Guest Segmentation

Segmentation is standard in email platforms and general customer relationship management tools, where a segment is assembled from list fields and email behaviour. The hospitality version sits in a different kind of product: a restaurant CRM, also called a guest data platform, which builds its segments from the reservation book and point-of-sale checks. That difference decides which questions a segment can answer.

How do you segment restaurant guests?

Restaurant guests are segmented on behaviour that arrives from the systems a venue already runs. Five axes carry most of the segments an operator actually uses.

• Visit behaviour. Visit count, recency, which venue in a group, day of the week, party size, and lunch against dinner.
• Spend. Total across visits, average per visit, and where the point of sale passes item detail, spend by menu category.
• Lifecycle stage. First-timer, second visit pending, established regular, lapsed.
• Stated preferences. Seating, dietary needs, wine style, the occasion that brought them in.
• Other revenue lines. Private events, catering, retail, and hotel stays in groups that run them.

A segment is never better than the connections behind it. A restaurant with the reservation platform connected and no point of sale can group guests by how often they come and not by what they order. This is worth checking before designing a segmentation scheme, because the axes that look most interesting on paper are often the ones the venue has no feed for.

Is a segment the same thing as a tag?

No, and the pair gets conflated constantly. A tag is a label attached to one guest. A segment is a population defined by a condition, so it has a size, it changes as guests move in and out of it, and it exists whether or not anyone has labelled anybody. Tags are frequently the inputs: a segment might be every guest carrying the wine tag who has not booked in ninety days.

The more practical difference is where each one is read. A tag travels well. Some restaurant CRMs write it back into Resy or OpenTable, so it appears on the booking a host already has open at 7:45 on a Friday. A segment is an audience, so its reader is a campaign or a journey rather than a person at a stand. Operators who want the floor to act on something should be building a tag; operators who want to send something should be building a segment.

Why does a segment reach the wrong guests?

A segment reaches the wrong guests when the identity underneath it has not been resolved. One regular who books with a work address on one platform and a personal address on another exists as two profiles holding half her visits each. Those two thin records can land her in the first-time-guest segment and the lapsed segment on the same afternoon, and both look perfectly reasonable from inside the segment builder.

Anonymous covers cause the opposite failure. A segment is computed over matched records only, so a guest whose visits never joined to a profile is absent from every segment rather than misfiled in one. Absence is the harder error to catch, because a segment count gives no signal that anyone is missing from it. The sequence that saves the most embarrassment is to look at the duplicate volume and the match rate first, then at segment sizes, because a segment is a query result and it inherits every gap in the records it runs against.

Can Mailchimp or Klaviyo segment my guests?

They can segment what they hold: email engagement, signup source, and online-order or e-commerce purchase history where a store is connected. Both are capable tools on that data. Neither reads a reservation book or a restaurant’s checks natively, and bridging either one to Resy or OpenTable is a custom integration project. The connectors that do exist tend to pull bookings across as marketing contacts.

So the segment an email platform produces describes a mailing list at the freshness of its last import. The segment a restaurant CRM produces describes a dining room, because it is computed over reservations and checks that have been reconciled against each other. The gap shows up fastest on anything time-sensitive. An email tool can tell you who has not opened a message in six months. It cannot tell you who has not walked through the door in six months, and those are different guests.