Chapter 04

The self-demonstrating channel

The demo is the product. Do not describe Housewarmly to an installer. Generate a pack for a house that has just sold on their patch, put their logo on it, and post it to them. It costs about two pounds and it is the entire go-to-market.

The move, in one paragraph

Pick a target installer. Find a live Rightmove listing within four miles of their showroom, with a tired kitchen and a floorplan. Build the pack in their name, with their logo and their phone number, exactly as their customer would receive it. Post it to the installer, with a covering note that says: this went out to nobody. It is a pack for a house down the road from you, built in your name in eleven minutes, and it cost us two pounds. If you want the real one to go to the real address, you know where we are.

4.1 Why this beats every other first touch

ChannelWhat the installer receivesWhat it costsWhy it loses to the pack
Cold emailA paragraph about softwareFree, plus a privacy notice obligationThey are on site. They do not read email on a phone in a van.
Cold callAn interruptionYour timeGatekeeper, or voicemail, or "send me something".
A demo videoA generic kitchen they have never seenFreeEvery AI render tool in chapter 02 has one, and they all look the same.
A free trialAn empty account and a job to doFree, plus a churn problemThey have to do the work to find out whether it is any good.
A pack, in their name, of a house on their patchThe finished product, addressed to them, that they did not have to makeAbout £2S1It answers "what is it", "is it any good" and "would it work for me" simultaneously, on paper, with no login.

The argument in one line: every other channel asks the installer to imagine the product. This one hands it to them. And because the marginal cost of a pack is compute and a stamp rather than a person's time, the demo can be given away to the entire target list before a single conversation happens.

4.2 What it costs to demonstrate to everybody

LineCostBasis
Renders, three styles63pmeasured at 20.9p a render, current settingsS1
Print and post, one item88ppublished Stannp A4 letter, Free plan, the honest low-volume priceS7
Listing fetch~5pderived from the fetch service's credit pricing
Covering note, printed as an insert5ppublished Stannp insert rateS7
Cost of one demo£1.61Call it £2 with the fourth style and the odd reprint
Cost of demonstrating to all 52 targets in chapter 05~£104Two pounds each, once
Cost of demonstrating to 500 installers~£1,000The whole first year's outbound budget
Sit with the second-last row. The complete, personalised, physical demonstration of this product to every single company on the target list costs about the same as one month of a Checkatrade Growth listing.S17 There is no other software category where that is true, and it is true here only because the product's output is the marketing.

4.3 The pack that goes to the installer is not the pack that goes to a homeowner

This is the part most likely to be got wrong. Three differences, all deliberate.

It is marked as a specimen

The demo pack must say, on the artefact, that it was not sent to the property. Anything else risks an installer believing we have already marketed in their name to a house they have not agreed to, which is both alarming and, if it were true, a genuine problem.

It carries a covering note, not a brochure

One side of A5, six sentences, written by a person. The pack does the selling. The note only explains why an unbranded envelope contains their own logo.

It goes to the named individual

The homeowner pack is addressed to "The Homeowner" for the legal reason in chapter 10. A business-to-business pack goes to a named director at a business address, which is a different and easier position.S40

The covering note, verbatim

Nobody else has seen this. Mr Darlison, The pack in this envelope is for 14 Oaklands Avenue, which sold three weeks ago about a mile from your showroom. It shows that house's actual kitchen, from the listing photographs, fitted out three ways in your name. It has not been sent to the house. It cost us two pounds to make and it took eleven minutes. If you would like the real one to go to the real address, and to the next forty that sell near you, the website is on the back. If not, keep the pack. It is yours either way. Simon
Why that note works. It is six sentences. It names a real street they will recognise. It gives the cost and the time, which is the only claim in it and the only claim we can prove. It removes the alarm ("it has not been sent"). It makes no promise about response rates, because we do not have one. And it ends with a genuine offer to walk away, which is what makes a stranger read to the bottom.

4.4 Choosing the listing, which is where the demo is won or lost

A demo pack of a beautiful kitchen is worthless. The whole point is the gap between what the room is and what it could be.

1

Within four miles of their showroom

Close enough that they know the street. If they do not recognise the address, half the effect is gone. Use the address of the showroom from chapter 05, not the registered office, which is often an accountant's.

2

Sold, or under offer, in the last month

A live listing works and is easier to find. Sold subject to contract is better, because the trigger story is true.

