The three layers of direct mail list selection
Think of a direct mail list brief as three concentric filters applied to a national consumer file. Each filter narrows the universe; together they define the households most likely to respond to your specific offer. Miss one layer and you will either over-spend reaching unsuitable addresses or under-reach by requesting a universe that is too small to test reliably.
Layer 1: geography
Royal Mail postcodes are structured in four levels of increasing precision. Areas (like BS for Bristol or M for Manchester) cover large city or regional zones. Districts (BS1, BS2, M1, M60) map to town or neighbourhood level. Sectors (BS1 1, BS1 2) divide districts into sub-areas with roughly 2,500 to 3,000 addresses each. Units (BS1 1AA) identify a single street or block of flats, typically 15 to 100 addresses.
Most campaign briefs operate at district or sector level. Sector is the most practical choice for campaigns targeting a defined local catchment, a solar installer covering the BS postcode area, say, would select sectors adjacent to their depot rather than the entire area. Unit-level selection is technically available but produces counts in the dozens, which is useful only for hyper-local canvassing rather than a print campaign. See our article on UK consumer data by postcode for a fuller explanation of how postcode geography maps to available counts.
You can also define geography by radius from a central postcode, a common approach for retailers, clinics, or franchised service businesses. A radius select is converted into a list of postcodes within that distance and then treated like any other geographic filter.
Layer 2: demographics
Demographic selects describe the household or individual, rather than where they live. The standard fields available on a UK consumer postal file include:
- Age band (typically in 5 or 10-year ranges: 25-34, 35-44, 55-64, 65+)
- Household composition (couple with children, single adult, couple no children, retired household)
- Property tenure (owner-occupier, private rental, social housing)
- Modelled income band (derived from property value, financial product holdings, and lifestyle indicators)
- Length of residence (recent movers being a distinct segment for products like home insurance or utility switching)
Property tenure is particularly useful. A financial services firm offering equity release should limit its selection to owner-occupiers aged 55 or over; including renters wastes print and postage on households that cannot convert. Homeowner data in the UK covers approximately 16 million addresses, which gives a deep enough universe for national campaigns even after layering in additional demographic criteria.
Layer 3: behaviour and lifestyle
Behavioural and lifestyle selects are drawn from declared interests, purchase history, and survey responses captured as part of the consent process on the consumer file. Common flags include mail-order purchaser (key for catalogue and DTC brands), charitable donor (important for charity fundraising and cause-related appeals), health and wellness interest, travel, gardening, home improvement, and financial product ownership.
These flags carry more predictive weight than demographics alone. Two 55-year-old homeowners in the same postcode sector can look identical on demographic selects but behave very differently: one is an active mail-order buyer, the other has never responded to a cold postal piece in ten years. Behavioural data separates them. For campaigns targeting high net worth households, combining a modelled income select with declared lifestyle flags (luxury travel, fine wine, premium car ownership) is consistently more accurate than income modelling alone. Our UK affluent consumer data article covers those selects in more depth.
How to write a direct mail data brief
A data brief is the specification document you send to your list supplier. The more precisely it describes your target, the more useful the count and the sharper the list. A weak brief returns an inflated count that falls apart after suppression and deduplication.
A complete brief covers six things:
- Target geography. List postcodes, sectors, districts, regions, or describe a radius from a named postcode. Avoid "the South of England" without specifying which areas; suppliers map that differently.
- Demographic criteria. Age range, household type, tenure, income band. Be specific: "owner-occupier, aged 45 to 70, household income modelled above £40,000" is workable; "middle-aged homeowners" is not.
- Behavioural or interest flags. Specify the declared interest categories relevant to your product. If you have no firm view, ask the supplier to model response propensity from past campaigns on the file.
- Required output fields. Minimum: title, first name, surname, address line 1, address line 2, town, county, postcode. Additional: telephone if you plan a follow-up call, or email for multi-channel sequencing.
- Intended mail date. This lets the supplier confirm that MPS suppression will be applied using the file closest to your dispatch date, which matters because MPS updates monthly.
- Volume required. Specify a target quantity and confirm you understand the count may be lower once MPS suppression, deduplication against your existing customer file, and any excludes are applied.
In our experience, the single most common gap in a data brief is the absence of any behavioural filter. A campaign selling walk-in bath conversions to elderly homeowners will produce a far higher response rate if it is limited to declared home-improvement intenders, rather than every owner-occupier over 70 in the target postcode. Adding even one lifestyle flag can halve the print run while doubling the conversion rate.
What sample sizes do you need for a reliable test?
The standard advice in UK direct mail is to test before rolling out. That principle is sound, but it is only useful if the test volume gives you a statistically valid response rate reading.
