Most DTC brands have a retention problem they have misdiagnosed as an email problem. The list has thirty or forty thousand people on it. Maybe eight thousand still open. The rest are sitting there as a line item in Klaviyo, quietly dragging down deliverability, getting suppressed, and eventually getting written off as churn. They are not churn. They are customers you have lost the ability to speak to.
Direct mail is how you speak to them. Not the 2009 version with a bulk list broker and a six-week lead time. The version that plugs into Shopify, triggers off a customer event, and posts a personalised card three days later. This is the breakdown of what the channel actually returns in 2026, which segments to mail, what it costs, how to measure it properly, and the four mistakes that make a working channel look broken.
Why The Numbers Favour Mail Right Now
The case for direct mail is a case about inventory, not nostalgia. Every DTC brand in your category is bidding for the same feed placements against the same CPM inflation. Almost none of them are posting anything. The competitive set in the letterbox is a utility bill and a takeaway menu.
That scarcity shows up in the response data. Direct mail to an in-house list averages around a 161 percent return, the strongest of any paid channel in the ANA benchmark set. Postcards specifically average a 5.7 percent response rate, with letter-sized envelopes at 4.3 percent, and that format gap holds consistently across campaigns and industries. Set that against 1 to 3 percent for email and 0.5 to 1.5 percent for Meta paid and the picture is not close.
Win-back is where it gets interesting. Cards sent to customers who last purchased 180 to 365 days ago have returned around 10x ROAS in reported ecommerce campaigns, and even cohorts last active three years ago have cleared 2x. One supplement brand reported a 1,052 percent return on a win-back campaign aimed at repeat purchasers past the 90 day mark. Those are not cold prospects. Those are people who already bought, already liked it, and simply stopped hearing from you.
Operators who dismiss direct mail as too expensive are comparing cost per impression against a channel that costs nothing to send. The only comparison that pays your wages is cost per acquired order.
The maths that changes the argument
A programmatic postcard costs somewhere between 60p and 1.20 GBP all in. An email costs a fraction of a penny. On a cost per impression basis mail loses by a factor of several hundred and the conversation ends there for most brands. Run it on CPA instead. Mail 2,000 lapsed repeat buyers at 90p a card and you have spent 1,800 GBP. A 5 percent response at a 55 GBP average order value returns 5,500 GBP in revenue from a segment your email programme had already given up on. The unit cost is irrelevant. The unit economics are the whole argument.
The Four Segments Worth Mailing
Direct mail is a retention channel with an acquisition side effect. Mail your own data first. Rented lists and neighbourhood targeting are a different, harder game and not where a DTC brand should start. Work down these four in order.
Lapsed Repeat Buyers (90 to 365 Days)
Orders >= 2 AND days since last order between 90 and 365
This is the highest-return segment in the channel and it is not close. These customers have proven product fit, proven willingness to reorder, and a purchase history that tells you exactly which product to put on the card. They have simply drifted out of your email engagement window.
For a consumable brand, size the window off your actual replenishment cycle rather than a round number. If the average reorder gap is 62 days, mail at day 85 to 95, not day 180. You are trying to catch them before they habituate to a competitor, not after.
Put the specific product they bought on the card. Not the range, not the bestseller. Personalisation lifts response for 88 percent of marketers who test it, and on a physical card the personalisation that matters is product recall, not their first name in a headline.
Unsubscribed And Unengaged Email Contacts
Email suppressed OR zero opens in 180 days AND >= 1 order
This is the segment that makes finance sit up. These people are unreachable by every other owned channel you have. They cost you nothing to hold and return nothing. Mail is the only door left.
The compliance position is different from email. An unsubscribe from marketing email is a preference against that channel, not a blanket instruction, though you should honour any explicit request to stop all contact and check your position under UK GDPR and PECR before you scale. Legitimate interest for postal marketing to existing customers is a well-trodden path, but get your own advice rather than mine.
Expect a lower response rate here than segment one and judge it on incremental revenue against a holdout, because the baseline you are comparing against is close to zero.
High-Value Cart And Checkout Abandoners
Cart value > 2x AOV AND no purchase within 72 hours AND known address
You only have a postal address for people who have ordered before, so this segment is smaller than it sounds and only works for returning customers who abandoned. That constraint is also what makes it profitable, because you are mailing high-intent people you already know.