3

A genuinely dated kitchen or bathroom, and a floorplan

The floorplan is a required input, and without it the measured plan page and the fit engine have nothing to work from.S6 Reject the listing rather than send a pack with no dimensions in it.

4

In their price bracket, not above it

Send a bathroom fitter in Ilkeston a pack for a £900,000 barn conversion and you have shown them a customer they do not have. Match the house to the work they photograph on their own site.

5

Their trade, and only their trade

A company is a kitchen installer or a bathroom installer in the product, never both.S5 Pick the one their website leads with. Getting this wrong is the single most obvious error a recipient can spot.

4.5 Getting their brand right without asking them

The demo has to look like theirs, and we are building it before they have given us anything. Three inputs, all from their own public website.

What we needWhere it comes fromThe trap
LogoTheir website header, at the largest resolution servedA logo lifted at 200px and printed at 300dpi looks like a forgery. If the only logo available is small and lossy, set it in type instead and say so in the note.
Brand colourSampled from their own siteSampling a stock photograph rather than their palette. Take it from a heading or a button.
Phone, email, addressTheir contact pageRegistered office instead of showroom. Chapter 06 flags every row where these differ.
Trade and stylesTheir portfolio pagesSending Shaker to a firm whose entire gallery is handleless matt. Spend two minutes looking.
The one thing not to do. Do not put their logo on a pack and then send that pack to a member of the public. A demonstration addressed to the installer is our marketing about their business. A pack addressed to a homeowner in their name, without their agreement, is us trading as them. The first is a good idea and the second is a serious one.

4.6 The sequence around the pack

The pack is the first touch, not the only one. Five steps over eighteen days.

4.7 Making the channel a machine rather than a stunt

Doing this fifty times by hand is a fortnight of work. Doing it five hundred times by hand is not possible. Three things turn it into an engine, in the order they are worth building.

1

A prospect brand kit that fills itself in

Paste an installer's website URL, and the system pulls the logo, the colour, the phone number and the trade, and creates a company row in a "prospect" state that can generate packs but cannot log in. Half a day of work and it is the highest-return half day in the plan.

Not built · chapter 08 week 3

2

A listing picker for a postcode

Given a showroom postcode, return recent sold-subject-to-contract listings within four miles that have a floorplan and a photographed kitchen. Today this is a person on Rightmove for four minutes per target. Read the terms-of-use position in chapter 10 before automating it, because this is exactly the activity the portals prohibit.

Not built · do the manual version first, deliberately

3

A demo pack template

The specimen watermark, the covering note as an insert, and a variant of the back page that addresses the installer rather than the homeowner. Two hours.

Not built · chapter 08 week 2

4.8 What we do not know, and how to find out cheaply

Nobody has run this campaign. There is no response rate for it, and there is no benchmark that fits. JICMAIL puts UK cold direct mail at about 0.9% and warm at 7.2%.S15 A pack showing the recipient their own trade, their own logo and a street they know is neither. It is also going to a business rather than a household, which is a different population from anything JICMAIL measures. Any number written down before the first fifty go out is invention.
Do not quote 4.4%. The figure that circulates as "the DMA direct mail response rate" is American, from the US ANA report. DMA UK publishes no direct mail response rate report at all. The UK cold benchmark is roughly five times lower.S16 A business case built on 4.4% is built on the wrong continent and the error is trivially checkable by anyone who wants to embarrass us.

The measurement that costs nothing

Fifty packs is £104 and it produces the only number that matters. Instrument it properly the first time, because there will not be a clean second chance with the same list.

RecordHowWhy
Which target got which listingA row per demo, before it is sentSo a reply can be traced to the pack that caused it, and so a good listing choice can be repeated
Whether the share page was openedThe QR code already resolves to a tokenised share URLS5The cheapest possible signal of interest, and it fires before anybody rings
Reply, by typeInterested, not now, wrong fit, no reply, angry"Angry" is a real category and its rate is the single most important safety metric in this channel
Days from send to first responseTimestampsSets the follow-up cadence for the next fifty. Guessing day 5 is a guess.
The bar to clear. Fifty packs, £104, and one paying Pro subscriber at £89 a month pays it back inside six weeks. Two say the channel works. Five and it is the only thing worth doing until it stops working.

4.9 The risks in this channel, stated before somebody else states them