Win-back emails for lapsed hotel and restaurant guests: timing and offers
You will be able to define when a hotel or restaurant guest has lapsed, using their last guest WiFi login and your own median return interval.
Part of the guide topicVenue email campaigns: what to send and when
Send a win-back email when a guest has gone roughly twice their usual return interval without a visit, measured from their last WiFi login. Open with a no-discount reminder, follow with a value offer around two weeks later, then a final message before you stop mailing. Match the offer to the guest's past value, not a blanket discount.
What does a win-back sequence do for your venue?
A win-back sequence goes to guests who used to visit and have stopped. Its job is a measurable return visit: a booked stay in a hotel, a seated cover in a restaurant. Opens and clicks are signals along the way, not the result.
The hard part is rarely the email itself. It is knowing who has lapsed, and when. E-commerce platforms such as Klaviyo key their win-back flows on the date of the last online order. A diner who pays at the table leaves no such date in your email platform. Neither does a hotel guest who booked through an online travel agent.
Guest WiFi closes that gap. Every login records a visit to a named venue on a named date. With Guest WiFi from Purple, guests who opt in to marketing at login join your list as consented, first-party data. Purple Engage then uses the date of each guest's most recent login to time the win-back email to lapsed guests. Mailchimp-style tools send campaigns, but they do not build the list or know when a guest was last on site.
The data behind this is large. Purple recorded 440 million logins in 2024 across 80,000+ live venues (Purple data). Hospitality operators running Purple include Premier Inn, Whitbread, Pizza Express and Stonegate Pubs.
A well-run sequence moves three numbers:
- Repeat visits: the share of lapsed guests who log in again within 30 days of a send.
- Revenue per send: incremental revenue from returning guests, divided by emails delivered.
- List health: fewer complaints and fewer dormant contacts, because the sequence ends by removing guests who no longer engage.
What do you need before you send win-back emails?
Contacts with a recorded marketing opt-in
GDPR defines consent as freely given, specific, informed and unambiguous (Article 4(11)). Article 7(1) requires you to be able to demonstrate that consent. In the UK, regulation 22 of the Privacy and Electronic Communications Regulations (PECR) also requires prior consent for marketing emails to individuals.
PECR has a narrow "soft opt-in" for contact details collected during a sale. The Information Commissioner's Office (ICO) interprets it tightly, so do not rely on it for WiFi-sourced contacts. Use the explicit opt-in captured at login instead. Purple's conscious-choice opt-ins ask guests to choose marketing as a separate step from accepting the WiFi terms. Guests who connected without opting in stay out of the sequence.
If your opted-in list is still small, start with our guide on how to build a guest WiFi email list that actually converts.
Enough visit history to see a pattern
You need at least one full return cycle of login data. A coffee shop may see a clear pattern within eight weeks. A leisure hotel needs more than 12 months, because many guests return once a year for the same holiday or event.
A holdout group
Set aside a random 10% of each lapsed segment and send them nothing. Some lapsed guests come back on their own. The holdout tells you how many, so you credit the email only with the extra visits it caused.
Authenticated sending
Since February 2024, Gmail and Yahoo have required bulk senders to authenticate mail with SPF, DKIM and DMARC. They also require one-click unsubscribe. Google asks senders to keep reported spam below 0.10% and never reach 0.30%. Win-back sends go to your least engaged contacts, so they carry the highest complaint risk. Our guide to Gmail and Yahoo sender requirements covers the changes in detail.
How do you use last-visit dates from guest WiFi to time win-back emails?
"Lapsed" means nothing until you define it against your own guests. A diner who visits weekly has lapsed after a month. A guest who stays at your resort every August has not lapsed in February.
Find your normal return interval
Take every guest with at least two logins at the same venue. Measure the days between consecutive visits, then take the median. Use the median rather than the average, because a handful of daily regulars will drag an average down.
Exclude staff devices before you calculate. Staff who connect to guest WiFi every shift look like your most loyal guests and distort every figure. Split hotels by segment too, because business travellers and leisure guests return on different cycles.
Set the lapse threshold at about twice that interval
Two intervals is a practical starting point. One interval overdue may mean a busy month. Two intervals suggests the habit has broken. Once your holdout is running, test 1.5 times against two times on split segments.
Sending too early spends an offer on a guest who would have returned anyway. Sending too late gives a new habit time to form at a competitor.
Handle one-visit guests separately
A guest with a single login has no personal interval. Apply the venue median for their segment and expect a lower response. Their first visit may have been a one-off: a wedding, a conference or a passing trip.
