A bought list
Every company matching an industry code and a headcount band. Sorted by nothing in particular.
Physical direct mail, B2B
We post paper letters to people who stopped reading email years ago. Inside each envelope there is an object they have to pick up with their hands, addressed to a named person at a street address we have verified, and delivered against a signature. Sometimes they reply.
You are buying a campaign. You are also watching it being thought through: which market, which companies, which people, which object, which sentence. Four stages, and you see the reasoning at each one, not only the result.
The record
1,000,000+
We have printed, packed and posted more than one million pieces of physical mail. Not a million impressions. A million envelopes that a courier or a postal service physically carried, most of them across a border.
That number proves two things and nothing else. It proves the production side is not a pilot. And it proves we have already made every mistake available at that stage: paper that curls in a sorting machine, an enclosure that rattles loudly enough to look suspicious, a customs declaration written by someone who had never seen one.
It does not prove your campaign will work. Volume is not a result. What decides the result happens before the print run — the market, the list, and the first sentence on the page — and those are the three parts we show you while they are still being decided.
letters posted ≠ replies earned
The market
Most direct mail fails before anyone writes it. It goes to a list somebody bought, filtered by industry code and headcount — a description of what a company is registered as, not of what it needs this quarter.
We start one step earlier. Before a list exists we work out which markets have a reason to answer you now, and which ones would read your letter politely and do nothing. This is the part clients understand least and pay for most, so it has a page of its own.
What you get is not a verdict. You see the hypotheses we began with, the ones that survived contact with evidence, and the ones we dropped, each with the reason it was dropped. A market we talk you out of is a result. Losing one on paper is cheaper than losing it in the post.
Every company matching an industry code and a headcount band. Sorted by nothing in particular.
The companies with a reason to answer you this quarter, each one carrying a note that says what the reason is.
The long version
How a market is read, how the hypotheses are named, what gets pruned and by what standard, and what we measured afterwards to find out whether any of it worked. It is long, and it is the honest version rather than the flattering one.
Read Market Scan Or start with what it cannot doCompanies and routes
A market is not a mailing list. Stage two turns one into the other, and most of it is done by hand.
For each company we find the person who owns the problem rather than the person whose title sounds nearest to it. Then we find where that person's post physically lands. A registered office is not an answer. It is often an accountancy firm two streets away who will bin your envelope without opening it, and they will do it efficiently.
Every address is confirmed against at least two independent sources. Where the sources disagree we either resolve it or delete the record. Records do get deleted, and the list you receive is shorter than the list we started with. A letter that arrives at the wrong desk costs more than a short list does, because printing was never the expensive part.
Company, and the reason it survived stage one
The recipient, by name, and their part in the decision rather than their job title
Street address, with floor or unit where the building needs one
How the address was confirmed, and on what date
Delivery method, and whether a signature is available on that route
Anything we could not confirm, labelled as unconfirmed
Some people cannot be reached this way at all. Anyone whose post is opened by a facilities team before they see it, anyone at an address that exists only as a post box we cannot tie to a floor, and anyone who has moved recently enough that no source agrees on where they sit. Where a name cannot be routed we tell you before production, not after.
The letter
An envelope with something solid in it does not end up in the pile with the invoices. It gets opened, because nobody can tell what is inside without opening it, and because a shape in an envelope is a small unanswered question.
That is the entire trick of the object, and it lasts about ten seconds. After that the letter is on its own.
So the letter opens by naming something true and specific about the recipient's operation — a sentence that could not have been posted to anyone else. Not a greeting, not a summary of their industry, not a paragraph about us. If the first sentence would survive a find-and-replace of the company name, it is the wrong first sentence and it gets rewritten.
One page. One ask. The ask is small enough to answer on the day it arrives: a reply, a short call, a sample sent back. We do not ask a stranger for an hour of their week.
The object is chosen for the argument, not for the budget. The expensive one is often the wrong one. We have talked clients out of the object they liked most, because it made the letter about us, and a letter about us is a letter about nothing.
the object buys ten seconds ≠ the letter earns the reply
Production and delivery
Production is the least interesting stage and the easiest to ruin. Paper behaves differently in a sorting machine than it does on a desk. An enclosure that rattles reads as a sample nobody ordered. A customs declaration filled in casually adds days at a border, and sometimes a charge the recipient is asked to pay, which is a poor way to introduce yourself.
