Direct answer: A diner can leave at any of four observable stages: Intent, Comparison, Validation or Action. There is no universal benchmark showing that one stage always causes the most drop-off. Testing each stage with pass/fail criteria lets an operator identify the largest local gap before deciding what to fix.

Why the Path from Search to Booking Breaks

A diner picking up their phone and searching "Italian restaurant open now near me" is expressing high-purchase intent. Yet a significant number of those searches do not end in a reservation at any restaurant, let alone at the one that appeared first. The reason is rarely a single failure; it is usually a chain of small friction points spread across four distinct stages. Operators who address only one stage — typically discoverability — while ignoring the others are optimising for visibility without capturing the value that visibility creates.

The framework below, called the ICVA Funnel (Intent → Comparison → Validation → Action), gives each stage a name, describes the diner's behaviour at that stage, and provides a pass/fail checklist an operator can complete in a single working session.

The ICVA Funnel Explained

Stage 1 — Intent: Being findable when the need arises

At the Intent stage a diner has a need but no specific restaurant in mind. They are entering keyword queries, using voice search, or opening a map application. The restaurant's job here is purely to appear. Google's guidance on improving local ranking identifies three primary factors that influence whether a Business Profile surfaces: relevance, distance, and prominence. Operators control relevance and prominence directly.

Relevance is shaped by how completely a Google Business Profile is filled out — category selection, services, opening hours, and attributes. Prominence is partly determined by the quantity and quality of reviews and by the consistency of business information across the web. Distance is a user variable that operators cannot change, but selecting the most accurate primary and secondary categories is the single highest-leverage action within operator control.

Stage 2 — Comparison: Standing out in a results list

Once a set of results appears, the diner scans thumbnails, star ratings, opening hours, price range, and cuisine type before clicking anything. At this stage the restaurant is competing visually against several alternatives. Structured data accelerates this process. Google's LocalBusiness structured data documentation explains that adding structured data to a restaurant's own website helps Google understand and display business information more accurately in rich results.

The Schema.org Restaurant type, which extends LocalBusiness, allows operators to mark up properties including servesCuisine, priceRange, openingHoursSpecification, menu, and address. A restaurant whose website communicates these signals in machine-readable form gives search engines a more confident basis for displaying accurate comparison data. A listing that shows incorrect hours or no price range during the comparison stage loses clicks to a competitor that has these fields populated.

Stage 3 — Validation: Building enough trust to proceed

After clicking through to a listing or website, the diner is not yet ready to book. They are validating: checking that the photos match expectations, that reviews are recent, that the menu looks appealing, and that practical details — parking, accessibility, whether they take large groups — answer their specific concern. This is the stage most frequently neglected by operators who consider their work done once the website is live.

Google's Business Profile guidance recommends that operators respond to reviews, add current photos, and keep menus updated. Each of these actions directly addresses the questions a diner is asking during Validation. A profile with no photos uploaded in eighteen months, a menu PDF last modified two years ago, and no owner responses to recent negative reviews fails the Validation stage for a meaningful proportion of prospective customers.

Stage 4 — Action: Completing the reservation without friction

The Action stage is where the guest attempts to complete a reservation. Schema.org's ReserveAction type can express a reservation action and its target URL in machine-readable form. Markup alone does not guarantee that a booking link will appear in a search product; Google documents separate eligibility and integration paths for direct booking actions. The restaurant still needs a visible, fast and mobile-tested reservation route.

Friction at this stage is often invisible to operators because it happens after the click. Common failure points include: a reservation widget that does not load on mobile, a phone number that goes to voicemail with no callback option, a third-party booking link that redirects to a generic homepage, and a reservation form that times out before the diner completes it.

The ICVA Funnel Audit Checklist

Run this checklist on a mobile device. A fail on any item represents a measurable gap in the funnel.