Starting timings by venue type
Use the values below as a first configuration, then replace them with your own medians.
| Venue type | Starting return interval | Email 1 (lapse threshold) | Email 2 | Email 3 | Stop marketing sends | Lead offer |
|---|---|---|---|---|---|---|
| Coffee shop or quick-service restaurant | 10 days | Day 21 | Day 30 | Day 45 | Day 75 | Free add-on with next order |
| Pub or casual dining restaurant | 30 days | Day 60 | Day 74 | Day 90 | Day 150 | Free dessert or drink, Monday to Thursday |
| Destination or fine dining restaurant | 90 days | Day 180 | Day 200 | Day 230 | Day 365 | New menu preview or chef's table invitation |
| Business hotel | 60 days | Day 120 | Day 134 | Day 160 | Day 240 | Late checkout or breakfast included |
| Leisure hotel or resort | 365 days | Day 300 | Day 320 | Day 345 | Day 450 | Room upgrade or added night, booked direct |
Leisure hotels are the exception to the two-interval rule. Annual guests plan ahead, so the first email has to land before they book. Sending at around 10 months reaches them while they are still choosing where to go.
Which offer should you send, and when?
Email 1: a reminder with no discount
Open with what has changed since the guest last visited: a new menu, refurbished rooms, a new season of events. Name the venue they visited, which the login tells you. Some lapsed guests only needed a prompt, and you keep your full margin on every one of them.
Email 2: a value offer, only to guests who have not returned
Around 14 days later, check for a fresh WiFi login. Anyone who has visited leaves the sequence. Everyone else receives the offer. Make it specific, time-limited and easy to redeem at the till or front desk.
Email 3: a last offer and a preference check
The final email repeats or slightly improves the offer. It also asks the guest to confirm they still want to hear from you. If they neither visit nor engage by the stop date in the table, end marketing sends to them.
That step is good practice and a compliance point. GDPR's storage limitation principle (Article 5(1)(e)) says personal data should not be kept longer than the purpose needs. Anyone who unsubscribes must stop receiving marketing at once. Under Article 21(3), an objection to direct marketing is absolute.
Match the offer to the guest's value
A blanket 20% discount costs you most on the guests who least need it. Use visit frequency from WiFi logins to set the offer instead.
| Guest segment | Definition from WiFi logins | Lead offer | Avoid |
|---|---|---|---|
| One-time visitor | 1 login in the past 12 months | Low-cost return incentive, such as a free drink with a main | Deep discounts that attract one-off deal seekers |
| Occasional guest | 2 to 4 logins in the past 12 months | Off-peak offer that fills quiet sessions | Offers valid on your busiest nights |
| Former regular | 5+ logins in the past 12 months, then none | Personal recognition: a manager's note or a tasting invitation | Generic percentage discounts |
| Repeat hotel guest | 3+ stays in the past 24 months | Late checkout, upgrade or breakfast on a direct booking | Rate cuts below your published direct rate |
For hotels, prefer added value to a lower rate. Breakfast or an upgrade keeps your published rate intact. The guest sees the benefit at the property, not on a price comparison site.
For restaurants, aim the offer at quiet sessions. A free dessert on a Tuesday fills a seat that would otherwise sit empty. The same offer on a Saturday gives away margin on a table you would have sold anyway.
What does a win-back sequence look like in practice?
The two scenarios below use illustrative figures, not Purple benchmarks. Replace them with your own guest counts, return rates and spend.
A 120-room city hotel winning back business travellers
The situation. A 120-room city-centre hotel has 6,000 opted-in contacts from guest WiFi logins over two years. Its business travellers return every 60 days on median. Of those contacts, 1,800 last logged in between 120 and 180 days ago.
What was done. The hotel held back 180 contacts (10%) and sent the three-step sequence to the remaining 1,620. Email 1 announced refurbished rooms and a new co-working lounge. Email 2 offered late checkout and breakfast on direct bookings. Email 3 repeated the offer and asked guests to confirm their preferences.
The outcome. Within 30 days of email 3, 81 of the 1,620 guests (5%) had logged in again. In the holdout, four of 180 (2.2%) returned with no email. The incremental return rate was 2.8 points, or about 45 stays the sequence caused. At an assumed £150 per room night and 1.5 nights per stay, that is roughly £10,100 in incremental room revenue. Guests who returned left the sequence early, so about 4,700 emails went out. Revenue per send was therefore about £2.
A 15-site pub group with guests who move between venues
The situation. A pub group runs 15 sites in one city and holds 40,000 opted-in contacts. Measured venue by venue, 9,000 guests have passed the 60-day lapse threshold at their usual pub.
What was done. Before sending, the group checked WiFi logins across the whole estate. It found that 2,100 of the 9,000 had logged in at another of its pubs in the past 60 days. Those guests had not lapsed from the brand; they had changed venue. The group suppressed them and sent the sequence to the remaining 6,900, minus a 690-contact holdout. The offer was a free dessert with any main, valid Monday to Thursday at any site.
