Guests rarely stop coming back in dramatic fashion. They drift, and by the time you notice, they've already booked their next stay somewhere else.
Bookboost data shows that 71% of guests who return for a second stay never make a third. It's the cliff most groups never see, because the data sits in three different systems and nobody owns the question. Here's why guests drift, and four signals in your data that catch it early.
The reasons are rarely dramatic, because a guest does not "churn" the way a SaaS user does, they just gradually stop choosing you. Here are five patterns explain almost every case:
- Post-stay silence. The guest checks out, gets a generic auto-response, and doesn't hear from you again until a random newsletter three months later. By that point, they've moved on.
- Generic communication. A leisure guest with a partner gets the same email as a corporate solo traveller. Open rates collapse, and the guest learns to ignore the brand.
- No recognition on their return. They came back, but nobody flagged them as a returning guest, so the welcome amenity went to a first-time guest in the suite instead of the loyal couple who had stayed with you three times before.
- Better offer elsewhere. The OTA retargets faster and with more relevant inventory than you do.
- Service issue, no recovery. A complaint at breakfast went into a paper notebook and nobody followed up, so the guest left polite and then booked a competitor next March.
McKinsey's research on personalisation puts a number on it: 76% of consumers feel frustrated when the experience isn't personalised. In hospitality, that frustration rarely surfaces as a complaint, it surfaces as the guest quietly booking elsewhere next time.
.jpg)
Most hotel groups find out too late. The point of a signal framework is to catch drift early enough to act, and each of the four signals below already exists in your data. The work is in connecting and watching them.
The pattern. A guest who used to come every nine months is now at fourteen, and although they have not cancelled, complained, or unsubscribed, they are simply slipping further out, stay by stay.
How to detect it. Calculate average days between stays per guest from PMS stay history, then compare each guest's current gap-since-last-stay to their personal historical average and flag anyone running 1.3x or more above their own baseline. This is not a population average, it is per-guest, because a corporate every-quarter guest and a leisure every-summer guest behave very differently.
What to do when it fires. Move the guest into a recency-triggered winback flow before the gap doubles. The earlier the trigger fires, the less you need to discount to bring them back. By the time the gap is two years, you're competing with someone else's brand affinity, not just a price.
The pattern. The guest used to score 9 or 10, then they scored 7 last time, or they did not respond at all when they previously always did. Cornell's Center for Hospitality Research has shown that management response to guest feedback raises the probability of becoming a loyal guest by roughly 50%. The reverse is worth watching too: feedback that gets ignored, or feedback that stops coming, often precedes drift.
How to detect it. Track the delta between a guest's current NPS and their historical mean, and treat non-response as data rather than as nothing. A guest who answered three of three previous surveys and ignored the fourth is telling you something.
What to do when it fires. A drop from 10 to 7 deserves a personal email from the GM, not a discount code. A non-response after a previously engaged guest deserves a check-in: did the stay meet expectations, is there anything we should know? Bain's classic finding still holds: a 5% increase in retention can lift profits by 25% to 95%.
.jpg)
The pattern. The guest used to open every newsletter and click roughly 12% of the time. Over the last four sends they've opened twice and clicked nothing, a sign your emails are becoming background noise.
How to detect it. Set a personal engagement baseline per contact across a rolling 90-day window and watch the trend rather than a population threshold. The MailerLite 2025 benchmark report puts travel and transportation at a 30.10% open rate, the lowest of any industry it tracks and well below the 43.46% all-industry average. Hospitality email consistently underperforms other categories on open rate, which is exactly why a per-guest trend matters more than a flat benchmark.
What to do when it fires. Reduce frequency before the unsubscribe, then switch the guest into a content track that matches what they actually clicked on historically, whether that is spa, family, restaurant, or business travel. Ask them to update their preferences rather than send another generic broadcast.
The pattern. This is the most expensive signal in the framework. The guest came back, they just came back through Booking.com, and you paid the OTA an 18% commission for a guest who was already yours. Worse, you've lost the email address, the marketing consent, and the chance to keep the relationship next time.
How to detect it. Match new OTA reservations against your existing CRM profile by name, phone, date of birth, and stay history, because a returning guest booking via OTA is a flag, not a win.
What to do when it fires. Two moves. First, on arrival, connect the OTA booking to the guest's existing profile in your system and get their consent to communicate directly. Second, post-stay, prove the value of booking direct next time with room upgrade availability, late check-out, or restaurant credit. The goal is to bring the next stay back to your direct channel.
A signal framework only works inside a lifecycle that can act on it. Bookboost data shows lifecycle-triggered emails average 50% open and 14% CTR, while the same brands sending generic broadcasts average 33% open and 4% CTR. Pre-arrival messages, the most contextually relevant moment in the journey, hit 25% CTR: roughly eight times the typical newsletter.
Winback emails are a slightly different story. They average a strong 43% open rate but only a 5% CTR. Lapsed guests are paying attention to the subject line, but the inside of the email isn't giving them a strong enough reason to click. The fix is making that reason specific to the guest, not a generic 10% discount.
A working lifecycle, in order, looks like this: pre-arrival to capture preferences and offer upgrades, in-stay to catch service issues live, post-stay to ask the right feedback question and route the answer to a human, and dormant winback the moment recency drift fires, not three months after the guest has already rebooked elsewhere.
.jpg)
The data already exists. The PMS knows when guests last stayed, the CRM knows when they last opened an email, and the booking engine knows when they last came direct.
The next step is small. Pull a list of guests whose last stay was more than 1.3x their personal average gap and who haven't opened an email in 90 days. That's your winback cohort for the quarter. Send them something specific and see what comes back. To track whether the winback is actually working over time, how to measure guest loyalty in hotels sets out the KPIs to watch.
Want a deeper read on this? Our Guest Relationship Health Check uses the same 6M+ guest dataset to give you a five-question diagnostic plus a self-assessment for your group. Download it here.
Are you ready to increase your revenue and build lasting guest relationships? Take the first step today.
.jpg)
.jpg)
