The Message That Breathes
A brewery with deep local loyalty and fading momentum replaced its promotional calendar with messages matched to each guest’s motivation and timed to their own rhythm. Revenue rose 32.6% in the first 90 days, and the next quarter ran higher still.
Revised edition of a post-pandemic engagement. Client name withheld.
The numbers
- +41.1%
- revenue in the following quarterSecond consecutive quarter, same baseline
- +18.4%
- repeat-purchase frequencyOver the campaign period
- 7.3 → 8.6
- Net Promoter ScoreOver six months
- +22%
- email open rateCross-channel read; other channels were not cannibalized
- 15–20%
- engagement above baseline in historically weak monthsSeasonal troughs flattened
Revenue, indexed to the pre-intervention baseline
Pre–post comparison across two consecutive quarters; no randomized holdout.
Net Promoter Score, before and after
- Before
- After
Measured over six months, on a ten-point scale.
Show the numbers as a table
| Pre-intervention baseline | 100 |
|---|---|
| First 90 days | 132.6 |
| Following quarter | 141.1 |
| Net Promoter Score | 7.3 before, 8.6 after |
The challenge
The brand carried real heritage and local affection, yet visits had lost momentum after the pandemic. Campaigns ran on a seasonal calendar, and the point-of-sale and loyalty systems held years of history that described what guests had done rather than what they would do next. Messages kept arriving out of step with the moments guests were actually planning a night out.
The insight
Guests were not leaving the brand; they were forgetting it amid crowded feeds and dense local competition. Timing is part of the product: the same invitation lands as hospitality on a Thursday afternoon and as noise on a Monday morning. And guests come for different reasons. Some want something new and an atmosphere to share; others want value they can count on. A message that fits the reason, at the moment the plan is forming, feels like a host remembering you.
What we did
- 01
Turn records into signals
Point-of-sale, loyalty, search and social traces were unified into consented, person-level profiles, so each history could predict the next visit rather than merely record the last one.
- 02
Group guests by why they come
Behavioral models recovered motivational archetypes, each with its own trigger, creative frame and daypart window.
- 03
Let the cadence breathe
Text messages served as a two-way read: the program paused when predicted receptivity fell and resumed when new signals appeared, instead of broadcasting on a schedule.
- 04
Reward at the right moment
Loyalty milestones and seasonal cues were timed to each guest’s habit window, in place of discount ladders, and every message kept the brewery’s own voice.
The cohorts
Each got its own reason, and its own message.
-
The Experience Seeker
Comes for what is new and for an atmosphere worth sharing. Responded to atmosphere-forward stories and first looks at new offerings.
-
The Value Pragmatist
Comes for a reliable, good-value night out. Responded to clear value rather than atmosphere.
-
The Mindful Moderate
Balances a night out with healthier habits. The case doesn’t report this group’s response separately.
How it was measured
A pre–post design across two consecutive quarters after go-live, with the period immediately before as the baseline. Revenue velocity, repeat-purchase frequency and Net Promoter Score were named as primary outcomes before launch, with email and social engagement as cross-channel checks. There was no randomized holdout, so seasonality, local competition and wider economic shifts can’t be fully ruled out; results by archetype and cost curves were not disclosed.
Results
Revenue rose 32.6% in the first 90 days and 41.1% in the following quarter, against the pre-intervention baseline. Repeat-purchase frequency rose 18.4%, and Net Promoter Score moved from 7.3 to 8.6 over six months. Email opens rose 22%, and engagement in historically weak months held 15–20% above baseline.
Results describe a specific engagement and are not a guarantee of future results. Platform-reported figures are labeled as such.