How to measure repeat visits to a restaurant
You will be able to measure repeat visits at every restaurant site using the guest WiFi access points you already own.
Part of the guide topicEmail marketing ROI for venues: measuring the visits it brings
Video overview
{{asset:header_image.png}}
Measure repeat visits to your restaurant in eight steps, on the access points you already own. Capture each diner's email at WiFi sign-in with a CCPA/CPRA-compliant opt-in. The login that captured the email then detects the return visit, so Purple Engage reports repeat rate and campaign-driven returns. Purple handles 500 million logins a year.
{"slot":"hero","alt":"A restaurant manager at the bar reviews a repeat visit report on a tablet while diners eat at tables behind her","caption":"Repeat visits measured from the WiFi login that captured the email","prompt":"Warm, natural-light photograph of a mid-size neighborhood restaurant during lunch service. In the foreground a manager stands at the bar holding a tablet showing a simple line chart and a percentage figure. Diners at wooden tables in soft focus behind. Muted purple accent in the tablet interface. No logos, no visible text, no emoji."}
What you will have once repeat visits are measured
At the end you will have a repeat visit report for every site, built from consented first-party data. It shows four numbers your finance lead can check: unique diners, returning diners, repeat rate and the extra visits each campaign produced. Revenue per send follows from those visits and your average spend per head.
The list behind the report builds itself. Each diner who signs in to guest WiFi and opts in becomes a profile with a first visit date and a site. When that diner comes back and connects, the login that captured the email detects the return visit. Mailchimp-style tools send campaigns, but they do not build the list. They also cannot see the diner walk back in.
You will also have a baseline. Without one, a busy month looks like a successful campaign. With one, you can separate the visits your email caused from the visits that would have happened anyway.
{"tone":"proof","title":"Harrods","text":"Harrods turned guest WiFi into a 57x ROI loyalty channel."}
What do you need before you measure repeat visits?
Most failed setups trace back to a missing prerequisite rather than a wrong step. Confirm each row below before you start.
{"caption":"Prerequisites for restaurant repeat visit measurement","asOf":"2026-06-10","columns":[{"key":"item","label":"Prerequisite"},{"key":"why","label":"Why it matters"},{"key":"check","label":"How to confirm"}],"rows":[["Guest WiFi on supported access points","Purple runs as a cloud overlay on Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme or Fortinet","Your IT contact confirms the make of each site's access points"],["Purple Engage on your account","Campaign-to-return-visit attribution reads from the WiFi login","You can open the measurement reports"],["Email sign-in with an unchecked opt-in","CCPA/CPRA consent must be a clear, affirmative act","A test login shows opt-in status on the profile"],["Staff on a separate network","Staff who connect every shift inflate returning diners","No staff names among guest profiles"],["A written visit definition","Marketing and finance must count the same thing","Both leads have signed it"],["Average spend per guest by site","Turns extra visits into revenue","Finance supplies the figure from the POS"],["Weekly covers by site","Shows what share of diners the WiFi data represents","Covers and WiFi diners sit in one sheet"],["A named owner per site","Someone notices when the sign-in page or network changes","An owner is listed for every venue"]],"footnote":"Purple is hardware-agnostic: no rip and replace is required on the vendors listed.","download":{"filename":"repeat-visit-prerequisites.csv"}}
Two items need a decision, not a setting. Agree on the visit definition and the spend per guest with finance first. If marketing and finance count visits differently, the first review will argue about the numbers instead of the results.
How does a WiFi login measure a repeat visit?
