Chapter 09
Metrics and unit economics
Six numbers matter. Four of them are measured today. Two of them nobody in this market has, including every competitor, and getting them is what the first quarter is for.
What matters
- Measured today: cost per render, cost per pack, cost per posted pack, cost per demo.S1
- Unmeasured, and the whole business turns on them: the demo reply rate and the homeowner response rate.S15
- Gross margin is 59% to 68% at the shipped prices. Software people should recalibrate: there is a printing press in this business.
- Payback on a demo is one month if it converts to Pro. That single fact is why the channel in chapter 04 is the plan.
- The worry threshold that matters most is not churn. It is the share of packs that are made and never posted.
9.1 The measured numbers
20.9p
Per render at current settings, 13 renders on the most recent dayS1
96p
Mean render cost per pack across 18 packs, range 62p to 143pS1
£1.84
A pack made and posted at low volumeS7
£1.61
One demo, three styles, posted with a covering noteS7
Every one of those four carries a date. The render cost moved 25% in two days when the model settings changed.S1 Print prices change on a published schedule, with a Royal Mail revision on 5 October 2026.S8 A unit-economics figure quoted without a date is a figure that will be wrong within a fortnight, and the person quoting it will not know.
9.2 Unit economics per customer
Assumptions are labelled. Nothing in this table is a forecast.
| Line | Starter | Pro | Basis |
|---|---|---|---|
| Price per month | £29 | £89 | shippedS3 |
| Packs included | 5 | 20 | shipped |
| Cost if every pack is posted | £9.20 | £36.80 | derived from measured and published figures |
| Cost at a 50% post rate | £7.00 | £28.00 | assumption the post rate is unmeasured |
| Gross margin, all posted | 68% | 59% | derived |
| Gross margin at 50% posted | 76% | 69% | derived |
| Cost to acquire, one demo | £1.61 | £1.61 | derived, and it excludes every hour of your own time |
| Demos per customer won | unknown | unknown | The whole point of the first quarter |
| Months to pay back one demo, if it converts | under 1 | under 1 | derived |
| Months to pay back at 20 demos per win | 1.6 | 0.6 | assumption a 5% demo-to-paid rate, chosen as a round number, not measured |
| Lifetime value | unknown | unknown | No customer has been through a renewal. Do not compute an LTV yet. |
The line that matters. Even at a pessimistic twenty demos per customer won, acquisition costs about £32 and a Pro subscription pays it back inside a month. The economics of the channel are not the constraint. The response rate is.
Two things this table deliberately does not do. It does not compute a lifetime value, because no customer has reached a renewal and an LTV built on an assumed churn rate is a number that flatters whoever wrote it. And it does not count your own time, which at fifty demos is the largest real cost in the business. Both belong in the table the moment there is data for them.
9.3 The two numbers nobody has
The demo reply rate
What it is: of fifty packs posted to named installers, how many reply at all.
Why nobody has it: nobody has run this campaign, in this trade, with this artefact.
What we must not do: bracket it against JICMAIL. UK cold direct mail is about 0.9% and warm about 7.2%, but those measure households, not businesses, and neither describes a pack carrying the recipient's own logo.S15 Any number written before the first fifty go out is invention.
Cost to find out: about £110 and three weeks.
The homeowner response rate
What it is: of packs posted to houses that have just sold, how many produce a phone call to the installer.
Why it is harder: it needs a paying or comped installer, a run of real packs, and an installer willing to tell us honestly what came in. It also has a lag of weeks to months, because renovations start after the boxes are unpacked.S46
What is known: 44% of UK homebuyers renovate within the first year and 64% within two.S46 That is the size of the pool, not the response rate.
Cost to find out: 500 packs at £1.84 is about £920, and it takes a quarter.
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.S16 Using it inflates every model roughly fivefold and the error takes anyone ninety seconds to find.
9.4 The weekly review, on one page
Fifteen minutes, every Monday. Six numbers, four questions. If it takes longer than fifteen minutes it will not happen in week five.
| Number | Where it comes from | Worry when |
|---|---|---|
| Demos sent this week | The demo table | Fewer than ten during weeks 4 to 6. The channel only works if it runs. |
| Replies, by type | The demo table: interested, not now, wrong fit, no reply, angry | Any "angry" at all. One is a signal, two is a pattern, three means stop the batch. |
| Packs created | The packs table | Flat week on week while subscribers rise. It means the product is bought and not used. |
| Share of packs actually posted | Dispatch rows over ready packs | Below half. This is the most important number in the table and the least obvious. A pack that is made and never posted is a customer about to churn, and it costs us the render either way. |
| Failed packs, by step | The pack state machine's failure reasonS5 | Any repeat of the same failure reason. Portal parsers drift and the failure is explicit by design. |
| Render cost per pack, seven-day mean | The renders tableS1 | Above £1.20. It means the styles or the settings have crept. |
And the four questions
- Which listing produced the best reply this week, and why? Listing choice is the single biggest controllable variable in the demo channel.
- Did anybody ask about the photographs? If yes, what was said, and does chapter 10 need updating?
- What did a customer ask for that does not exist? One line. Do not build it. Just count how many times it comes up.
- Is the agency conversation further forward than last week? It is the gate on everything, and it is the item most easily displaced by busy work.S43
9.5 Vanity metrics, named so they can be ignored
| Metric | Why it flatters | Use instead |
|---|---|---|
| Sign-ups | A free trial with no pack made is nothing | Packs created in the first seven days |
| Packs created | A pack downloaded and never posted did not reach a homeowner | Packs posted |
| Share-page views | Mostly the installer checking their own work | Share-page views from a device that is not the installer's, if you can tell, and honestly say when you cannot |
| Renders generated | It goes up when quality goes down, because failures retry | Renders per ready pack |
| Total addressable market | There is no published census of UK installers, so any figure is oursS19 | Named firms in the target list with a verified source per row |
| Website traffic | The buyer is on site with a phone in a van | Packs posted |
9.6 What a good quarter looks like, in numbers
| At week 12 | Poor | On plan | Good |
|---|---|---|---|
| Demos sent | 20 | 50 | 50 |
| Replies of any kind | 1 | 6 | 15 |
| Discovery calls held | 0 | 4 | 10 |
| Paying subscribers | 0 | 2 | 5 |
| Comped design partners | 2 | 6 | 10 |
| Packs posted to real homes | 0 | 60 | 200 |
| Monthly recurring revenue | £0 | £118 | £385 |
| Agency or CRM conversations open | 0 | 1 | 2 |
| Supplier conversations open | 0 | 1 | 1 |
Read the revenue row honestly. "Good" at week 12 is £385 a month. This is a business that will take a year to be worth anything at £29 and £89 a month, and the chapter 02 ceiling says the whole UK pack market at £5 a pack is under £4m. The quarter is not being run to produce revenue. It is being run to produce two numbers and one signed data agreement, because those are what make the supplier conversation in chapter 07 possible, and that conversation is where the business actually is.
9.7 The three things that would change everything
- A signed data agreement with an agency group or CRM vendor. It lifts the ceiling in chapter 10 and it is the precondition for every large partnership.S43
- A measured homeowner response rate above about 1%. At 1% and a £17,500 median kitchen, one pack in a hundred producing an enquiry makes the arithmetic obvious to any installer, and to any supplier.S12 Below about 0.5% the channel is a curiosity.
- One supplier paying for packs on behalf of dealers. It changes who the customer is, removes the £5 ceiling, and turns print volume from a margin problem into a volume discount.S48