Personalisation can be tested, but channel averages do not establish its effect. The ANA's 2023 report explicitly cautions that its self-reported results should be used for information rather than as benchmarks. Compare personalised and control mailings within the same audience before forecasting a lift. The reason to be careful is that published averages mix audiences, offers, formats and industries, so a figure drawn from one context predicts very little about another. Self-reported figures also tend to over-represent campaigns their owners were willing to report. A within-audience comparison avoids both problems: split one list at random, send the personalised and control versions at the same time, and hold everything except the personalisation constant. Decide the sample size and the response window before the drop rather than afterwards, and treat the result as evidence about your own list and offer rather than as a number to generalise from.
This guide explains variable-data preparation. PDFImpose arranges the resulting PDF for printing; create and verify the data merge in another tool first.
What do the response-rate benchmarks actually say?
The Association of National Advertisers published its 2023 Response Rate Report in February 2024. Its methodology describes survey responses and warns about limited sample sizes, estimates, and sparse ROI data. Those limitations matter more than a headline percentage when planning a particular campaign.
A comparison between mail, email, and social advertising combines differences in audience, offer, format, cost, and measurement. It cannot tell you how much adding a name changes the response to an otherwise identical mail piece.
Do different mail formats get different response rates?
Different formats can have different costs and audience responses. Compare equivalent audiences and offers when evaluating postcards, letters, or envelopes. Do not attribute a format or list difference to personalization unless the study design actually isolates it.
| Measurement | Record explicitly | Why it matters |
|---|---|---|
| Audience | House list or prospects | Prior customer relationships affect response |
| Format and offer | Mail piece, incentive, and timing | These can change along with personalization |
| Response definition | Scan, inquiry, purchase, or other action | Different actions produce different rates |
| Denominator | Sent, delivered, or reachable recipients | Rates must use the same basis |
| Profit measurement | Contribution after all campaign costs | Revenue is not profit |
| Study reliability | Sample size, estimates, and uncertainty | A small survey result is not a forecast |
What's the evidence specifically on personalization?
Widely repeated claims about a large response lift from adding a recipient’s name need a traceable study and methodology. This guide does not establish such a universal lift. A claim without a suitable comparison group should not become a production forecast.
Personalization and targeting are also different interventions. A relevant offer sent to existing customers is not a controlled comparison with generic mail sent to unrelated prospects. To estimate the effect of personalization, keep the list, offer, format, timing, and response definition comparable.
How does direct mail ROI compare to digital channels in the same study?
The report includes a high reported ROI for house-list direct mail, but its own methodology cautions against treating results as benchmarks. Compare profit definitions and all campaign costs before comparing channels. A channel ranking does not establish the effect of personalization.
It's also worth noting that reported ROI and reported response rate aren't the same measurement, and a campaign can score well on one without scoring equally well on the other. A low-response, low-cost postcard drop can still post a strong ROI if the cost per piece is low enough; a high-response, high-production-cost personalized mailer can post a weaker ROI despite the stronger response number. Read the two figures together, not as substitutes for each other.
What are the honest caveats here?
- Self-reported survey data. The ANA report is built from respondent-submitted campaign results, not a controlled experiment — it reflects what marketers report, aggregated.
- Averages hide variation. Audience, offer, format, and response definitions differ between campaigns. A published average is not a forecast for your mailing.
- No universal personalization lift is established here. A repeated percentage needs a traceable comparison and methodology before it can support a claim.
- Format and list type are confounded. Benchmark comparisons across format or list type aren't holding personalization constant, so isolating personalization's individual contribution from published aggregate data isn't possible with precision.
How should you use these numbers in practice?
Use external reports to frame questions, then build estimates from your own comparable campaigns. Define the response event and attribution window before mailing. Randomly allocate recipients to a control and a personalized version, record delivery failures, and keep the remaining production variables consistent.
Variable data printing automates the placement of record-specific content. It can reduce manual transcription, but bad data or field mapping can still repeat the same error across a run. Verify names, identifiers, and output records before approving the mailing.
In the illustrative comparison above, 40 responses from 1,000 recipients is 4%, while 60 from 1,000 is 6%. The observed difference is two percentage points, or 50% relative lift. Those calculations alone do not establish a reliable effect: account for sampling uncertainty, repeat the test where appropriate, and compare incremental profit after personalization costs.
Create names, numbers, codes, and other variable content in a data-merge tool first. Bring the completed PDF into PDFImpose for cut-and-stack imposition.
Common questions
What is a typical direct mail response rate?
There is no single response rate suitable for every campaign. Match the audience, format, offer, event definition, and measurement window. The ANA’s 2023 report cautions that its survey results are informational and should not be treated as benchmarks.
Does adding a name reliably improve response?
This guide does not establish a universal lift from adding a name. Use a comparable control group and measure the actual response and profit of your campaign before assuming personalization is effective.
Is direct mail's reported ROI higher than email or social media?
Reported channel comparisons depend on the sampled campaigns, cost definitions, and audience. They do not prove that one channel will earn more for your business, or that personalization caused the difference.
Why do different sources report different direct mail response rates?
Response rate depends on list type (house vs. prospect), format (postcard vs. letter vs. oversized envelope), industry, and offer strength. A single average figure without those variables attached should be treated as a rough industry-wide midpoint, not a prediction for a specific campaign.
Why not just use a published industry average?
Because published averages mix audiences, offers, formats and industries, so a figure from one context predicts little about another. Self-reported figures also over-represent campaigns their owners chose to report, which biases the result upward in a way the headline number does not show.
How should a personalisation test be set up?
Split one list at random, send the personalised and control versions at the same time, and hold everything except the personalisation constant. Decide the sample size and the response window before the drop rather than afterwards, so the result is a measurement rather than an interpretation.
1. Association of National Advertisers, Response Rate Report, 2023, published February 22, 2024. Read the methodology, especially the warning about using the survey as a benchmark. The response example in Fig. 1 is invented to explain arithmetic.