Direct Mail vs Google Ads vs Email: What Is Actually Published in 2026

What the mail data actually says

There is no published response-rate benchmark for saturation mail, and that is not a gap in our reading. The report everyone cites carries no saturation, EDDM or carrier-route category, so a rate taken from it was carried over from a different kind of mailing. Vendors who publish a number are publishing their own field experience. That is worth reading and it is not a benchmark.

What USPS does publish is attention, measured in a household survey it files with its regulator. 36% of mail from a business the household has no prior relationship with gets read, and 3% of those households say they are very likely to respond. Mail to a past customer reads at 64%, with 25% very likely to respond. A saturation drop to a neighborhood is the prospect row. Plan against that one, not the customer row, and not a percentage off a vendor page.

The survey also measures what an offer does. Put a coupon on the piece and stated intent to respond among prospects moves from 2% to 5%. Note the verb in all of this: the survey measures what people say they are likely to do, which is not the same as a response rate, and it says so itself.

• Gets read, no prior relationship: 36% (USPS Household Mail Survey, FY 2025, Table 5.5)
• Very likely to respond, no prior relationship: 3% (USPS Household Mail Survey, FY 2025, Table 5.5)
• Gets read, past customer: 64% (USPS Household Mail Survey, FY 2025, Table 5.5)
• Very likely to respond when the piece carries a coupon: 5% (USPS Household Mail Survey, FY 2025, Table 5.6)

Cost per lead: one disclosed number, and one you compute yourself

Google publishes no lead price. Its own help page says lead prices vary by location, job type, lead type and bidding mode, which is a polite way of saying there is no rate to quote. The single digital figure in this guide with a disclosed method comes from a vendor aggregating its own book: $53 per lead and $233 per paying customer on Google Local Services Ads, observed across 888 contractors, $6.72M spend, 126,650 leads (SearchLight Home Services LSA Benchmark, February 2026). That is observed spend from one set of advertisers, not an industry average, and it is worth exactly what its sample is worth.

LocalAd publishes no cost per lead of its own, because it has not measured one. What it publishes is the price. A category-exclusive slot on a 5,000-home card starts at $349, and every card carries a QR code whose scans land in your dashboard with a timestamp. Divide the price by the responses you counted and the number is yours, with a sample size you can see, from the first drop on.

The arithmetic that needs no benchmark

Run this before any drop. Take the price. Take your average job value and your gross margin on it. Divide. That is how many jobs the drop has to produce to wash its face, and you can judge whether that is plausible in your own market better than any report can. A $349 slot against a $400 job at half margin needs 2 jobs out of 5,000 homes. A plumber knows within five seconds whether that is a reasonable ask.

This is the only comparison that survives a skeptical reader, because nobody can argue you out of your own average ticket. It is also the reason we would rather publish a price than a response rate: the price is a fact about us, and the response rate would be a guess about you.

Where each channel actually wins

Search wins on speed and on intent. A campaign is live in a couple of hours, in front of people typing the problem into a box right now. If the phone has to ring this week, that is the channel and mail is not.

Mail wins on persistence and on the moment before the search. A card has to be picked up and thrown away to leave the house, and USPS measures 10% of advertising mail as deliberately set aside to deal with later (USPS Household Mail Survey, FY 2025, Figure 5.7). How long a card then sits on a fridge is not measured by anyone, and we are not going to invent an average for it. What you can say is that the card is in the kitchen when the water heater fails, and the ad auction is not.

Email wins on cost per send and loses on the list. The send is nearly free only because the list was paid for somewhere else, and if it was bought, you have a CAN-SPAM problem rather than a channel.

The portfolio answer is the boring one and it is right. Search for emergency intent, email for a list you already own, mail for the neighborhood you want to own. See the plumbing playbook, the roofing guide and the HVAC seasonal strategy for how each trade times its drops.