We print, pack and dispatch, and you get the same manifest we work from.
Print run, paper stock, and the proof you signed off
Object per recipient, packed and weighed
Envelope, addressing, and any customs paperwork
Dispatch date and route, per record
Tracking reference wherever the route provides one
Delivery comes back as evidence rather than as a summary. Every record ends in one of four states, and the states are counted as they are. It is not done when it leaves us. It is done when delivery is recorded.
Delivery today runs through a service that records receipt by the named recipient. That works, and it is what the paragraph above describes. It is not what we want.
Three things we are building. None of them exist as you read this.
When any of this becomes true, this section gets shorter. Until then it stays here, listing what we owe.
None of this tells you whether anybody read it. A signature can belong to a post room. We report delivery, replies, and the routes the replies came back through, and we have deliberately not built a number that claims to measure attention.
Every record keeps the reason it was chosen. When a reply arrives it is still attached to the person, the address, the object and the hypothesis from stage one that put that company on the list. That is what makes a second campaign cheaper than the first: you are not starting from a blank market, you are starting from the part of it that answered.
Four stages, and none of them can be skipped by paying more for the last one.
The objects
Every campaign carries one object. It is not a gift and it is not merchandise. It is the reason the envelope gets opened, and the reason the letter is still on the desk on Thursday.
These are the ten we produce now. The right one is the one that carries your argument — the one a recipient can hold up and explain to a colleague without reading the letter twice.
For letters about money, profit, savings, cost of error, investments, surveys, invitations, and the first conversation.
Money in the hand works harder than the word “money” on a page.
A coin-shaped token sets the subject immediately: this is about real money, not an abstraction. A printed paper token can honestly buy the first two minutes of attention—and that is fair if a strong offer follows.
An old-looking token or unusual coin-shaped piece works especially well when the subject is inflation, devaluation, financial memory, rarity, a nonstandard solution, or the cost of delay.
For proof: a property, warehouse, production line, product, empty shelf, document, person, or real condition.
A photograph is useful where a claim would be easy to dismiss.
It says: this exists. This happened. Not a mockup. Not a promise. Not a polished story from a presentation.
For a complex product, a general view is usually weaker than one precise detail: a seam, connection, edge, material, package, finish, defect, or result.
For land, beach, resort, construction, plot, warehouse, industrial zone, infrastructure, and any projects where the main thing is place.
The letter puts a small part of the location into the recipient’s hands.
The plot stops being a line in a PDF. The beach stops being a picture. The construction site stops being a promise.
Road access, neighbours, footfall, route, future development, and land value stop being abstractions. They become physical context.
For letters where an answer is needed: qualification, survey, signal of interest, application, consent to a conversation, or a reason for a second touch.
A prepaid reply removes one excuse. The recipient does not have to work out how to answer.
A questionnaire or check-style form works even harder: there is an action to complete. It no longer feels like ordinary paper.
When a person sends back a form, questionnaire, or short answer, you get a clean signal: they noticed, understood, and acted.
And they will almost certainly remember who asked the question.
For letters that begin with a specific pain: expensive leads, an empty sales pipeline, debt, overload, delay, advertising that burns budget, production errors, or lost enquiries.
The blister-shaped card does not solve the problem. But it names the pain immediately.
Before the first paragraph, the recipient understands: this is about my headache.
A stress card makes the same move differently. It turns pressure into a small physical test.
For contrast between the old and new way. Or for the idea of hidden value.
A feather can stand for the quill—once normal, now absurdly slow.
Coal shows raw material before transformation. It looks rough, black, and boring. But there is value inside it.
Such an object works well in letters about transition, renewal, modernisation, revaluation of old assets, and opportunities that still look unattractive.
For access, a closed opportunity, real estate, negotiations, entry to the right person, a private offer, or a limited circle of participants.
A key does not open anything by itself.
