Why lifecycle marketing is the next growth lever, what it is worth, and where every journey stands.
The next phase grows how many customers upgrade, and how often: win a customer their first upgrade, turn one win into a habit, step in at the first sign of a slip, and win back the lapsed. Ten journeys in four programmes, each shipped behind a holdout so every claim is measured, building the data and decision rules that become a one-to-one AI engine.
The funnel the board saw on 23 September 2026: the plan, and what a bigger lift or more sign-ups would add.
Counts are Marketing's August 2026 warehouse read; net revenue is Finance's booked August figure. The plan and the three cases are the Q4 2026 Marketing Model as presented to the board on 23 September 2026, at benchmark lift rates: modelled, not measured, with the first holdout reads in November. The 10,000 extra sign-ups are an illustration, not a forecast.
What the programme is worth, journey by journey
Every figure is the Q4 marketing model’s (exported 29 Sep 2026), net to Seatfrog, incremental of the customers who would have returned anyway. Q4 is phased by month from each journey’s launch date in the model: the first slip (Journey 8) and the never-transacted seasonal wave (Journey 9) run the full quarter from 1 October; the losing-bidder, first-transaction and churned-buyer waves (Journeys 4, 5, 9) from 1 November; the first win, the abandoned upgrade and the win-back (Journeys 1, 3, 7) from 1 December. Each programme ladders to one Q4 V2MOM measure; all four ladder to monthly transacting users, 2.24% in August; the Q4 north star 3.3%, the growth model’s own forecast; 4.5% the 2026 anchor.
ACTIVATETurn intent into a first transaction, and keep live bidders in the game.Drives: the first win · V2MOM target: New user to first purchase within 30 days 20.4% (19.8% today); existing user to first purchase 4.8% (3.7% today)
In-app funnel friction work (the 48% bid-step drop-off) is scoped for Q4 as its own programme; until then the funnel-abandon moments sit in this programme’s journeys.
Audience: 415k marketable (1.75M total); 908k on last quarter’s definition. Decides who, with: intent shown: a lost bid, an abandoned upgrade, a first search. New user to first purchase within 30 days: 19.8% today → 20.4% (the Q4 V2MOM measure and target; today from the model’s KPIs tab).
GROWTurn one win into a habit: recognise and reward repeat buying, with offers sized by the propensity-to-return model.Drives: the habit loop · V2MOM target: First-time buyers who bid or buy again within 30 days of the first win 33.6% (26% today)
Audience: about 5k new buyers a month; about 85% reachable after the win via soft opt-in. Decides who, with: the win itself; the propensity-to-return model sizes every credit (the churn-risk score stands in until it is built). First-time buyers who bid or buy again within 30 days of the first win: 26% today → 33.6% (the Q4 V2MOM measure and target; today from the model’s KPIs tab).
DEFENDStep in at the first sign of a slip, while keeping the customer is still cheap.Drives: reactivation · V2MOM target: Month-three retention 18.0% (13.8% today)
Audience: 89.3k marketable (156k total): the current active base. Decides who, with: the churn-risk model, re-scored nightly. Month-three retention: 13.8% today → 18.0% (the Q4 V2MOM measure and target; today from the model’s KPIs tab).
RECOVERWin back the lapsed with contact and offers spent only where the propensity-to-return model says they are needed.Drives: reactivation · V2MOM target: Reactivation rate 1.1% (0.6% today)
Audience: 481.1k marketable (1.92m total): the current lapsed and churned base. Decides who, with: the propensity-to-return score, re-scored daily. Reactivation rate: 0.6% today → 1.1% (the Q4 V2MOM measure and target; today from the model’s KPIs tab).
ACQUIREGrow sign-ups through our own channels: the rebuilt website, boosted organic social, the app store page, search and AI visibility, referral.Drives: efficient acquisition · V2MOM target: monthly new users 30.9k (30.1k a month, July and August)
Four tests, new in Q4 and in neither earlier forecast; Marketing runs no paid media. Sized on installs that register and transact within 30 days; the cost per install and the store-page volumes are assumed until AppsFlyer and the store consoles report.