{"kind":"flow","direction":"lr","caption":"How a WiFi login becomes a measured return visit","alt":"Flow from a diner's first WiFi sign-in through consent, profile and campaign to a return login and a repeat visit report.","nodes":[{"id":"a","label":"Diner signs in to guest WiFi","sub":"Email entered on the sign-in page","tone":"neutral"},{"id":"b","label":"Conscious-choice opt-in","sub":"Unchecked box, consent recorded under CCPA/CPRA","tone":"brand"},{"id":"c","label":"First-party profile created","sub":"First visit date and site stored","tone":"brand"},{"id":"d","label":"Campaign sent from Purple Engage","sub":"90% of the segment; 10% held back","tone":"neutral"},{"id":"e","label":"Diner returns and logs in","sub":"The same login detects the return visit","tone":"success"},{"id":"f","label":"Repeat visit report","sub":"Repeat rate, uplift and revenue per send","tone":"success"}],"edges":[{"from":"a","to":"b"},{"from":"b","to":"c","label":"consented"},{"from":"c","to":"d"},{"from":"d","to":"e","label":"within the window"},{"from":"e","to":"f"},{"from":"c","to":"e","label":"organic return","style":"dashed"}]}
A diner connects to guest WiFi and enters an email on the sign-in page. This is the captive portal: the page a phone shows before it reaches the internet. If the diner checks the marketing opt-in, Purple creates a first-party profile with the date and site of that first visit. Purple calls these conscious-choice opt-ins, because nothing is pre-selected.
Every later login from that diner adds a visit to the same profile. The identity comes from the login the diner made. You are counting people who chose to connect, not anonymous devices passing the window.
The metrics, defined
| Metric | How it is calculated | Decision it informs |
|---|---|---|
| Unique diners | Distinct opted-in profiles with a visit in the period | Size of the list you can reach |
| Returning diners | Unique diners with two or more visit days in the lookback window | Whether people come back at all |
| Repeat rate | Returning diners divided by unique diners | Health of each site over time |
| Days between visits | Average gap between a diner's visit days | When to send a win-back email |
| Campaign uplift | Recipient return rate minus holdout return rate | Whether the campaign caused the visits |
| Revenue per send | Extra visits x spend per head, divided by emails sent | Budget and frequency of sends |
The formulas
Keep these in one shared sheet so marketing and finance read the same numbers.
Repeat rate (%) = returning diners / unique diners x 100
Campaign uplift (pts) = recipient return rate - holdout return rate
Extra visits = uplift x number of recipients
Campaign revenue = extra visits x average spend per head
Revenue per send = campaign revenue / emails sent
How to measure repeat visits to a restaurant, step by step
Work through the eight steps in order with Purple open. Each step ends with a check. Do not move on until it passes, because every later number depends on the one before.
{"title":"Measure repeat visits to a restaurant in 8 steps","steps":[{"title":"Capture email at WiFi sign-in","body":"Make email the sign-in field on your guest WiFi page. Add a separate, unchecked marketing opt-in with plain wording.","check":"Sign in on a test phone. A profile appears with your email and opt-in status."},{"title":"Name each site and remove staff","body":"Give every venue its own name in Purple. Move staff devices to a separate Staff WiFi network.","check":"Each site shows its own logins, and no staff names appear in guest profiles."},{"title":"Agree the visit definition","body":"Count one visit per diner per day. A returning diner has two or more visit days in the lookback window, such as 90 days.","check":"Marketing and finance have signed the same written definition."},{"title":"Record a baseline","body":"Run four to eight weeks with no campaigns. Note unique diners, returning diners and repeat rate per site each week.","check":"Weekly repeat rate stays within a narrow band for three weeks running."},{"title":"Split the segment and hold back 10%","body":"Pick opted-in diners who have not returned in 30 days. Randomly split them: send to 90%, hold back 10%.","check":"Both group sizes are written down before anything is sent."},{"title":"Send and fix the attribution window","body":"Send the campaign. Choose a window for counting returns, such as 14 or 30 days, and keep it for every send.","check":"The send log shows delivery to the 90% group only."},{"title":"Read campaign-attributed returns","body":"Purple Engage matches each returning recipient's WiFi login to the campaign they received.","check":"The report lists return visits from recipients inside the window."},{"title":"Compare with the holdout and report","body":"Uplift is the recipient return rate minus the holdout return rate. Multiply extra visits by spend per head.","check":"Uplift is positive on two consecutive sends before you report it to finance."}]}
{"tone":"tip","title":"Pilot one site first","text":"Run the baseline and first send at one site with steady trade. Settle the definitions there before you roll out across the estate."}
Worked example: a 140-cover neighborhood restaurant
The figures here are illustrative, to show the arithmetic end to end.
