Cost per enquiry is useful only when a business knows what an enquiry represents. A venue might count requests to hire a room, catering enquiries and general availability questions together, even though they need different amounts of staff time and have different chances of becoming bookings. Artificial intelligence can help organise those requests, but cheaper inbox administration is not automatically a lower cost of acquiring worthwhile business. A sound approach measures acquisition spending, handling effort and later outcomes separately.
Small UK service firms do not need a complex prediction model to begin. They need clear definitions, a reliable record of where enquiries came from and a practical way to check the work created by each channel. Google's conversion reporting documentation distinguishes primary conversion actions and shows how advertising cost per conversion is calculated. That is a useful reporting definition, not a guarantee that every tracked conversion is a good sales opportunity.

Decide which contact is worth calling an enquiry
A visitor opening a page is not equivalent to submitting a detailed quotation request. Even completed contact forms differ: somebody asking for a brochure may be exploring, while somebody specifying venue date, attendance and accessibility requirements has a concrete requirement. Choose a definition for each meaningful enquiry type and use it consistently across phone, email, referrals and paid campaigns. Keep general information requests in a separate category when that distinction helps decisions.
For a venue, record the event type, rough date, expected guest numbers and stage of the enquiry without requiring unnecessary personal information. A well-designed form can encourage useful details without pretending that every incomplete request has low value. Document what happens when the same customer calls and then submits a form: two messages should not silently become two independently acquired leads. This definition work is more valuable than buying software that makes an unstable metric look precise.
Calculate advertising cost and full handling cost separately
For a simple illustrative calculation, a campaign costing £400 and yielding 20 consistently counted enquiries has a media cost per enquiry of £20. That calculation is arithmetic, not a claim that any actual venue achieves those numbers. Google's cost-per-conversion metric uses eligible advertising costs divided by recorded primary conversions, which can include modelled conversions. It therefore need not match a service firm's own count of received calls and emails.
A second calculation considers staff effort, tools and rework. If many initial enquiries require repeated calls or duplicate data entry, the operational cost of handling them may be substantial even when the advertising dashboard looks healthy. Keep both measures instead of blending them into an unexplained total. A software subscription should appear as a real cost in the business case, and a claimed saving should include time spent supervising, correcting and maintaining automation.
Remove confusion before the customer makes contact
Some expensive enquiries are created because the website fails to answer ordinary questions. Venue customers may not know the room capacity, whether their event type is permitted, when access begins, how accessible the entrance is or what the quoted hire fee includes. Improving those pages can reduce unsuitable enquiries and make suitable ones more informed. An assistant may help visitors locate maintained information, but it cannot replace the business's obligation to keep that information accurate.
Make the limits visible. A published capacity may depend on room layout, and an advertised package may exclude equipment or staffing. Where details require a site assessment, say that explicitly. Do not invite a chatbot to invent a figure to keep a visitor engaged. Count useful outcomes such as fewer clarification emails and better-completed briefs, not just a lower number of contact forms. A reduction in enquiries may be positive if irrelevant demand falls while qualified bookings remain steady.
Route new requests without losing human responsibility
A common failure happens after acquisition: a form sends a generic notification, no one owns the message and the prospect looks elsewhere. AI-assisted classification can identify subject, location, intended date and urgency, then propose a queue or responsible department. Keep the original message accessible, especially when a summary could drop a crucial condition such as wheelchair access or a strict deadline. A routing rule should state who checks unassigned or ambiguous requests.
Define a narrow automatic response that confirms receipt and explains the real next step. Acknowledgement is not acceptance, and an automatically suggested slot is not a booking. For events with complex safety or legal questions, a specialist should respond. A documented handover reduces duplicated effort and may make the existing marketing investment more productive, but neither the assistant nor a shorter response time proves the enquiry will convert.
Compare lead quality without hiding the difficult cases
A low-cost campaign producing many contacts can look attractive until the team sees how many requests concern services it does not supply. Compare the share that is relevant, quoted, booked and ultimately completed. Track outcomes over a meaningful period rather than treating an unreturned call as a definite lost sale. Some large events take months to decide; others are urgent. Segment them if the difference would materially change a decision.
AI classification can highlight patterns but should remain explainable. A prospective customer who has an unusual requirement should not be discarded solely because it resembles a historic low-value pattern. Use objective criteria such as service area or date availability when those really govern eligibility; flag exceptions for review. Do not assign invented lifetime values merely to improve a dashboard. The strongest evidence combines real outcome records with a clear account of what the team was able to serve.
