This page applies Steve Blank’s customer-development method to House-works. The idea is simple: a business plan is a stack of untested assumptions. So we write them down — organised across the nine blocks of the Business Model Canvas — then leave the building to replace each guess with evidence, until every assumption is validated, invalidated, or iterated into a better one.
The left column holds the assumption. The right column holds the proof — what we have found, or what would settle it. It is deliberately a working document: most cells start Untested, and the value is in watching them change.
Status key: Validated · Testing · Untested · Iterating · Invalidated
1 · Customer segments
| Assumption | Proof point / evidence | Status |
|---|
| Homeowners accumulate a real property record — surveys, certificates, warranties, invoices, project notes — but keep it scattered across paper folders and loose PDFs. | Untested. Proven by: problem interviews in which owners describe the scattered-folder pain unprompted. | Untested |
| There is a reachable early-adopter segment: owners of improvement-heavy or higher-value homes who feel the disorganisation most acutely. | Untested. Proven by: identifying and interviewing a cohort who have done recent works and actively wish they had kept better records. | Untested |
| The addressable market of record-caring homeowners is large enough to build a venture on. | Untested — being sized separately. Proven by: the market-sizing and segmentation work quantifying a reachable, willing segment. | Testing |
2 · Value propositions
| Assumption | Proof point / evidence | Status |
|---|
| At the point of sale, most of a property's work-history is lost — a felt pain for sellers (value they cannot evidence) and buyers (provenance they cannot see). | Untested. Proven by: sellers/buyers citing missing history as a real friction in a recent transaction; agents confirming it recurs. | Untested |
| Homeowners value a structured lifecycle record enough to adopt it — turning a document pile into a searchable record is a compelling first value moment. | Untested — the core hypothesis. Proven by: users completing the upload-and-organise flow and returning to it unprompted. | Testing |
| The 3D building model / digital twin adds meaningful value beyond a simple document store. | Untested and at risk. The brief warns House-works is 'not a feature list'. Validate only after the core record value is confirmed, to avoid over-building the differentiator first. | Iterating |
3 · Channels
| Assumption | Proof point / evidence | Status |
|---|
| House-works can reach homeowners cost-effectively — directly, and via referral from transaction intermediaries (estate agents, conveyancers, surveyors). | Untested. No channel tests run. Proven by: a channel producing sign-ups at a cost below the value of a customer. | Untested |
4 · Customer relationships
| Assumption | Proof point / evidence | Status |
|---|
| Homeowners will keep the record current across the long, infrequent gaps between property works (get / keep / grow). | At risk by design — the brief itself notes people 'do works infrequently' and 'are not good at record keeping'. Proven by: a cohort still updating records 6+ months after onboarding. | Untested |
5 · Revenue streams
| Assumption | Proof point / evidence | Status |
|---|
| Someone in the chain will pay — the homeowner directly, or an intermediary on their behalf. | Untested. No pricing tests run. Proven by: a willingness-to-pay signal — a letter of intent, pre-order, or paid pilot. | Untested |
| The value recovered at sale — recognising work 'hidden behind the walls' — is large and legible enough to justify the price. | Untested. Proven by: evidence that documented works measurably lift sale price or speed, enough to anchor pricing against. | Untested |
6 · Key resources
| Assumption | Proof point / evidence | Status |
|---|
| House-works' defensibility rests on a proprietary, structured data model — the lifecycle plus BIM-ready building schema — that is hard for a flat document store to replicate. | Untested. Proven by: evidence the structured model produces value competitors' flat stores cannot — i.e. it is a moat, not a commodity. | Untested |
| Authoritative property data — Land Registry, EPC, planning, warranty registries — is accessible on workable terms to enrich the record. | Untested. Proven by: confirming the key data sources are obtainable at acceptable cost and licensing. | Untested |
7 · Key activities
| Assumption | Proof point / evidence | Status |
|---|
| The make-or-break activity is reliable document ingestion and automatic categorisation by lifecycle phase — accurate enough that users trust it rather than re-filing by hand. | Untested. Proven by: ingestion and auto-categorisation accuracy high enough that users stop correcting it. | Untested |
8 · Key partnerships
| Assumption | Proof point / evidence | Status |
|---|
| Intermediaries and data holders (agents, conveyancers, insurers, mortgage/utility providers, warranty registries) will integrate or co-build — feeding data in and embedding House-works out — rather than treat it as a threat. | Untested. No partner conversations logged. Proven by: one signed data-sharing or embedded-distribution agreement. | Untested |
9 · Cost structure
| Assumption | Proof point / evidence | Status |
|---|
| House-works can be delivered at a cost — document storage, 3D rendering, ingestion — low enough to leave margin at a price homeowners or their intermediaries will accept. | Untested. Proven by: a unit-economics model showing positive gross margin at the target price point. | Untested |