By Shannendoah Gallagher, Social Linus Web Services. Last updated: October 21, 2026. Next scheduled update: March 2027, when USPS publishes its 2026 Household Mail Survey.
Search for the average direct mail response rate and you will see 4.4%, 9%, 4.9%, 3.63% and “2% to 5%,” sometimes on the same page. We went looking for where each number comes from. A few trace to a real survey. Several trace to each other.
This page sorts what we could verify from what we could only find secondhand and what we could not trace at all. It also includes our own 2021 pilot data. The sample is small, so we show you exactly how small.
The short answer
Plan a cold prospect list at roughly 1% to 3% response and a list of existing customers higher than that. Treat anything above those ranges as a result worth checking before you celebrate it. That range is our judgment, built from the sources below and our pilot. It is not a measured industry average, and nobody has published one you can verify for free.
What the public sources say
| Source | Year | What it reports | What it actually measures | Our check |
|---|---|---|---|---|
| USPS Household Mail Survey | FY2025 | Advertising mail was 62% of household mail | Share of mail volume, about 5,200 households | Confirmed through an independent analysis. |
| USPS Household Mail Survey | FY2025 | Existing customers read 64% of ad mail they receive vs 36% for prospects; 25% vs 3% “very likely to respond” | Reading and stated intent, not response | Found in one secondary write-up only. |
| ANA Response Rate Report (with Demand Metric) | 2025 | Headline direct mail average of 4.4% circulates widely, credited to this report | Survey of 324 marketers | The full report is open only to ANA members and survey participants. We could not find 4.4% in the public summary |
| DMA Response Rate Report | 2018 | 9% house list, 4.9% prospect list | Marketer-reported response | Secondhand only. |
| DMA Response Rate Report | 2016 and 2015 | House list 5.3% (2016) and 3.7% (2015); prospect 2.9% (2016) and 1.0% (2015) | Marketer-reported response | Secondhand only |
| Lob consumer surveys | 2025 | 67% of Gen Z and Millennial respondents say they have taken action after receiving mail | Recall and stated intent | Published by a company that sells mail automation. Not response rates |
| Social Linus pilot | 2021 | 1.97% tracked response across 1,317 mailpieces to a public-records prospect list | Tracked form submissions and tracked phone calls | Our own data. See below |
Why the numbers disagree
The same label means different things. “Response” can mean a call, a visit, a coupon redemption, a form fill or a purchase. A postcard that asks for a phone call and a catalog that asks for a credit card do not share a denominator in any useful sense.
House lists and prospect lists get blended. People who already buy from you respond at several times the rate of strangers. Any single “average” that mixes the two describes nobody’s campaign.
The surveys are self-selected, and it shows. The DMA’s reported house-list rate went from 3.7% (2015) to 5.3% (2016) to 9% (2018). Customer behavior did not more than double in three years. Who answered the survey changed.
The numbers get copied, and the copies lose their labels. This is the part we find hard to ignore. The same industry-by-industry table, with open rates and response rates to two decimals, shows up on more than one vendor page. Some credit the ANA/DMA. Some credit the vendor’s own data. We could not match it to anything the ANA publishes in the open. Direct mail “open rates” in particular cannot be measured, because no one knows whether a piece was opened. When you see one, ask who counted.
Our 2021 pilot, and what it can and cannot tell you
We tracked responses from five mail campaigns. The campaigns ran in 2021. The audience was deed holders of land in Central Texas, pulled from public records, and the offer was a cash offer for their land. Each piece carried a QR code to a landing page with tracked form submissions, plus a tracked phone number. This was a targeted prospect list. The recipients had no prior relationship with us, so the fair comparison is with prospect-list figures, not house-list figures. The 26 responses below count both form submissions and tracked calls. The two “LUSurvey” campaigns asked for more than a click. A response there meant completing a survey that confirmed the recipient owned the land, had an intended use for it, could give an estimated offer range, and was open to selling for the right cash offer. That makes those responses qualified leads, and a higher bar than a basic form fill. The survey campaigns drew 6 responses from 322 pieces (1.86%) and the other three drew 20 from 995 (2.01%). Samples that small cannot tell those two rates apart. The other three campaigns asked for a simple form fill to schedule a callback about the cash offer printed on the card. Campaign 1E2AB tested the market. Campaigns 1E1AA and 1E3AC used what the survey responses revealed to choose the cash offer printed on the postcard. The offer changed between campaigns, so a difference in rate reflects the offer as well as the mailing, which is one more reason not to rank them.