Set a value floor. Mailing a 25 GBP abandoned cart at 90p a card is a rounding error on margin. Mailing a 180 GBP one is straightforward maths. Anchor the floor at roughly twice your AOV and revisit it monthly.
Accept the lag. The card lands three to five days after the abandon, which means it is not competing with your abandoned cart email. It is the follow-up nobody else sends in week two.
Subscription Cancellers And Churned Members
Subscription cancelled 30 to 120 days ago AND no reactivation
Cancellation flows are almost always email-only, which means the entire save attempt happens in the inbox that person was probably ignoring in the first place. A card arriving three weeks after cancellation catches them at the point the product actually runs out.
The offer here should be about friction, not discount. Reinstate at the price they had, skip the first delivery charge, let them set the interval themselves. Discounting a churned subscriber teaches them to churn again.
Track this cohort separately from one-time buyers. Subscription reactivation carries a far higher lifetime value than a single win-back order, so it will justify a higher cost per piece and a premium format.
How To Run It Programmatically
The operational shift that made this channel viable again is programmatic triggering. Instead of exporting a list, briefing a printer, and waiting six weeks, a platform sits on top of Shopify and Klaviyo, watches for a data condition, and prints and posts a card when a customer crosses it. PostPilot and Birdseye are the two most DTC brands land on. Functionally it behaves like a Klaviyo flow with a three to five day delivery delay baked in.
Build it as flows, not campaigns. A one-off blast to your whole list is a bad first test because you cannot separate the effect of the audience from the effect of the offer. Two or three evergreen triggers running continuously will teach you more in six weeks than four blasts will in six months, and they keep working once you stop paying attention.
Suppress aggressively. Anyone who has ordered in the last 30 days, anyone already inside an active email win-back sequence, anyone with a pending return or open support ticket. The fastest way to lose money on this channel is posting cards to people who were about to buy anyway, which is the same mistake brands make with retargeting ads and then blame on the platform.
On format, postcards beat envelopes on response and cost, so start there. Save the higher-cost formats for the subscription reactivation cohort where the lifetime value justifies it.
The Four Mistakes That Burn The Budget
First, measuring on code redemption alone. Most people who respond to a card type your domain straight into the browser and never touch the code. If you treat code redemptions as total attribution you will undercount by a wide margin and kill a profitable programme. Run a holdout instead: randomly exclude 10 to 20 percent of the target segment, compare revenue per customer across a 30 to 45 day window, and call the difference your incremental lift.
Second, calling the result too early. Email spikes within 48 hours. Mail response is slower and flatter, spread across weeks. Brands that judge at day seven consistently conclude the channel failed while most of the revenue was still in the post.
Third, mailing dirty addresses. Shopify address data is entered by customers at checkout and decays at roughly 10 percent a year through moves alone. Run address validation before every send. Every undeliverable card is your full unit cost with zero chance of return, and it is the single easiest cost to eliminate.
Fourth, putting a brand advert on the card. A postcard has one job and about three seconds to do it. One product, one reason to come back now, one offer, one URL. The cards that fail are the ones that look like a mini catalogue.
What This Looks Like In Practice
A wellness brand I worked with had 34,000 people on the list and 6,800 of them engaged. The other 27,000 were treated as dead weight, suppressed for deliverability, and never spoken to again. Their entire retention strategy was a welcome flow and a replenishment reminder, both aimed at the 6,800.
We cut a test segment of 2,400 lapsed repeat buyers between 120 and 400 days since last order, held back 15 percent as a control, and posted a single card featuring the exact product each customer had bought, with a reinstate-your-price offer rather than a discount. Delivery took four days. We measured at day 42.
The mailed group generated meaningfully more revenue per customer than the holdout, and the incremental margin covered the send several times over. The part that mattered more was structural: a segment that had contributed nothing for a year became a repeatable monthly line item, and roughly a third of the responders re-entered the email programme by opting back in at checkout. Direct mail did not just recover orders. It rebuilt permission.
