Reading demand for licensing help through friction rather than search volume

Search counts reveal visits, not the task behind them. Someone may be testing a business idea, checking a rule, preparing paperwork or repairing a returned form; each behaviour points to a different service need. Read demand by the obstacle and the decision it could inform, not by traffic alone.

What can demand data actually tell me?

Demand becomes useful when it shows what blocks a person from taking the next appropriate step. A page view cannot tell whether someone is exploring, checking a rule or preparing paperwork, so each behaviour needs its own interpretation and a service decision it might inform.

Use the channel strategy to distinguish orientation from application preparation. That separation can help decide whether a change belongs in guidance, a handover or case support. A reading of demand should not turn an observed action into an authority decision or result.

Where should I stop tracing a licensing journey?

Stop at the last event your organisation can observe reliably. An information service may follow readers as far as departure to an official route, but should not claim to know whether an authority later accepted an application.

A practical path can run through discovery, authority identified, application started, evidence completed, submission acknowledged and decision received. The case workflow can help separate those stages. An acknowledgement records submission, not approval, so keep the two events distinct in reports.

The Department for Business and Trade's Find a licence service asks for activity and location, offers filters covering all four UK nations, and says its catalogue may not contain every licence. A publisher can count readers moving towards that service, but cannot infer that every visit became a valid application.

Which obstacles should I record from enquiries?

Record what people say is getting in their way, rather than grouping every contact under one demand total. Relevant categories include uncertainty about activity, the wrong council, missing evidence, a premises issue, an accessibility problem and unclear responsibility for a case.

Repeated questions about one field may call for clearer wording or a better handover. Ten detailed cases pointing to the same blockage can offer a more actionable lead than a much larger page-view total without a clear reason behind it. The obstacle gives a service team something specific to examine.

What does a completed food registration step mean?

It signals stronger intent than a general search, but it does not prove that registration was accepted or that no other permission is needed. Treat completed steps as evidence of progress, not as an authority decision.

The Food Standards Agency's food business registration service asks for operator, establishment and activity information so a case can be routed. Completing those steps indicates stronger intent than a general search, but it does not prove registration was accepted or show that other permissions are unnecessary.

How should I define a signal before counting it?

Define each measure before counting it, including its source, owner and review frequency. This prevents a completed search, a support interaction and a correction request from being treated as equivalent evidence.

For example, count discovery only after an activity-and-location search is completed, an assisted case only after a genuine support interaction, and a return when an authority requests a correction. Keep missing data visible. Do not replace an unknown decision status with the most favourable assumption. See each signal's definition.

The Government Digital Service's guidance on using data to improve a service connects measures with user needs, baselines and decisions. It is service-practice guidance rather than licensing law, and supports measurement design, not any claim about national demand levels alone.

What can assisted contact tell me?

Keep supported cases in their own group and record why someone contacted the service. Frequent support can indicate a useful service need, but it can also expose unclear information or a routing defect.

Record whether the person reached the correct official route afterwards, then review a small sample before deciding what a high assistance rate means. The contact count alone cannot distinguish valuable demand from a problem in the content or handover.

Keep repeat visitors separate from new users when the data allows. A person returning to check a condition is not equivalent to a founder making a first enquiry. For privacy, use aggregated analysis unless case-level linking is necessary, justified and controlled.

Why might enquiry numbers rise or fall?

A rise in enquiries does not by itself mean underlying demand has grown. Rule changes, campaigns, seasonal trading patterns or a broken form can also change what gets recorded.

Compare the same definition over time and mark changes to channel, geography or eligibility. Separate England from other nations where their routes differ, and review abandonment alongside accessibility feedback before concluding that people simply lost interest. A market outlook should make those limits visible.

Choose one question for the next reporting period, such as whether applicants find the correct council without assistance. Collect only the events needed to answer it, review a sample of recent journeys and state what the resulting signal cannot establish.

What should I do first?

Start with one service question, then trace only the events needed to answer it. Check a recent sample, keep unknown outcomes visible and record what the evidence cannot establish.