Use clean records and appropriate privacy boundaries
Linking advertising to downstream business requires discipline. Use a consistent enquiry identifier and preserve channel attribution only where it is reliable. If people contact the venue on multiple devices or by telephone after seeing an advert, some attribution will remain uncertain. Avoid presenting estimates as observed facts. Note missing consent or measurement limitations when comparing campaigns; uncertainty does not justify filling every gap with a fabricated AI-generated label.
The ICO's data-minimisation guidance requires organisations to hold no more personal data than necessary for the purpose. A detailed event brief may include contact details and sensitive accessibility requests. Avoid sending those details to a third-party scoring system without a proper data-handling assessment. Limit access, specify retention and ensure staff can correct an incorrect lead record or respond to a data-protection complaint.
Run a small experiment before declaring savings
Choose one fix, such as automated assignment of email enquiries, and compare it with a recent period using the same definitions. Record how many messages arrived, reached the correct owner, progressed to a useful conversation and needed manual correction. Note staffing changes and seasonal demand, as these can affect results independently of the tool. A reliable test can show that routing improved, even where there is insufficient evidence to claim the advertising campaign itself became cheaper.
Document the operational exceptions: a lead without a date, two people enquiring on behalf of one organisation, a customer who withdraws interest and an enquiry requiring a different venue. If automated rules mishandle these cases, correct the workflow rather than counting them as failed prospects. Recheck whether follow-up messages respect customer preferences and applicable marketing rules. A bounded trial protects the business from turning a plausible efficiency idea into an unsupported financial promise.
Judge the whole enquiry journey by useful bookings
Look at the funnel from discoverability and clarity to first response, qualification, quotation and booking. Fix each handover separately. A company may find that the real opportunity is removing repeated availability checks, explaining its offer more clearly or preventing messages from sitting unassigned over a weekend. AI tools can assist with those tasks, while people retain the commercial judgement about pricing, fit and contractual commitments.
Cost per enquiry should be a decision aid rather than a target pursued regardless of quality. Keep arithmetic, assumptions and business results distinguishable. If improved routing saves time but does not reduce media costs, say so. If fewer unsuitable contacts arrive while profitable bookings remain stable, examine that benefit without assuming a universal conversion gain. The lasting advantage comes from a traceable, measured process, not a promise that automation alone makes leads inexpensive.
Cost per enquiry or cost per won customer: which figure answers the decision?
A low advertising cost per enquiry can hide poor fit or expensive administration. A small service company should follow enquiry records through to verified wins before treating lower lead-acquisition cost as better business performance.
| Check | Advertising cost per enquiry | Cost per won customer |
|---|---|---|
| Numerator | Relevant advertising spend for the reporting period; state which costs are included. | Relevant acquisition spend, and handling costs separately where the business wishes to model full acquisition cost. |
| Denominator | Consistently defined valid enquiries, or platform conversions if labelled as such. | Verified new customers or jobs won from the same defined cohort and observation window. |
| Illustrative result | £400 media spend ÷ 20 valid enquiries = £20 per enquiry. | If only 4 of those enquiries became won customers, £400 ÷ 4 = £100 media cost per won customer, before handling costs. |
| Main limitation | Low value can reflect many unsuitable, duplicate or unqualified contacts. | Slow sales cycles, incomplete CRM records or uncertain attribution can distort the denominator. |
Practical decision: review enquiry quality and verified won work before moving money to a campaign with the cheapest initial contact. The £400, 20 and 4 figures are transparent hypothetical arithmetic, not observed results or a guaranteed AI benefit. Advertising dashboard conversions should not be silently relabelled as won customers.
Primary source A — campaign conversions: Google Ads, About conversion columns and reporting (platform measurement documentation, checked 26 October 2026). Reported conversions depend on the action configuration, attribution and relevant reporting rules; they are not automatically qualified enquiries or signed service contracts.
Primary source B — later lifecycle events: Google Analytics, Recommended events (Google Analytics 4 event model, checked 26 October 2026). Its lead-generation events distinguish stages such as generate_lead, qualify_lead and close_convert_lead where deliberately implemented. A named event does not guarantee that the underlying business record is correct. These sources define measurement methods, not a cross-company performance comparison or AI-caused saving.