At a 1% response rate (typical for cold consumer direct mail), a test of 5,000 records will yield roughly 50 responses. The confidence interval on that is wide: your true response rate could plausibly be anywhere from 0.7% to 1.4%. You cannot confidently scale on that. A test of 10,000 records doubles the response count, narrows the interval considerably, and crosses the threshold where you can distinguish a genuine 0.9% response rate from a genuine 1.1% one. At 25,000 records, the reading is solid enough to forecast the economics of a national rollout.
The 10,000 floor applies to a single cell. If you are A/B testing two creative treatments or two list segments, each cell needs 10,000 records, which means a minimum total of 20,000.
There is a practical exception: campaigns with very high conversion values (equity release, legal services, vehicle sales) can occasionally make a smaller test work because a response rate of even 0.3% generates enough revenue to justify rollout. But the uncertainty in the rate estimate remains; you are simply accepting a wider margin of risk.
How does postcode targeting interact with lifestyle data?
Geography and lifestyle data work as multipliers, not alternatives. Postcode targeting defines the territory; lifestyle data determines which households within that territory are worth mailing. Used together, they produce a list that is both physically reachable within your distribution model and behaviourally aligned with your offer.
A practical example: a solar panel installer operating in South Yorkshire wants to run a door-drop campaign in the S postcode area. Geographic selection alone would return around 250,000 households. Adding owner-occupier status (excluding flats, which cannot have roof panels) brings this to roughly 140,000. Filtering further to households with a declared home-improvement interest reduces it to perhaps 60,000. Each filter costs something in print savings and gained response rate. The art of list selection is deciding where the optimal trade-off sits for your specific margins.
Postcode geography also matters for modelled income data. Income band models are calibrated at sector level using property values, council tax bands, and financial product data, so a sector selection and an income band select are naturally complementary: the geography drives the model accuracy, rather than conflicting with it.
MPS suppression, pricing, and the selection checklist
MPS suppression as standard
The Mailing Preference Service (MPS), administered by the Data and Marketing Association (DMA), allows UK consumers to register their address as opted out of unsolicited direct mail. Under DMA guidance and as a general industry expectation, any reputable supplier of a UK consumer postal file will suppress against MPS before delivering the list. You should confirm this explicitly before purchase. A supplier who does not mention MPS suppression is either failing to apply it or expecting you to apply it yourself; in either case, ask the question directly.
MPS typically removes 3 to 8% of a consumer file, varying by demographic: older, owner-occupier households register at higher rates than younger or rental households. The MPS file is updated monthly, so timing your mail date to the most recent update gives you the cleanest suppression.
How tightness of selection affects pricing
Direct mail list pricing in the UK is almost universally quoted as cost per thousand (CPM) records. Broad national files with minimal demographic filtering typically cost in the range of £70 to £120 per thousand. Adding demographic layers narrows the universe, which tends to push CPM upward as you move into more specialist territory: a file of declared charitable donors aged 55 to 74 with an owner-occupier flag in four specific postcode districts might cost £180 to £250 per thousand.
The pricing logic is straightforward. Tighter selection produces a smaller universe. A smaller universe is harder to replace if the initial count does not hit your target volume. The supplier is also applying more processing to generate the output. None of this means tight selection is bad value: a higher CPM paid for a list that converts at 1.8% will nearly always beat a lower CPM list that converts at 0.6%. The comparison that matters is cost per response, not cost per record.
List selection checklist
Before approving any direct mail list purchase, confirm every item in this checklist with your supplier. Gaps at the pre-purchase stage almost always cost more to fix than the time spent addressing them before you sign off.
| Checklist item | What to confirm | Why it matters |
|---|---|---|
| Geographic spec | Postcodes, sectors, or radius are listed explicitly | Avoids ambiguous region definitions that inflate counts |
| Demographic filters | Age band, tenure, household type are confirmed | Removes structurally ineligible households before print |
| Lifestyle/behaviour flags | At least one declared interest or propensity flag applied | Strongest predictor of response beyond demographics |
| MPS suppression | Supplier confirms MPS applied (date of last wash) | Industry standard; applies suppression before fulfilment |
| Deduplication | Existing customers and recent contacts removed | Avoids mailing current customers as cold prospects |
| Postal cleansing | File has been processed against PAF (Postcode Address File) | Reduces undeliverable mail and wasted postage |
| Output format | Fields confirmed (name, address, any additional selects) | Avoids reformatting delays at the print fulfilment stage |
| Volume and count | Supplier has provided a post-suppression count, not a gross count | Gross counts overstate the usable universe by 5 to 15% |
| Lawful basis | Consumer file is fully opt-in under UK GDPR and PECR consent | Ensures the data is lawfully usable for direct mail marketing |
| Test volume | Minimum 10,000 records per test cell confirmed | Below this threshold, response rate data is statistically unreliable |