Situation. A 140-cover neighborhood restaurant knew its weekly covers but not who came back. Its email list sat in a spreadsheet nobody trusted.
What was done. The team put email sign-in on guest WiFi and ran a six-week baseline. It recorded 1,800 unique WiFi diners, of whom 360 returned: a 20% repeat rate. Next, it picked 1,200 opted-in diners who had not returned in 30 days. It sent a midweek offer to 1,080 and held back 120.
Outcome. Within 30 days, 14% of recipients returned against 9% of the holdout. That is a five-point uplift, or 54 extra visits. At $32 spend per head, the campaign caused $1,728 of revenue: $1.60 per email sent. A 120-diner holdout is small, so the team repeated the send before reporting.
How do you check your repeat visit measurement works?
Run three checks before the report goes to anyone outside marketing.
Reconcile against the register. Compare WiFi diners per location with covers for the same week. The share will never be 100%, but it should hold steady. A sudden drop usually means the sign-in page changed or an access point went offline.
Inspect the top returning profiles. Open the 20 most frequent returners at each location. Staff, test accounts, and delivery drivers show up here first. Remove them before you calculate the repeat rate.
Repeat the uplift. One positive send can be chance. If the holdout returned faster than recipients, check that the split was random rather than, say, alphabetical.
{"title":"Repeat visit measurement sign-off","groups":[{"heading":"Data quality","items":[{"text":"WiFi diners as a share of covers is steady week to week","why":"A drop means the sign-in page or an access point changed"},{"text":"The 20 most frequent returning profiles contain no staff or test accounts"},{"text":"Visits count once per diner per day"},{"text":"Every location reports its own logins"}]},{"heading":"Consent","items":[{"text":"The opt-in box is unchecked by default"},{"text":"Marketing consent is separate from the WiFi terms"},{"text":"The privacy notice mentions visit measurement"}]},{"heading":"Attribution","items":[{"text":"Holdout size was recorded before the send"},{"text":"The attribution window matches the previous send"},{"text":"Uplift is positive on two consecutive sends"},{"text":"Finance has confirmed spend per head"}]}],"download":{"filename":"repeat-visit-measurement-checklist.csv"}}
What goes wrong with repeat visit measurement most often?
Counting logins instead of visit days
A diner can register more than one login during a single meal. If you count logins, one lunch becomes a repeat visit. Fix: apply the one-visit-per-day rule from step three to every report.
Letting staff into the guest count
Staff who connect every shift look like your most loyal diners.
Situation. A 12-location pub-restaurant group saw one location report a 38% repeat rate. The other 11 sat between 19% and 23%. The regional manager wanted to copy that location's offers across the portfolio.
What was done. The CRM manager opened the most frequent returning profiles. Most belonged to staff using guest WiFi on every shift. The location moved staff to a separate Staff WiFi network and removed those profiles from reporting.
Outcome. The location's rate settled at 21%, in line with the portfolio. The group avoided rolling out a playbook built on a counting error. (Figures illustrative.)
Pre-checked or bundled consent
{"tone":"warning","title":"Pre-ticked boxes void the list","text":"Under CCPA/CPRA regulations, silence, pre-checked boxes or inactivity do not count as valid consent. A list built that way cannot be used, and every return visit you attribute to it is in doubt."}
Consent must also be separate from the WiFi terms. Bundling them means a diner cannot connect without agreeing to marketing, which fails the "freely given" test under the CCPA/CPRA. Fix: one unchecked box, worded in plain English, separate from access.
