Incrementality

Incrementality is a measurement concept from advertising rather than a feature of any restaurant system. It becomes answerable in hospitality only once visits are attached to identified guests, because the comparison requires knowing who was messaged and who came in. That join is what a restaurant CRM produces.

How do you know whether a campaign caused a visit?

You cannot know it for one visit, and this is the honest starting point. A single guest who received an email on Tuesday and booked on Thursday may have been coming regardless. Nothing in the data distinguishes the two stories for that guest.

Causation only becomes visible across a group. Take a segment, withhold the message from a randomly chosen slice of it, and compare the visit rate of the two halves over the same period. If the messaged half visits at 12% and the held-back half at 9%, the incremental effect is three percentage points and the other nine were coming anyway. The comparison group is the whole method, and it is the thing a standard campaign report does not have.

Why do restaurant campaign reports look better than the reality?

Restaurant campaign reports overstate results because of who ends up in the numerator. Three effects push the same way.

• Regulars are in every segment. High-value lists select the people most likely to visit for reasons unrelated to marketing. The better the targeting, the more baseline visits get absorbed into the result.
• Engagement gates the credit. Campaign attribution generally starts from an open or a reply, and the guests who open restaurant email are disproportionately the loyal ones. The measurement is biased toward the audience least in need of persuading.
• The window is generous. Any window wide enough to catch a considered booking is also wide enough to catch a routine one.

None of this makes attribution useless. It makes attributed revenue an upper bound on the value of a campaign rather than an estimate of it.

When is it worth testing incrementality?

Incrementality testing is worth the effort whenever the campaign gives something away. A 20% discount redeemed by guests who had already booked is margin handed over for nothing, and it will still show a strong return in an attribution report. Free-to-send announcements matter less, because the only cost of over-crediting them is a slightly flattering slide.

Two practical constraints apply at restaurant scale. The comparison needs enough guests in each half to distinguish a real difference from an ordinary quiet week, so a single venue running a small list will often not get a clean answer from one campaign. And holding a slice of guests back means real people do not receive a real invitation. That is easier to justify on a discount test than on a message the restaurant thinks of as hospitality. Groups with several locations can run the same test across venues and reach a usable sample faster.