| Campaign | What it was | Recipients | Tracked responses | Rate | 95% range |
|---|---|---|---|---|---|
| 1E2AB | Market test | 432 | 14 | 3.24% | 1.94% to 5.37% |
| 1D2AC | LUSurvey | 144 | 3 | 2.08% | 0.71% to 5.95% |
| 1D1AA | LUSurvey | 178 | 3 | 1.69% | 0.57% to 4.84% |
| 1E1AA | Offer set from survey findings | 389 | 5 | 1.29% | 0.55% to 2.97% |
| 1E3AC | Offer set from survey findings | 174 | 1 | 0.57% | 0.10% to 3.18% |
| All five | 1,317 | 26 | 1.97% | 1.35% to 2.88% |
Read the right-hand column before the middle one. Our best campaign (3.24%) and our worst (0.57%) look like a near six-fold gap. With 14 and 1 conversions behind them, a standard significance test does not separate them (p of about 0.08). Five campaigns of a few hundred pieces each can show you a range. They cannot rank anything.
Grouping them tells a more interesting story, with a warning attached. The market test (1E2AB) drew 14 responses from 432 pieces (3.24%). The two campaigns that used survey findings to set the offer drew 6 from 563 (1.07%, range 0.49% to 2.31%). That gap clears the usual significance bar (p of about 0.02). But we picked the comparison after seeing the data, the offers differed, and the mailings did not run side by side, so treat it as a hypothesis to test and not as a finding. Response rate may also have been the wrong yardstick by then, because a tighter offer can pull fewer responses and still make the economics work. We do not publish offer amounts or cost per deal.
Two limits matter more. We counted only responses we could track, meaning form submissions and calls to the tracked number. Anyone who responded another way, such as calling a different number or replying in person, is missing, so 1.97% likely undercounts. Referrals are missing too. Some recipients pointed us to other properties, and we did not track those. And this is one business’s mail to one audience (Central Texas landowners) from 2021. We publish it as one data point with its error bars, not as an industry benchmark.
What the responses turned into
A response rate is a means to an end, so here is the end. The program overall produced 22 rental doors (units) under management. It also produced a 5-acre parcel of undeveloped land that was later developed into 20 rental units and a 33-lot RV park. Some recipients also referred us to other properties, and we did not track those referrals, so these totals may include results the 26 tracked responses do not account for. These are our own records and have not been audited.
What we do not publish
We do not publish an industry-by-industry table. We have not found a public, verifiable source for one, and we will not invent one. When our own campaign data reaches a sample size that supports industry cuts, we will add them and show the sample size next to every figure.
Want your own number? Every SignalForge postcard carries a tracked phone number and a QR code, so you can measure your own response rate. See how SignalForge follows up with website visitors by mail.
How we grade sources
- A: primary source, free to read, method visible.
- B: primary source, but paywalled or published by a company with something to sell.
- C: found secondhand only. We could not open the original.
- D: no traceable source.
How to cite this page
Gallagher, S. (2026). Direct Mail Response Rate Benchmarks, Source-Traced. Social Linus Web Services. https://sociallinus.com/growth-hub/direct-mail-response-rates/
Update log
- October 21, 2026: first publication.
- March 2027 (planned): refresh after the USPS 2026 Household Mail Survey release, plus any new campaign data.