Inside the system
How we build this for brands
The hard part of direct mail is not the card. It is knowing which 2,000 people out of 34,000 are worth 1,800 GBP of postage this month, and having the copy on the card sound like the brand rather than a printer template. We run that with lifecycle flows built and deployed in Klaviyo by AI, sitting alongside a VOC engine that mines real customer reviews and support messages so the reason-to-return on the card comes from the customer's own language, not a copywriter's guess. The segment logic is rebuilt against live Shopify and ad data by a reporting agent that also flags when a cohort's response rate starts drifting.
On top of that we run email-monitoring agents that watch what happens after the card lands, so a customer who replies or comes back gets a genuinely personalised follow-up rather than dropping into a generic post-purchase sequence. For the highest-value reactivated customers we go further and invite them to something real, a small tasting or a run club rather than another discount. Part of this runs live for portfolio brands today; the full system is what we deploy when we take a brand on.
Retention Audit
Find Out What Your Lapsed List Is Actually Worth
I will size your unengaged and lapsed segments against real replenishment behaviour, model what a mail programme would return against a holdout, and tell you plainly whether it is worth running for your margin structure. Numbers first, then the plan.
Book Your Retention AuditFrequently asked questions
Does direct mail still work for ecommerce brands in 2026?
Yes, and the numbers are better than most operators expect. Direct mail to a house list averages around a 161 percent return, the highest of any paid channel in the ANA benchmark data. Postcards average a 5.7 percent response rate against 1 to 3 percent for email and 0.5 to 1.5 percent for Meta. The reason is inventory. Every brand competes for the same feed impressions, and almost nobody competes for the letterbox. It works best as a retention and win-back channel against your own customer list, not as cold acquisition.
How much does direct mail cost per piece for a DTC brand?
A programmatic postcard typically lands between 60p and 1.20 GBP per piece including print and postage, depending on volume, format, and country. That is 50 to 500 times the cost of an email. The comparison that matters is cost per acquisition, not cost per impression. If a postcard converts at five to ten times the rate of an email to the same lapsed segment, the CPA can come out lower even at 100 times the unit cost. Model it on a 500 to 1,000 piece test first.
What is programmatic direct mail and how does it work with Shopify?
Programmatic direct mail triggers a physical card off a data event rather than a manual campaign upload. It connects to Shopify and Klaviyo, watches for a condition such as 120 days since last order or a cart abandoned above a value threshold, and prints and posts a personalised card automatically. Platforms including PostPilot and Birdseye run this. Operationally it behaves like a Klaviyo flow with a three to five day delivery lag instead of an instant send.
Which customer segments should DTC brands mail first?
Start with lapsed repeat buyers, meaning customers with two or more orders whose last purchase was 90 to 365 days ago. This segment has proven intent, known product fit, and no active email engagement, which is exactly where mail outperforms. Reported win-back ROAS in the 180 to 365 day window sits around 10x, and even three-year lapsed cohorts have returned 2x or better. Mail unsubscribed and unengaged email contacts second, because they are unreachable by any other owned channel.
How do you measure direct mail attribution accurately?
Use a holdout group. Randomly exclude 10 to 20 percent of the target segment from the mailing, then compare revenue per customer between the mailed group and the holdout over a 30 to 45 day window. The difference is your true incremental lift. Unique codes and vanity URLs undercount badly because most recipients type your domain into the browser instead, so treating code redemptions as total attribution will make a profitable programme look broken.
How long does it take to see results from a direct mail test?
Allow three to five days for delivery and then a 30 to 45 day response window. Direct mail response is slower and flatter than email, which spikes within 48 hours. Brands that call a result at day seven consistently conclude the channel failed when the majority of the revenue had not landed yet. Budget for one full test cycle of roughly six weeks before deciding whether to scale.
Can you mail customers who unsubscribed from your emails?
An email unsubscribe is a preference against that specific channel rather than a blanket instruction, and postal marketing to existing customers under legitimate interest is a well established practice. That said, you must honour any explicit request to stop all contact, keep a postal suppression list, and check your own position under UK GDPR and PECR before scaling. Take proper advice rather than relying on a blog post.
About the author
Caner Veli is a DTC operator who has helped 350+ brands fix broken growth engines. He built Liquiproof from zero to 3,000+ global retailers in under 6 years. He now runs the same playbook, supported by AI systems he built himself, for DTC and CPG brands.