By Xcel Media Group ·
Two numbers, and only one of them matters
Almost every automotive direct mail benchmark you'll find online reports a single "response rate." That number is misleading, because in car retail there are really two different measurements:
- Tracked response — the people who used a response mechanism printed on the piece: a call-tracking number, a QR code, or a personalized URL (PURL). It's real and immediate, but it badly undercounts, because most buyers just drive in or search the store online without ever mentioning the mailer.
- Sold units (matchback) — the households from the mailing list who actually bought or serviced, found by matching the full list back against sales and repair-order records in the dealership's Dealer Management System (DMS).
Both are benchmarked below. The first table reports tracked response by campaign type. The second reports sold units per 10,000 mailed across every campaign Xcel put in home in a defined window — 436 campaigns and 3.0 million pieces — rather than a hand-picked sample.
One thing had to be handled honestly to make that second table mean anything. About a third of campaigns run at stores whose DMS is not connected for matchback. Those campaigns report zero sold units and zero repair orders no matter how the mail actually performed — they are missing data, not failures — and averaging them in as zeros drags every figure down. For credit mail it nearly halves the result. So the sold-unit medians are taken across matchback-connected campaigns only, and the table shows how many campaigns that is in each row.
The practical consequence of having both: campaign types with lower tracked response can still produce strong sold-unit counts, which is exactly why judging automotive mail on response rate alone leads dealers to cut the wrong campaigns.
Response and sold units by campaign type
The benchmarks below are drawn from real, authorized Xcel campaigns. The figure shown is the median campaign of each type with the full range of individual campaigns beside it, so one unusually strong result cannot stand in for the type.
A few patterns hold up across the data:
- Handwritten mail posts by far the highest tracked response — a hand-addressed, real-pen piece gets opened and often kept — but it is a premium format best used on high-value lists rather than sprayed at a market.
- Service and recall mail punches above its weight because the reason to act is concrete and personal: a due service or an open recall, not a general offer.
- Credit and special-finance mail swings widest (0.9% to 7.2%), because response depends almost entirely on how well the offer matches the household's actual financing situation.
- Conquest posts the lowest tracked response — it reaches households with no relationship to the store — but it is the only one of these that grows the customer base, so it is judged on cost per net-new buyer rather than response.
- Owner/database mail looks modest on response and is not — these recipients already know the store, so many simply drive in without using a tracked path, which is precisely the undercount this metric suffers from.
How to read these against your own store
Treat these as directional, not guarantees. Real automotive results vary widely with your make, market, offer, list quality, and season, and this is a sample of campaigns rather than the entire universe of mail. A luxury store's numbers won't match a high-volume domestic store's, and a tight equity list will always beat a broad conquest blast.
The honest way to use benchmarks is to set expectations by campaign type, then measure your own campaigns the same way — tracked response for the immediate signal, and DMS matchback for the truth. Xcel's UpDash platform runs this matchback automatically against your DMS, so every campaign reports both its tracked responders and the sold units and repair orders attributed to it — which is how these benchmarks were produced in the first place. Xcel has mailed 250M+ pieces for 2,700+ dealerships across 46 states since 2009, so the underlying dataset is large; the figures here are a published, verifiable slice of it.
Three planning cautions. First, these are medians rather than averages, deliberately. One exceptional campaign would drag an average somewhere no dealer should budget against; the median describes the campaign in the middle, and the range beside it shows how much spread sits behind that single number. The response table rests on a much smaller sample than the sold-unit census, so treat its per-type figures as directional.
Second, tracked response is a floor, not a forecast. It counts only the people who used a path printed on the piece. Owner and database campaigns are the clearest example — those recipients already know the store, so many just drive in, and the tracked figure understates them more than it does conquest.