The outcome. Suppression removed 23% of the list before a single offer went out. Without it, 2,100 active guests would have received a discount for visiting pubs they already used. The group counted a return login at any site, not only the original venue. Over 30 days, 6.5% of the 6,210 guests in the send group returned, against 3% of the holdout. That is an incremental 3.5 points, or about 217 extra visits. At an assumed £28 average spend per visit, the sequence added about £6,100.
How do you know win-back emails are working?
Measure return visits, not opens
The primary metric is the incremental return rate. Take the share of the send group that logs in again within 30 days. Subtract the same share for the holdout. A WiFi login is a visit you can see, which an open is not.
Opens are a weak signal for this job. Apple's Mail Privacy Protection, introduced in 2021, loads email content in the background for Apple Mail readers who switch it on. That records opens that may never have happened.
Read the return rate as a floor
WiFi logins only show guests who connect. A guest who returns and stays on mobile data will not appear. Your measured return rate is a floor. The holdout comparison stays fair, because both groups are measured the same way.
Watch the guardrails
- Spam complaint rate: Google asks bulk senders to stay below 0.10% and never reach 0.30%. Pause any sequence that pushes you towards that line.
- Unsubscribe rate: compare it with your regular campaigns. A win-back unsubscribe is not a failure. It removes a contact who was not going to return.
- Offer cost per incremental visit: divide total redemption cost by incremental visits. If the result exceeds your margin on a visit, change the offer.
Run one test at a time
Change the lapse threshold, the offer or the gap between emails, never all three together. Keep the holdout running for every test, or you lose the baseline.
What goes wrong, and how do you avoid it?
- Discounting regulars. Mixing former regulars and one-time visitors into one send gives your most valuable guests the cheapest message. Segment by login count first.
- Defining lapse by calendar, not by guest. A fixed 90-day rule is too slow for a coffee shop and too fast for a resort. Use the median return interval for each venue type.
- Counting staff as guests. Staff devices inflate visit frequency and shorten your intervals. Exclude them before you calculate anything.
- Ignoring the rest of the estate. In a multi-site group, check logins at every venue before you call a guest lapsed, as the pub group above did.
- Emailing without recorded consent. Only guests who opted in to marketing belong in the sequence. Under GDPR Article 7(1), the burden of proving consent sits with you.
- Never stopping. A sequence with no sunset keeps mailing dormant contacts. That raises complaint risk and holds data longer than you need it.
- Undercutting your own direct rate. A hotel offer that beats your published rate teaches guests to wait for the email before they book.
Frequently asked questions
Do we still need Purple Engage if we already use Mailchimp or Klaviyo?
Yes, if you want win-back sends timed on real venue visits. Mailchimp and Klaviyo send campaigns, but they do not build a list from the guests who walk through your door. They also hold no record of a diner's last visit. Purple Engage builds a consented list from guest WiFi logins and uses each guest's last login date to trigger the sequence. Purple also integrates with CRM and marketing platforms.
Does Purple work with the WiFi hardware we already have?
Yes. Purple is hardware-agnostic and runs as a cloud overlay on access points from Cisco Meraki, HPE Aruba, Ruckus, Juniper Mist, Ubiquiti UniFi, Cambium, Extreme and Fortinet. You keep your existing network and add Purple's login and opt-in experience on top. Every login then records a visit date your win-back sequence can use, with no rip and replace of hardware at your venues.
Is it GDPR compliant to email guests who logged in to our WiFi?
Yes, provided each guest gave marketing consent at login and you can prove it. GDPR requires consent to be freely given, specific, informed and unambiguous, and Article 7(1) puts the burden of proof on you. Purple's conscious-choice opt-ins separate marketing consent from accepting the WiFi terms. Guests who connect without opting in receive no marketing. Purple is ISO 27001 certified and GDPR compliant.
How many emails should a win-back sequence contain?
Three is a practical default. The first is a reminder with no discount, the second a value offer around 14 days later, and the third a final offer with a preference check. Each email goes only to guests who have not logged in again since the last send. After the third, stop marketing sends to anyone who neither visits nor engages, which protects your sender reputation.
Should a win-back offer always be a discount?
No. Start with a reminder that carries no offer, because some lapsed guests only need a prompt. When you add an offer, match it to the guest's history. A low-cost incentive fits one-time visitors. Former regulars value recognition, such as a tasting invitation. For hotel guests, choose late checkout, an upgrade or breakfast over a rate cut that undercuts your published direct price.
How does Purple know a lapsed guest has come back?
Purple sees the return when the guest logs in to your guest WiFi again. In a multi-site estate, a login at any of your venues counts, so a guest who switches site is not treated as lapsed. Guests who return without connecting will not show up. Treat the measured return rate as a floor and compare it against a holdout group measured the same way.
How much login history do we need before we start?
You need at least one full return cycle for your venue type. A coffee shop or quick-service restaurant typically has enough data after eight to 12 weeks of logins. A business hotel needs around six months. A leisure hotel or resort needs more than a year, because many of its guests return annually and the first email must land before they rebook.