The Anatomy of a Postcard That Makes the Phone Ring: Offers, Tracking, and the A/B Test That Settles Arguments

Show a room full of contractors two postcard designs and you'll get a strong opinion from every chair. Mail both cards to identical audiences with separate tracking numbers and the opinions stop mattering — the phone counts the votes. In our case study, two cards to identical 2,000-household audiences pulled 31 calls and 67 calls respectively. A 2.1× gap, decided by the offer, settled by the phone numbers.

This article is the checklist behind cards like the winner: what goes on them, how responses get counted, and how to run the test that keeps making them better.

Part 1: The offer does most of the work

The largest response differences we see between cards to the same audience come from the offer — not the photo, not the font. Four rules, all supported by industry benchmark analysis:

Specific beats vague. "$79 drain clearing" outperforms "10% off plumbing services" even when the dollar value is similar. Benchmark analysis of promotional mail finds that specific, bounded, low-effort offers measurably beat percentage discounts — a percentage forces the reader to do math about an unknown price; a flat number is a decision they can make at the mailbox.

Bounded beats open-ended. "This month only" or "through June 30" creates a decision window. Open-ended offers get set on the counter to "deal with later," and later never calls. (Mail does hang around — about 17 days in the home on average, per the ANA — but a deadline is what converts the lingering into a call.)

Value-add can beat discount. The 2.1× winner above wasn't a deeper discount — it was "free camera inspection with any drain job" beating "$79 drain clearing." A value-add costs you less than a price cut, positions you as thorough rather than cheap, and legitimately surfaces bigger repairs the homeowner didn't know about. Test both types; let the numbers pick.

One offer per card. A card with three offers is a card with no offer. Pick the one that matches the season and give it the whole billboard.

And time it to the season. Home services mail benchmarks highest — 1.5–4.5% response — when a seasonal, urgency-driven offer lands 4–6 weeks ahead of peak demand: pre-summer AC tune-ups in April, drain campaigns ahead of spring, water-heater checks before the first freeze. The same card off-season benchmarks at roughly half that.

Part 2: The format — make it impossible to miss

Postcards are the trade's workhorse for a reason: no envelope stands between your offer and the reader, and the ANA's format data has postcards averaging around 5.7% response — they're used in 76% of house-list campaigns, the most of any format.

Go oversized. Benchmark guidance for home services specifically calls for 6×9 or larger to command attention in the mail stack, and USPS accepts 6×9 cards at the First-Class postcard rate — double the standard card's real estate without letter-rate postage.

Layout rules that survive contact with a real mailbox:

• Front: the offer, readable at arm's length in under two seconds. Headline = the offer, not your company name. One strong photo (a real tech and truck beats stock art).
• Back: the trust stack — license number, years serving the area, review count and rating, guarantee — plus the offer repeated, the deadline, and the response paths.
• Personalize what you can. ANA/DMA data credits adding the recipient's name with up to a 135% response lift, and variable-data printing that tailors pieces pulls 2–3× the response of generic ones. At minimum, make it local: "Serving [neighborhood] since 2009" on a card that actually goes to that neighborhood.

Part 3: The response path — where good cards go to die

A card can win the mailbox and still lose the campaign in the ten seconds after the reader decides to act. Two non-negotiables:

A dedicated tracking phone number on every card version. It forwards invisibly to your regular line, and it turns your campaign from a feeling into a count: calls per card, per neighborhood, per week. This is now standard professional practice — 82% of direct mail marketers track response through online tracking and codes, per the ANA — and it's the entire mechanism that makes the A/B test in Part 4 possible. It also gives you recordings, which is how you find out whether the calls your card generated were actually booked (industry average: only ~38% are — see Your CSR Is Half Your Marketing Budget).

A QR code that lands somewhere built to convert. Pairing print with a digital response path lifts overall response 20–30% per the ANA — but only if the digital side holds up. A QR scan is a mobile visit by definition, and Google's research found 53% of mobile visits are abandoned when the page takes over three seconds to load. Never send the code to your homepage. Send it to a fast, single-purpose page with exactly two buttons above everything else: Call now and Pick a time, wired to your real calendar. (In our case study, fixing this one leak took scan-to-booking from roughly 1 in 6 to 4 in 10 — on responses already paid for with postage.)

Part 4: The A/B test that settles arguments

Here's the practice that separates mailers who improve from mailers who repeat themselves. The structure is simple and worth doing exactly:

1. Split one audience in half, randomly. 2,000 and 2,000 from the same neighborhoods — not "Card A to the north side, Card B to the south," which tests neighborhoods, not cards.
2. Change one variable. Start with the offer; it's the biggest lever. (Design, headline, and format tests come later.)
3. Give each version its own tracking number and QR code. Otherwise you've mailed two cards and learned nothing.
4. Let it run 6–8 weeks. Mail responses arrive on a long tail; calling a winner at week two is how you crown the wrong card.
5. Count calls and booked jobs per version — then keep the champion and write a new challenger. Every future mailing tests something against the reigning winner, so the creative improves on a schedule instead of by argument.

One piece of statistical honesty most vendors skip: at a 2% response, a 2,000-card cell produces about 40 responses. That's plenty to detect a big difference — a 2× gap like 31 vs. 67 calls is unambiguous — but not enough to detect a 10% difference. So test bold differences: a different offer, a different mechanism, a different season hook. Don't burn a mailing testing button colors; the sample can't hear differences that small.

The pre-flight checklist

Before any card prints, it should pass all ten:

1. One specific, flat-value offer (not a bare percentage)
2. A real deadline
3. Offer readable at arm's length on the front
4. 6×9 or larger format
5. Trust stack on the back: license, years, reviews, guarantee
6. Local relevance (right neighborhood named, right season)
7. Dedicated tracking phone number for this version
8. QR code to a fast, two-button booking page — never the homepage
9. A challenger version, split randomly, testing one bold variable
10. The people answering your phones briefed on the offer before the first card lands

Ten yeses and the card has done everything a card can do. From there, the campaign is won or lost by targeting — which streets you mail — and by the phone — how the calls get handled. Cards, targeting, phones: instrument all three and marketing stops being a debate.

We design the cards, split the tests, assign the numbers, and score every call automatically on each campaign we run — and the targeting starts with a free market analysis of your own job history. No contract. Pay per card. The phone counts the votes.