How to measure direct mail ROI by location

To measure direct mail ROI by location, mail some markets, hold matched markets back, and compare customer counts at the mailed locations against a weighted blend of the unmailed ones, starting from the in-home date. The gap is the extra customers. Value them at margin, subtract every cost including the discounts you gave away, and do it store by store. Coupon counts alone will mislead you in both directions.

Mail is one of the few channels where you decide exactly which households get the ad. USPS Every Door Direct Mail lets you pick carrier routes around a store and send up to 5,000 pieces a day per ZIP Code, with retail postage listed at $0.26 a piece for marketing flats (Source: USPS, accessed 2026). That control is a gift for measurement. Most brands waste it by mailing every store at once and counting coupons.

Why direct mail ROI by location is harder than it looks

• Regulars use the coupon too. A $5-off postcard lands in the kitchen of someone who orders every Friday anyway. They redeem it. That’s a discount, not a customer.
• New customers often skip the coupon. Someone sees the card, remembers you exist, and orders online without the code. Mail did its job and the redemption report never hears about it.
• Timing is fuzzy. Pieces drop over several days, sit on counters, and get acted on over weeks. A read that starts on the mail date and stops after two weeks catches the front of the wave.
• Trade areas overlap. Mail routes for one store often reach households closer to another. If the neighbouring store is in your comparison group, the comparison is contaminated.

The direct mail ROI formula, per location

ROI = (extra customers × contribution per customer − total mail cost) ÷ total mail cost. Three things go into total mail cost, and the third is the one people forget.

1. Print and design. Including the second version if you split-test creative.
2. Postage and list or route fees. EDDM keeps this simple. Addressed mail adds list rental and data processing.
3. The discount on every redemption. Every coupon redeemed gave away the discount, whether the customer was new or a regular. Count all of it.

Contribution per customer is ticket times margin, not revenue. A $34 order at 35% contribution is worth $11.90. Our guide to cost per incremental customer covers how to set that number so finance agrees with it.

Three ways to build a holdout for direct mail

The design you can use depends on what data you have. Pick the strongest one your systems support.

• Market split (works for almost everyone). Mail every store in some markets and none in matched others. You only need weekly counts by location from Toast, Square or your POS. Keep a buffer: no mailed routes within reach of an unmailed store.
• Route holdout within a trade area. Mail half the carrier routes around each store and hold back the other half. This needs a customer address or ZIP on each order, which delivery and booking businesses often have and walk-in businesses usually don’t.
• Address-level holdout. For addressed mail to your own customer list, randomly hold back 10–20% of names and compare their purchases. The cleanest design there is, and only possible with customer-level data.

For a multi-location brand, the market split is usually the right starting point. It reads what the mail did for the location, which is the decision you actually make. The comparison of matched-market tests and geo holdouts goes deeper on picking markets.

Time the read to the in-home date

Ask your mail house for the expected in-home window for each drop, and start the read there. Then read for about six weeks. With two drops three weeks apart, that covers both waves and the tail of the second.

Hold everything else still in both groups for those six weeks: no new LTO in the mailed markets only, no paid social push, no price change. If the brand runs a national promotion, fine. It hits both groups and the twin absorbs it. (If you are testing the offer rather than the mail, measuring limited-time offer lift is a different design.)

Worked example: direct mail ROI across 16 pizza stores

This example is illustrative. A 40-store pizza chain mails EDDM postcards with a $5-off coupon to 8,000 households around each of 16 stores in six markets, in two drops three weeks apart. That is 16,000 pieces per store, 256,000 in all. Print plus postage comes to $0.38 a piece ($0.26 postage plus about $0.12 print), so $6,080 per store and $97,280 in total. The other 24 stores, in ten markets, get no mail and form the twin. Each store takes about 1,500 orders a week.

The working, for the suburban group: 6 stores × 1,500 orders × 9.0% × 6 weeks = 4,860 extra orders. Cost is 6 × $6,080 of mail plus 1,560 coupons × $5 = $44,280. Value is 4,860 × $11.90 = $57,834. ROI = ($57,834 − $44,280) ÷ $44,280 = +31%. The other rows follow the same steps, with 1,200 and 600 coupons redeemed.

Notice two things. First, the chain-wide answer (−20%) would have killed a program that is clearly profitable in six stores. Second, coupons redeemed (3,360) and extra orders (7,650) are nowhere near each other. Some redeemers were regulars cashing in. Many extra orders came without a coupon. Neither number stands in for the other.

What to do with a direct mail ROI result by location

Keep mailing the suburban family stores. In the example, mailing only those six would have turned a $23,045 loss into a $13,554 profit on the same creative. For the mixed group, try something cheaper before giving up: one drop instead of two, or a tighter route selection filtered to households with kids. For the dense urban stores, stop. Apartment buildings, shared mailrooms and heavy delivery-app use are a bad mix for postcards.

One caution. A four-store group has a wide range. If the urban read is +1.0% with a 95% range from −2.5% to +4.5%, the honest call is “no evidence it works,” not “proven worthless.” That’s still enough to stop spending there, but retest if the mix of stores changes. The guide to allocating ad budget across locations covers how big each move should be.

Do coupon codes measure direct mail ROI?

No, but keep them. Unique codes per drop or per creative version are a cheap way to compare which postcard pulled harder, because both versions suffer the same biases. A code count is also a quick sanity check that the mail actually arrived.

What codes can’t do is tell you how many customers the mail created. They overcount where regulars redeem and undercount where new customers ignore the code. In most programs both happen at once, so you can’t even say which direction the error runs. Treat “ROI by coupon” as a creative metric, and get the business answer from the location comparison.

Common mistakes when measuring direct mail ROI by location

• Mailing every store at once. With no unmailed comparison, a good week and a good postcard look identical.
• Reading from the mail date. You lose the first few days of response to the postal pipeline and stop before the tail.
• Leaving the discount out of the cost. At $5 a coupon, 3,360 redemptions is $16,800. That was 15% of the example’s total cost.
• Comparing mailed stores with their neighbours. If the next store over received spillover mail, it can’t serve as the comparison.
• Reporting one chain-wide number. Mail economics vary with housing type, household size and distance to the store. The average hides the stores that make the program worth running.
• Over-reading small groups. Two or three stores per group gives ranges too wide to act on. Group stores by trade-area type before the test, not after.

When this approach doesn’t fit

A single location has no unmailed twin to compare against. If you record a ZIP code for each job or appointment, a route or ZIP holdout around the one store can work instead. Countlift runs ZIP-level tests for single-location service businesses in early access; for walk-in businesses with one store, the single-location measurement guide covers the realistic options.

And if a drop is a few thousand pieces for one store, you won’t see it in weekly counts against normal noise. Mail it if the coupon math works, and don’t spend a week analysing it.

Questions people ask about direct mail ROI

Direct mail gives you more control over who sees the ad than any other local channel. Use that control to build a fair comparison, and you’ll know which stores to mail next quarter and which to leave alone.

Written by The Countlift team at First Byte in The Woodlands, Texas. Questions or corrections? Email hello@countlift.com.