Sending with no holdout
Without a holdout, every return after a send looks like campaign revenue. Many of those diners would have come back anyway. Fix: hold back 10% on every send and report uplift, not raw returns.
Moving the attribution window
A 30-day window will always show more returns than a 14-day one. Switching between them makes later campaigns look better or worse than they were. Fix: pick one window per campaign type and keep it.
Counting hotel guests as repeat diners
In a hotel restaurant, a guest who connects every night of a three-night stay is one guest, not three repeat visits. Fix: count returns across separate stays, or report guests and non-guest diners apart.
Comparing locations with different WiFi take-up
A location where few diners connect has a smaller, less typical sample. Its repeat rate can swing on a handful of regulars. Fix: report WiFi diners as a share of covers beside every repeat rate, and flag locations below your brand average.
If a check fails and the cause is not listed here, go back to the step it belongs to. Re-run that step's check before you move forward, because each later figure inherits the error.
Where repeat visit measurement sits in email marketing ROI
Repeat visit measurement is the evidence layer of email marketing ROI. The pillar guide to email marketing ROI for venues covers the rest: revenue per send, opt-in rate and lifetime value. Once repeat rate and uplift hold steady over two sends, those figures feed straight into a lifetime value model your finance lead will accept.
Purple Engage measurement reports campaign-to-return-visit attribution from the WiFi login. For the wider view of what WiFi data tells you about your venues, read Purple's WiFi analytics guide.
The approach scales. Purple runs across 80,000+ live venues in 90 countries and handles 500 million logins a year (Purple's own data).
Frequently asked questions
How much does repeat visit measurement cost?
It needs no new hardware, so the cost is a Purple subscription rather than a network project. Purple runs as a cloud overlay on the access points you already own. Plans run as Connect, Capture and Engage, and campaign-to-return-visit attribution sits in Purple Engage. Ask Purple for a quote based on your sites and the plan you need; the walkthrough above shows what that plan delivers.
Will it work with the access points we already have?
Yes, if they come from a supported vendor. Purple is hardware-agnostic and works with Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme and Fortinet. It layers on top of existing infrastructure, so there is no rip and replace. Your IT contact can confirm the make from the network dashboard, and the marketing setup then follows the eight steps above.
Do we have to leave Mailchimp, Klaviyo or HubSpot?
No. Purple integrates with CRM and marketing platforms, so you can keep your current sending tool. The difference is where the list and the visit data come from. Those tools send to lists you already hold. Purple builds the list from guest WiFi logins and sees the diner return. Sending from Purple Engage keeps the campaign and the return visit in one report.
Is measuring repeat visits through WiFi GDPR compliant?
Yes, when the consent is valid. GDPR Article 4(11) requires consent that is freely given, specific, informed and unambiguous, and Recital 32 rules out pre-checked boxes. Purple's conscious-choice opt-ins record that choice at sign-in. Purple holds ISO 27001, Cyber Essentials and B Corp certification and is CCPA/CPRA compliant. Your privacy notice must still explain visit measurement to diners.
How accurate is a WiFi-based repeat rate?
It is precise for diners who connect and a sample for those who do not. Every return login from an opted-in diner counts, so the figure for connected diners is not an estimate. Diners who never connect stay outside it. Compare WiFi diners with covers each week. If the share holds steady, trends in the WiFi repeat rate are a fair guide to the whole room.
How long before the numbers are reliable?
Plan on about three months. Spend four to eight weeks on a baseline with no campaigns. Then run at least two sends with a holdout before you report uplift to finance. A single send with a small holdout can swing several points by chance. Two consistent results are the minimum evidence for a budget decision, and they give you a defensible revenue per send.
Who needs to be involved, and how much effort is it?
Three roles: a marketing owner, a finance lead and a manager at each site. Marketing runs the sign-in page, segments and sends. Finance signs off the visit definition and the spend per head. Site managers keep staff off guest WiFi and report changes on the floor. After the baseline, reading the report and running a send fits inside a normal monthly campaign cycle.