Building House Price Derry: from public data to a useful local product

Good digital products often begin with a very ordinary question. For homeowners in Derry, one of those questions is simple: what might my house be worth?

The information needed to give a useful starting point exists, but it is spread across official datasets, property types, reporting periods and local areas. Raw data alone does not answer the homeowner's question. It needs to be checked, explained and turned into an experience people can use without translating a spreadsheet.

That was the thinking behind House Price Derry, a local product we built to give homeowners a clear guide range based on official house-price data. It is free to use, makes its limits visible and does not trigger a sales call unless the homeowner asks for one.

Start with the decision, not the dataset

A common mistake in data projects is to begin with everything the dataset can show. That often produces a large dashboard with plenty of filters but no obvious next step.

House Price Derry starts with the user's decision instead. A homeowner wants an initial sense of value before deciding whether to research further, speak to a professional or do nothing for now. The product therefore asks only for the details needed to produce a useful guide, including postcode and property type, before showing a range and a confidence label.

The lesson applies well beyond property. A useful data product should help somebody make a decision, complete a task or understand an exception. If the user still has to download the result and rebuild it in a spreadsheet, the product has probably stopped too early.

Make the source visible

Trust matters whenever a product presents financial or operational information. House Price Derry starts from the UK House Price Index for Derry City and Strabane. The site names that source, explains the reporting period and links its method to the figures it presents.

That source trail is not a footnote added for compliance. It is part of the product. A visitor should be able to understand where the information came from, when it was updated and what the result can and cannot tell them.

The same rule should apply to an internal reporting tool, a customer portal or an AI-assisted workflow. Keep the source record, transformation rules and update date close to the output. A confident answer without provenance is much harder to trust.

Be honest about uncertainty

An online guide cannot see an extension, a recent renovation, the condition of a kitchen or the exact position of a home on its street. House Price Derry therefore presents a guide range rather than pretending to offer a formal RICS valuation. It also uses a confidence label and explains the method behind the estimate.

This is a useful product-design principle. Uncertainty should be designed into the interface, not buried in small print. Where information is incomplete, show the range, confidence or reason for human review. That is better than giving a precise-looking answer the underlying evidence cannot support.

Turn information into a complete workflow

The product does more than display an area average. It moves from a short set of property details to an on-screen guide, then offers a fuller report by email. The visitor can stop there. Contact with a professional remains optional.

That sequence matters because a useful digital product needs a beginning, a clear output and a sensible next step. It should also preserve the user's control. Asking for every detail before showing any value would create unnecessary friction. Passing the visitor into a sales process without clear consent would damage trust.

When we design business systems, we use the same questions. What starts the workflow? Which information is genuinely required? What can happen automatically? Where does a person need to review the result? What should happen next if the user chooses to continue?

Use AI where it helps, not where rules are clearer

Not every part of a modern data product needs AI. Source validation, calculation rules, consent and the definition of an approved reporting period should be explicit and testable. They should not change because a language model produced a different answer.

AI can be useful around that dependable core. It can help classify incoming questions, summarise supporting information, prepare plain-English explanations or identify records that need review. The important distinction is between flexible language work and the rules that protect the integrity of the result.

Our guide to AI data readiness for business covers the checks needed before connecting business information to an AI workflow.

Local focus can make a product stronger

A national product may have more data and a larger audience, but a focused local product can answer questions more directly. House Price Derry uses the language, areas and concerns familiar to homeowners in and around the city. The content can explain the local market without forcing visitors through a national portal designed for several different jobs.

That does not mean every business needs a separate product for every town. It means scope should follow a real audience and a real need. Starting narrowly can make the data model, user journey and content much easier to test. Expansion should come after the first version is useful.

Build the update process at the same time

A data product is not finished when it launches. House-price figures change as new official releases are published. The site therefore needs a repeatable process for updating the source period, recalculating the guides, checking the output and refreshing the explanation shown to visitors.

This is where many business tools quietly decay. The first version works, but nobody owns the next data import, the exception list or the wording that explains a changed result. A practical build should name the owner, source, review step and update frequency from the beginning.

What other businesses can take from the build

Many SMEs hold useful information that customers or staff cannot easily use. It may sit in job records, stock files, inspection logs, service reports, price lists or public datasets. The opportunity is not always another dashboard. It may be a focused calculator, guided report, search tool, customer portal or internal decision workflow.

A sensible first version should:

• answer one clear question for one defined audience;
• use a named and reviewable source;
• show uncertainty and limitations honestly;
• collect only the information the workflow needs;
• leave sensitive or consequential decisions with a person;
• include a repeatable update and quality-check process; and
• give the user a useful result before asking for a bigger commitment.

House Price Derry is a good example of a small, local product doing one job clearly. The useful work was not simply accessing the data. It was turning that data into a trustworthy journey a homeowner can understand.

Sources and further reading

• UK House Price Index reports
• House Price Derry methodology

If your business has useful data trapped in spreadsheets, systems or public sources, Blue Canvas can help turn it into a practical tool or workflow. Book a free 15-minute call.