Stage Audit Item Pass Criteria Fail Indicator
Intent Google Business Profile completeness All core fields populated: name, address, phone, category, hours, website Any field blank or marked "Add missing info"
Intent Primary category accuracy Primary category matches the restaurant's main cuisine or concept Generic category (e.g., "Restaurant") used when a specific one applies
Intent Opening hours current Hours reflect today's actual schedule including special hours for holidays Diner arrives to find restaurant closed despite "open now" status
Comparison LocalBusiness / Restaurant structured data on website Validated via Google's Rich Results Test with no errors No structured data present or validation returns critical errors
Comparison Schema fields populated servesCuisine, priceRange, menu, openingHoursSpecification, address, telephone all present One or more key fields absent from structured data
Comparison Review recency At least one review response from owner within the last 60 days No responses, or all responses are older than six months
Validation Photo currency Interior, exterior, and food photos added within the past twelve months Oldest photos show décor or dishes no longer present
Validation Menu accuracy Menu on website and linked from GBP reflects current pricing and items Items listed that are no longer available; prices differ from in-restaurant
Validation Attributes set Relevant attributes confirmed: outdoor seating, accessibility, payment types, reservations accepted Attributes not reviewed or set to "don't know"
Action ReserveAction markup Schema.org ReserveAction implemented with a working target URL No reservation markup; booking link absent from structured data
Action Mobile reservation flow Reservation can be completed end-to-end on a mobile browser in under 90 seconds Widget fails to load, redirects incorrectly, or requires more than three screens
Action Phone number functional Calling the listed number during service hours reaches a person or a clear voicemail with callback promise Engaged tone, no answer, or number disconnected

How to Prioritise Fixes: A Simple Decision Rule

If your audit returns multiple fails, use this sequence to prioritise: fix Intent fails first (you cannot convert a customer who never finds you), then Comparison fails (you cannot win the click if your listing is incomplete), then Validation fails (you cannot earn trust with outdated information), and finally Action fails (you cannot capture the booking if the path is broken). Any Action fail should be treated as urgent regardless of sequence because it destroys value already created by the earlier stages.

Measuring Progress

Structured data changes can be verified immediately using Google's Rich Results Test. Google Business Profile provides an Insights panel that shows search queries, views, and the actions users take on the listing (calls, direction requests, website clicks). Monitor these weekly after making changes. Website-side changes to the reservation flow should be tested by completing a real booking on a mobile device after every update. Operators should treat a drop in "website clicks" from GBP alongside stable "views" as a signal that the Comparison or Validation stage has a new problem.

Explicit Limitations of This Framework

The ICVA Funnel describes observable on-page and on-listing factors. It does not account for offline word-of-mouth, social media discovery, or the effect of paid advertising. Structured data implementation does not guarantee rich result display; Google's documentation notes that structured data is eligible for but does not guarantee enhanced presentation. Review volume and star rating affect perception but this framework does not prescribe any particular review acquisition strategy. The audit is a point-in-time snapshot and needs to be repeated whenever menus, hours, or the reservation system changes.

Where a Platform Like ChefNet May Fit

Restaurant management and visibility platforms can reduce the manual work involved in keeping profile information consistent across channels, which is directly relevant to the Intent and Comparison stages of this funnel. ChefNet is designed to support restaurant operators in managing their digital presence and reservation-related information. Operators evaluating such platforms should verify whether a given tool supports structured data output, profile synchronisation, and direct reservation pathways before assuming those capabilities are present — and should audit the results using the checklist above regardless of which tools they use.

Primary sources

FAQ

How can a restaurant find where its discovery-to-booking funnel breaks?

Test each stage with restaurant-specific evidence: listing accuracy for Intent, profile completeness for Comparison, current trust information for Validation, and a complete mobile booking test for Action. Compare the failures you observe rather than assuming a universal weakest stage. A broken Action path is urgent when testing shows that guests cannot complete a reservation.

Does adding structured data guarantee that my restaurant appears in rich results?

No. Google's own LocalBusiness structured data documentation states that structured data makes a page eligible for rich results but does not guarantee them. Correct, error-free implementation is a prerequisite, not a promise of enhanced display. Use Google's Rich Results Test to confirm your markup is valid and monitor Search Console to see whether rich results are being served.

Which Schema.org properties are most important for a restaurant to mark up?

The Schema.org Restaurant type, which extends LocalBusiness, includes properties most relevant to diners in the comparison stage: servesCuisine, priceRange, menu, openingHoursSpecification, address, and telephone. Adding ReserveAction with a working target URL supports the booking step directly. Prioritise accuracy over completeness — an incorrect value is worse than an absent one.

How often should a restaurant repeat this funnel audit?

The audit should be run whenever a significant change occurs: menu or pricing updates, change of opening hours, renovation affecting photos, or a switch to a new reservation system. As a minimum baseline, a full audit once per quarter is a reasonable maintenance cadence. Individual checklist items such as review response recency and photo currency should be reviewed monthly.

Editorial disclosure: ChefNet publishes this guide and develops products for restaurant discovery and operations. General operating guidance is separated from product claims. Capabilities can change as pilots progress. Published 2026-07-22.