Does direct mail still work for home service companies? The break-even math

Direct mail still works for home service companies, but not the way it's usually sold. The question isn't "what response rate does mail get?" Nobody can tell you that for your town, your offer and your list, and the industry averages that get quoted are mostly from other industries. The useful question is what response rate do I need to break even, and the answer is often surprisingly low when you mail the right people, and surprisingly high when you mail everyone.

This post prices both approaches with USPS rates in effect October 2026, gives you a break-even formula you can run on a napkin, and shows how to test whether mail actually added jobs, rather than just collecting credit for customers who would have called anyway.

What mail costs in October 2026

Postage is the one number in direct mail you can verify exactly. USPS Price List Notice 123 took effect October 4, 2026.¹ Here are the prices that matter for a contractor, per piece.

Sources: USPS Notice 123¹ and the USPS EDDM page, which lists EDDM Retail at $0.26 per piece with a 200-piece minimum and a 5,000-piece daily maximum per ZIP Code.² Printing, design and mail-house fees are extra and vary widely, so the examples below use clearly labelled assumptions for them.

For scale: USPS delivered to 170.4 million delivery points in fiscal 2025, about 157.8 million of them residential.³ A metro service area of 80,000 homes is a lot of postcards. That's the core problem with saturation mail. Most of the homes on a route don't need you this month.

The break-even response rate

Here's the formula that decides whether a mail piece can work.

It answers: what share of the homes I mail need to respond for this drop to pay for itself on the first job alone? The beauty is that it doesn't require knowing anybody's average response rate. You compare the break-even number to what you think is plausible, and if the break-even is absurd, you don't mail.

Worked example: two very different mailers

All figures except postage are hypothetical assumptions. Replace them with yours.

Mailer A: EDDM tune-up offer to 5,000 homes. Postage $0.26, printing an assumed $0.14, so $0.40 per piece and $2,000 per drop. Say 70% of people who call book, and a booked tune-up yields an assumed $120 of gross margin on the first visit, counting the occasional repair.

Break-even = $0.40 ÷ (0.70 × $120) = 0.48%, or about one booked response for every 210 homes.

Mailer B: personalized postcard to 600 of your own customers whose system is 12 or more years old. Postage $0.453 presorted, printing and personalization an assumed $0.30, so about $0.75 per piece and $450 per drop. These customers already trust you. Say 45% of responders buy a replacement, and a replacement yields an assumed $3,200 of gross margin.

Break-even = $0.75 ÷ (0.45 × $3,200) = 0.05%. One sold replacement from the whole drop of 600 pays for it about seven times over.

Mailers C and D in the chart use the same method. C is the personalized postcard ($0.75) sent to 250 customers with unsold estimates, assuming 50% of responders buy and $3,500 of margin per sale. D is the EDDM piece ($0.40) carrying a replacement offer, assuming 40% of responders buy at $3,200 of margin.

Mailer D is the trap. On paper, an EDDM replacement offer has the lowest break-even because replacement margin is so large. In practice, almost nobody on a random route needs a new system this month, and the people who do will get three quotes. A break-even that looks easy can still be unreachable if the list is wrong. The formula tells you the threshold; your list decides whether you can clear it.

Where mail earns its keep: lists from your own job history

The reason Mailer B works isn't the postcard. It's the list. A contractor's job history is a better targeting engine than any purchased list, because it knows things no list broker does: what's installed, how old it is, what you quoted and what the customer said.

A mailed piece also has a legal advantage over texts: you don't need the customer's prior consent to mail them, as you would for a marketing text sent with automated technology.⁴ Our post on texting customers legally covers why that matters. Mail still has rules: claims must be honest, and federal law makes a solicitation nonmailable if it suggests a government connection without the required disclaimers.⁵ Skip the seal-and-"final notice" look entirely; it erodes trust even when it's legal. And you should keep a do-not-mail flag for anyone who asks to stop. Our companion post, the best marketing list you own is your job history, goes deeper on building and cleaning these segments.

In TradesWork, ask the business flags thin weeks ahead and suggests audiences to work, such as unsold estimates, maintenance due and replacement candidates, and every list behind a number exports to CSV for a mail house. TradesWork doesn't print or send mail itself.

Measure it with a holdout, not a coupon count

Here's the uncomfortable truth about mail attribution. When you mail 600 loyal customers and 9 of them call, some of those 9 would have called anyway. Coupon codes and tracking numbers count responses; they don't tell you what the mail added. If you only count responses, mail to your best customers will always look brilliant.

The fix is old and simple: a randomized holdout.

1. 1

   Build the eligible list

   For example, 750 customers with equipment 12 or more years old, excluding anyone with an open estimate or a job in the last 60 days.

2. 2

   Randomly hold out about 20%

   Sort by a random number and set aside 150. Don't cherry-pick. They get nothing.

3. 3

   Mail the other 600

   Same piece, same week. Use a dedicated phone number and URL so you can also see direct responses.

4. 4

   Wait 60 to 90 days

   Replacements have a sales cycle. Count booked jobs and sold revenue in both groups, from any source, not just the tracking number.

5. 5

   Compare rates, not counts

   Incremental jobs = (mailed group's booked rate − holdout's booked rate) × mailed count. That's what the mail bought.

A worked example with hypothetical results: the 600 mailed homes produce 15 sold replacements (2.5%) over 90 days. The 150 held-out homes produce 2 (1.3%). Incremental sales = (2.5% − 1.3%) × 600 ≈ 7 replacements the mail actually caused, not 15. At $3,200 of margin each, that's about $22,400 of incremental margin on a $450 drop. Still a great result, and now you know it's real.

Small lists make noisy tests. With a few hundred homes, a difference of one or two jobs can flip the result. Run the same segment test across two or three drops before you draw firm conclusions, and keep the holdout random each time.

Tracking mail into the job record

For any of this to work, each job needs to carry where it came from.

In TradesWork, a phone number can be tied to a campaign so calls to that number land on the job with the campaign attached, and online booking tags jobs to its own campaign. Revenue then rolls up by campaign in reports. The holdout comparison is a spreadsheet step: export both lists and their jobs, and compare rates.

app.tradeswork.app · Demand versus capacity forecast by week

When direct mail is the wrong tool

Mail is slow, physical and expensive per touch, which is exactly why it still gets read. Used on a list you built from your own work, timed to a thin week, and measured honestly, it's one of the few marketing channels whose results you can prove. For timing, see our shoulder-season playbook, and for broader channel planning, the HVAC marketing guide.