But it instantly sets the meaning: there is a door, there is access, there is something not everyone is allowed into.
In a letter, the key says: this is not a mass mailing. It is an invitation through a specific door.
Mr Adeyemi —
There is a key taped above this line. It opens nothing.
I had it cut to a blank that matches no lock anywhere, which cost about a pound and took a morning. I did that because a letter from a stranger has roughly four seconds to survive the walk from your post room to your bin, and an envelope with a key in it does not feel like paper. It feels like an object, and objects get opened.
That is the whole trick. I would rather tell you than have you work it out on your own.
For manufacturing, construction, finishing, packaging, furniture, equipment, locations, routes, and logistics.
Material is better touched than described.
You can write “dense,” “expensive,” “reliable,” “pleasant to the touch” as much as you want. One real sample settles the question faster.
A map fragment works when the main thing is place, movement, route, delivery radius, footfall, proximity to an object, or point of entry.
The map forces the recipient to see the situation.
For private meetings, presentations, clubs, access, a premium offer, a choice, or a personal reply.
A ticket turns an invitation into an object.
A plastic card gives a feeling of belonging even before activation. The person is not holding a description. They are holding an almost-ready pass.
A blank card shows free space: for a name, decision, signature, choice, status.
Sometimes blank space asks for a decision more strongly than another paragraph.
For letters about repair, replacement, insurance, quality, production, safety, errors, and the cost of stoppage.
A broken thing can make risk more believable than a polished image.
The opening can be brutally simple: “I included this broken part for a reason. Big expenses start with small things like this.”
First a crack. Then a stoppage. Then urgent repair. Then people explain why no one noticed the problem earlier.
In letters about quality, safety, and losses, a broken part does the main thing: it removes abstraction.
The error is no longer in the report. The error lies on the table.
If none of these carries your argument, we make something else. That is a conversation for the brief, not for the print queue.
We would rather post four hundred letters that land than four thousand that arrive.
How we work together
01
You tell us what you sell and who you think ought to answer. We come back with the markets worth writing to, the ones we would skip, and the reason for each. Some briefs end at this step. That is a cheap place for a brief to end.
02
You approve the list, the letter and the object, and you approve them separately. Approving the letter means reading it word by word: the first sentence decides more than the paper does.
03
We print, pack and post. You get delivery evidence for every record, and the replies as they come back, still attached to the reason that record was chosen.
We do not start a print run to protect a date in a calendar. If approval slips, dispatch slips with it.
The obvious questions
There is no price on this page because there is no self-serve product behind it. Every campaign is quoted on its own, and the quote moves on three things: how many people you want to reach, which object goes in the envelope, and which routes the post has to travel. We quote before any work starts, and the quote does not change afterwards unless you change the brief.
Deciding who is worth writing to takes as long as that market takes to read. Confirming where those people physically sit takes longer than anyone expects it to. Production and dispatch are scheduled after you have approved the list, the letter and the object.
We write them. You approve every word before anything is printed, and you can rewrite any of it. What we will argue about is the first sentence, because a first sentence that could have been posted to anybody else has already lost the reader.
Yes, after we check it. Expect it to get shorter. Bought lists lose records at the verification stage, sometimes most of them. A short list you can trust costs less than a long one you cannot.
Then nobody replies, and we do not have a clause that pretends otherwise. What you are guaranteed is delivery evidence per record and a visible reason behind every choice that led to it, which means a campaign that fails tells you where it failed. A campaign that fails still narrows the market, which is worth something, though a good deal less than a reply.
Most of them, but not on the same terms. Some routes come back with a signature and a name, some come back with a delivery scan and nothing else, and a few we cannot confirm at all. The list says which route each record travels and what evidence that route can return.
Anyone who needs volume this week. Anyone whose product sells perfectly well over email, because then this is an expensive way to do something cheap. Anyone who cannot name the person who would say yes, because that is the whole job and we cannot invent them for you. And anyone who wants a price without a conversation.
Start here
Three lines is enough to begin: what you sell, who you think should answer, and what you have already tried. If we think direct mail is the wrong instrument for the job, you will hear that in the reply rather than in a proposal.