Ask most people to measure a mail campaign and you get one number: response rate. It is a fine number. It is also nowhere near enough to run a program on. After building tracking software and watching campaigns report in, here are the seven KPIs I actually make decisions with, plus three I have learned to ignore.
The seven that matter
1. Scan rate by segment. Not your overall rate, your rate per neighborhood, per list source, per format. The ANA's benchmark research puts typical direct mail response well above email, with house lists running several times stronger than cold prospect lists. Your segments will vary even more. The segments are where the budget decisions live.
2. Cost per response. Total campaign cost divided by tracked responses. A 2% response at 80 cents a piece beats a 3% response at $1.60 a piece. This number keeps format upgrades honest.
3. Cost per acquisition. Responses are nice; customers pay the bills. Track scans through to booked jobs or closed deals, even roughly. This is the number that survives a conversation with your accountant.
4. Repeat scan rate. Someone who scans your card three times over two weeks is telling you something no response rate can. I treat repeat scanners as a hot list for personal follow-up.
5. Time to first response. How long after delivery do scans start? JICMAIL's panel data shows mail staying active in the home for over a week, with items driving repeated interactions across days. Your response window is longer than you think, and knowing yours prevents premature verdicts.
6. Delivery rate and timing. If 8% of a list never arrives, that is a list quality problem masquerading as a creative problem. Piece-level delivery events separate the two.
7. Response by creative version. When you split-test two offers, per-recipient tracking scores the winner with real names attached instead of a hunch.
The three I ignore
Impressions. Mail volume is not an outcome. Five thousand pieces delivered is a cost line, not a result.
Vanity scan totals. Total scans without unique scanners hides one enthusiast scanning nine times. Track both, decide on uniques.
Industry average response rates, in isolation. Benchmarks are for context, not verdicts. A 1.5% response on a $40 postcard campaign selling a $12,000 service is a phenomenal outcome, and an average response rate on an unprofitable offer is still unprofitable.
If you can only track three things, make them scan rate by segment, cost per response, and repeat scans. Those three, watched monthly, will out-coach any dashboard full of vanity numbers. And if you want them computed for you, that is literally what Relay's report page does.