Three arguments about physical outreach, at length

One. Direct mail is an account strategy, not a channel

The instinct when a team buys sending software is to treat it as email with stamps: build a list, write one letter, press go. It fails, predictably, and the failure gets blamed on the channel.

It fails because the economics are inverted. Email costs nothing per send and everything in aggregate attention. Post costs real money per send and buys attention that email cannot: three minutes of it, over nine days, per the UK benchmarks. A channel with a high unit cost and a high attention return is not a volume channel. It is a targeting channel, and the only sane way to run it is to decide who deserves it before deciding what to send.

The practical version, borrowed from account based marketing and stripped of the jargon: sort your accounts into three tiers. The top tier, perhaps twenty accounts, gets something genuinely one to one, researched, referencing something specific to them. The middle tier, perhaps two hundred, gets a strong letter with real personalisation of the first line and nothing else bespoke. Everything below that does not get post at all, and pretending otherwise is how budgets get wasted.

The uncomfortable consequence is that if your target list is five thousand accounts and your average deal is four thousand pounds, post is probably not your channel. We would rather say that on our own website than take the money and watch it fail.

Two. Sequencing, and the thing nobody else has

Every outbound sequence template online assumes email and LinkedIn, because those are the channels where the tooling reports back. Post gets bolted on as step nine, sent blind, and followed up on a guess.

That is a waste of the one advantage post has over every other channel: a delivery event and, if you put a QR code on it, a scan event. A scan is not an open. An open is somebody's mail client loading an image. A scan is a human being holding a piece of paper, deciding to reach for their phone, and pointing it at your letter. In terms of intent it is closer to a demo request than to an email open, and almost nobody treats it that way.

So the sequence should branch on it rather than run on a calendar:

• Scanned. Call within the hour if you can, or the same day. Reference the letter directly, not obliquely. They already know who you are, which removes the hardest part of the call.
• Delivered, not scanned. Wait for the post to have landed, then email or call referencing the letter explicitly, because the letter is your permission to interrupt. Something physical on their desk beats nothing every time.
• Returned. Stop, fix the address, decide whether the account is worth a second attempt. A returned letter is data about your list, not a failure of the channel.

The reason so few teams run it this way is that most sending platforms hand you a print confirmation and nothing else. If there is no page you control at the end of the QR code, there is no branch to make, and the sequence goes back to being a calendar.

Three. Proving it worked, when the usual methods do not apply

Direct mail attribution content online is almost entirely about retail: matchback files, holdout groups, control cells, hundreds of thousands of pieces. Every one of those methods needs volume to produce significance, and B2B post is deliberately low volume. Run a holdout on two hundred accounts and you have learned nothing except that statistics needs bigger numbers.

What works instead, in rough order of how convincing a finance director finds it:

1. A tracked page per recipient. The QR code goes to a page built for that person. A scan is timestamped, attributable to one account, and requires no cookie, no form and no consent banner. This is the closest thing to a clean signal that exists in this channel.
2. Meetings booked within a window. Count meetings from targeted accounts in the fortnight after delivery, against the same accounts' baseline. Not causal proof, but honest, and directionally hard to argue with.
3. Pipeline influenced, stated as influence. Log the send as a touch on the account and report it as influence rather than source. Overclaiming here is what gets a channel killed at the next budget review, not underclaiming.
4. The cost per meeting comparison. All in cost of a send, divided by meetings it produced, against the same figure for your other channels. It is the only number most executives actually want.

The trap to avoid is attributing revenue to the letter. You did not close the deal with a letter. You bought a conversation that would not otherwise have happened, and that is a defensible claim you can make repeatedly without anybody rolling their eyes.

Four. Getting it opened

All of the above assumes the envelope survives the building. That is a subject of its own, written up as what happens to your letter between the postbox and the desk, and it is the cheapest of all these improvements: a real stamp, a named person and a job title, a plain envelope, nothing showing through a window.