Case Study · Pinucci Ltd · 2010-2017
Founder · Marketing Automation Owner
Lifecycle marketing: segmentation, campaign automation, suppression and frequency governance.
The Brief
The hard part was the category's economics. Acquisition in luxury is expensive; the business only works if customers return for the next collection. In the early years, campaigns were generic, the same newsletter to the whole list, timed to the marketing calendar rather than the customer's behaviour. Engagement was flat, and nothing could be reliably attributed. The brief, set by the founder for the founder: replace broadcast with governed segmentation and behaviour-triggered journeys, build the content engine to feed them, and measure whether retention actually moved. This was 2010s rule-based automation, hand-built, long before AI entered the marketing stack.
The Approach
Customers were segmented by value and behaviour on unified online and in-store purchase data (the collectors separated from the one-time gift buyers, the engaged from the lapsing) with segment definitions written down and kept stable so a retention number meant the same thing season after season. Journeys were built around the moments that mattered, with frequency caps and suppression rules protecting the brand's positioning. Because creative direction and automation sat with the same person, the content calendar and the journey logic were designed together. Results were read at cohort level, in a weekly loop of insight, change and re-measurement.
Fig. 01 · Four journey families, governed
The automation runs the loop so the customer comes back.
All journeys run with manners: frequency caps and suppression rules throughout, based on global luxury brand patterns.
Mailchimp journeys between 2010 and 2017 did by rule what Journey Builder and Account Engagement do by configuration today: segmentation, triggered journeys, suppression and frequency logic, sustained across seven seasonal cycles before any of it was native.
Recreated from the journey architecture. Segment definitions and volumes withheld.
Judgement Call
Retention and campaign efficiency were the numbers the business was steered by, so their definitions were fixed and guarded: what counted as a retained customer, what counted as a conversion, which channel got attribution. The lasting lesson, applied in consulting work since: be sceptical of metrics that flatter (including ones you designed yourself) and build benefit cases that must survive a two-year horizon, not a launch quarter.
What Was Delivered
- A value-and-behaviour segmentation model on unified purchase data, with stable, documented definitions
- Four families of triggered lifecycle journeys with frequency caps and suppression rules
- A content engine (seasonal newsletters, digital lookbooks, campaign creative) aligned to the collection rhythm
- A/B testing where volume allowed; cohort-level retention analysis; revenue attribution by journey
The Outcome
Retention improved 10% year on year and campaign efficiency 15%, sustained across seven years of collection cycles rather than achieved once. Every rule was written by hand, and what AI-era tooling automates today still depends on exactly the design questions this programme answered manually: who should hear from you, when, about what, and how you will know if it worked.