes to a fixed schedule, and the schedule is public. That gives you a legitimate reason to publish a regional email story at a particular moment.

Check the ONS release calendar before you plan a data-led campaign or a client report. Releases are announced in advance, and the regional breakdowns you want usually arrive with a named release rather than on their own.

Timing works in your favour in two ways. A fresh release gives your story a news hook. It also means the figures you quote are the current ones, which matters when a reader checks your source.

Plan backwards from the release date. Draft the analysis in advance, leave the numbers as placeholders, and fill them on the day. Do not publish a regional claim before the release it depends on.

Be explicit about reference periods. An ONS release published in one month often describes data collected months earlier. Say so, or a reader will assume the figures are current to the day.

Where a release is revised, note the revision. Regional estimates do get revised, and a claim built on a superseded figure is easy to discredit.

Statistics produced by the ONS operate under the governance of the UK Statistics Authority, which is why citing the release and its reference period carries weight with informed readers.

The same discipline applies to your own numbers. Separating platform attribution from causal impact is what stops a well-timed regional story from turning into an unsupported one.

Limits of ONS statistics for email engagement claims

The limits matter because ONS data cannot provide a table of regional open rates.

First, the ONS does not measure email engagement. It measures internet access, online activity and business internet use. Opens and clicks are not in scope.

Second, regional estimates carry sampling uncertainty. Differences between United Kingdom regions can sit inside the margin of error, particularly for smaller nations and regions.

Third, the ONS reports by nation and region of the UK, which does not always match how a marketing list is organised. Postcode, region field and ONS geography can disagree.

Fourth, the releases are periodic, not live. Between publications, the regional picture you are describing may have moved.

Fifth, ONS figures describe populations and businesses, not subscribers. Your list is a self-selected subset with its own biases, and it is not a sample of the region.

None of that makes the ONS useless. It makes it a context source. Use it to describe who is online, how they connect and what kind of work they do.

Use your own platform data to describe what happened when you mailed them. Keep the two in separate sentences and separate columns.

The Home internet and social media usage release is the right starting point for household context, and the Digital economy collection covers internet and email use on the business side.

For demographic layers, People, population and community holds the regional breakdowns you will need to describe an audience before you mail it.

And before you publish anything with a regional claim in it, check the Release calendar so the figures you quote are the current ones.

Common questions