Third, response is not revenue. For budgeting, size on sold units, not response. The census median across all four data sources is 29 per 10,000, so planning on the order of 20-30 remains the right conservative anchor — but the data source matters more than the average suggests: owner and equity mail ran a median of 44 per 10,000 while untargeted saturation mail managed 3. Replace the assumption with your own DMS matchback as soon as you have one campaign's worth of it. A high response rate on a list that cannot buy is not a win.
| Campaign type | Campaigns | Pieces mailed | Tracked response — median (range) |
|---|---|---|---|
| Handwritten | 4 | 24,080 | 8.9% (3.3-10.6%) |
| Service & recall | 3 | 5,388 | 2.9% (2.3-7.1%) |
| Credit / special finance | 4 | 27,642 | 2.6% (0.9-7.2%) |
| Database / equity (owner) | 4 | 18,847 | 1.6% (1.3-2.8%) |
| Bankruptcy / Fresh Start | 4 | 10,488 | 0.9% (0.7-2.3%) |
| Conquest | 3 | 13,337 | 0.8% (0.7-1.8%) |
| Data source | Campaigns | With DMS matchback | Sold per 10,000 — median (IQR) |
|---|---|---|---|
| Database / equity (owner) | 191 | 141 (74%) | 44 (27-68) |
| Bankruptcy / Fresh Start | 22 | 5 (23%) | 39 (14-93) |
| Credit / special finance | 186 | 125 (67%) | 20 (12-31) |
| Saturation (untargeted) | 37 | 11 (30%) | 3 (1-6) |
| All four combined | 436 | 282 (65%) | 29 (16-52) |
Frequently asked questions
What is a good response rate for automotive direct mail?
It depends entirely on the campaign type, and "response" itself is ambiguous. Tracked response (calls, QR, PURL) on real campaigns ran roughly 1-3% for database and conquest, 2-7% for service, and 3-10%+ for handwritten. Across 22 real Xcel campaigns (99,782 pieces), median tracked response was about 8.9% for handwritten, 2.9% for service and recall, 2.6% for credit, 1.6% for owner/database, 0.9% for bankruptcy, and 0.8% for conquest. All of those undercount, because plenty of recipients drive in without using a tracked path. Judge each campaign type against its own benchmark rather than one industry number, and judge the campaign itself on sold units from DMS matchback, not response alone.
How many cars should a direct mail campaign sell?
Across 436 campaigns and 3.0 million pieces, the median was about 29 sold units per 10,000 mailed — but the data source drives it more than anything else. Owner and equity mail ran a median of 44 per 10,000, credit and special-finance mail 20, and untargeted saturation mail just 3. Those medians count only campaigns at stores whose DMS is connected for matchback; a third of campaigns run at stores with no DMS feed and report zero sold regardless of performance. For budgeting, planning on the order of 20-30 per 10,000 is a reasonable conservative anchor for targeted mail, then replace it with your own matchback once you have a campaign of your own to measure.
Which direct mail type sells the most cars?
On tracked response, handwritten mail leads by a wide margin — a median of 8.9% against 0.8% for conquest — because a real-pen, hand-addressed note gets opened and often kept. But response and sold units are not the same thing. Owner, equity, and handwritten lists reach people who already own a vehicle you can trade, so they tend to convert well and are also the most undercounted by tracked response, since those recipients often just drive in. Conquest and bankruptcy mail convert less often per piece but reach net-new and credit-challenged households the other lists cannot, so most dealers run a mix and judge each type on its own cost per sold unit.
Are these numbers typical for any dealership?
They're directional, not guaranteed. They come from real, DMS-matched campaigns, but automotive results vary widely by make, market, offer, list quality, and season, and this is a sample rather than every campaign ever run. Use them to set expectations by campaign type, then measure your own store's results with tracked response plus DMS matchback.
Sources
- Sold-unit census: every direct-mail campaign Xcel put in home in the stated windows, pulled from the UpDash platform (database/equity Q1 2025, credit H1 2025, bankruptcy and saturation full-year 2025) — 436 campaigns, 3,033,628 pieces, non-automotive clients excluded. Medians are computed across the 282 campaigns at stores with DMS matchback connected.
- Xcel Media Group / UpDash — 22 real, authorized dealership direct-mail campaigns (99,782 pieces), each matched to sold units and repair orders in the dealer's DMS; response and sold-unit figures published on xcelmediagroup.com. Individual results vary.
- Aggregate program scale: 250M+ pieces mailed for 2,700+ dealerships across 46 states since 2009 (Xcel Media Group).
- ANA/DMA Response Rate Report — external context: house/customer lists respond at meaningfully higher rates than cold prospect lists (approximate; varies by year and list type).
- Automotive attribution relies on DMS sales matchback because most influenced buyers never present or reference the mailer at the dealership (established industry methodology).
See how Xcel runs this in practice
Explore ROI matching in UpDash