The test
Basis
V2MOM measure
Customers a year
Revenue a year, net
Oct
Nov
Dec
Q4 2026, driven
Landing in Q4
Where it is
Boosted organic social, the test
Installs a month = the boost budget divided by the cost per install (both on the Inputs tab)
Monthly new users
3,600
£24.8k* £108.0k gross
–
£2.0k*
£2.1k*
£4.2k*
£4.2k*
TEST, FROM 1 NOV
The app store page test
Installs a month = store page views x the extra 2 points of store conversion
Monthly new users
9,000
£62.1k* £270.0k gross
–
£5.1k*
£5.3k*
£10.4k*
£10.4k*
TEST, FROM 1 NOV
Search and AI visibility on the current site
Installs a month = web sessions x the extra 10% organic by December x the measured 10.2% session-to-install (17.39% session to store click x 58.73% click to install; Sam, 10 Sep)
Monthly new users
3,198
£22.1k* £95.9k gross
£0.1k*
£1.8k*
£1.9k*
£3.7k*
£3.7k*
TEST, FROM 31 OCT
The rebuilt website, phase 1
Installs a month = web sessions x the extra 5 points of handoff; today's measured handoff is about 10.2% (Sam, 10 Sep)
Monthly new users
15,660
£108.1k* £469.8k gross
–
–
£3.8k*
£3.8k*
£3.8k*
TEST, FROM 19 DEC
Acquire, as sized
31,458
£217.1k
£0.1k*
£9.0k*
£13.1k*
£22.1k*
£22.1k*
Marketing, as sized: £130k net driven by Q4 launches; £60k lands inside the quartercustomers a year50,205revenue a year, net£697.4k(£3,032k gross)Oct£32.7k*Nov£43.6k*Dec£53.9k*the figures in the Q3 board pack for Marketing and the Q4 V2MOM, 21 Sep 2026
The four lifecycle programmes, as sizedcustomers a year18,746revenue a year, net£480.4k(£2,089k gross)Oct£32.7k*Nov£34.6k*Dec£40.8k*Q4 2026, driven£108.1k*landing inside the quarter£37.9k*the sized floor: about 5,800 additional transacting customers in an average month against the north-star gap
PROPOSED · SCOPE RULING OPENTrip state: disruption and refundswould be an eleventh journey; whether it is in scope, and whose it is, is an open ruling. Sizing follows the ruling.
*Every figure is the Q4 2026 Marketing Model’s, Business Case and Sizing tabs, exported 29 Sep 2026, the planned case (a 30% lift in conversion) as the board saw it on 23 Sep 2026: volumes checked by Analytics (Ivi, 8 Sep 2026); organic and target rates measured against holdout and confirmed by Analytics; a buyer’s month valued at £30 an upgrade, 1.29 upgrades in a first-purchase month and 1.67 in later months (the 20 Sep change; the earlier pre-read’s £172k counted one upgrade a month). Net is gross at the 23% take rate (agreed 8 Sep 2026). The totals and the programme measures and targets are the board’s, as the customer funnel presented on 23 September 2026 states them (the planned case, a 30% lift in conversion); the programme audiences are the Q3 board pack’s (21 Sep 2026); today’s figures are the model’s KPIs tab. Each Sizing row is assigned to one journey once, so the journeys sum to the programme totals. Two Q4 bases, never blurred: driven is the revenue the Q4 launches drive, months active credited to Q4 pro rata by days live, the basis the Q4 V2MOM and the Initiatives tab carry; landing in Q4 counts only the months inside October to December (Tony, 10 Sep 2026), the forecast Marketing submits: £37.9k lifecycle, £60.0k with the acquisition tests. Finance banks about £125k net from lifecycle in Q4; Analytics reconciles the two. Journey 2 is with Analytics, Journey 6 counts inside Journey 1 and Journey 10 is Product-led, so they carry no standalone figure. Audiences: the model’s worked volumes (monthly flows) and the Sept pool; TBC rows name their owner. Stage positions: the delivery record, 22 Aug 2026. The logic, assumptions and formulas: the Q4 marketing model (the live Google Sheet).
Where this goes: foundational to AI 1:1
TODAY
Stage 1: Foundational
“You don't know me”
One size fits all: one message to everyone, sent in batches, with the channels working separately.
In place: the baseline every stage above is measured against.
SHIPS Q4 2026
Stage 2: Segmented
“You know I'm new”
Right stage: the right message for where the customer is in their life with us, across email, push and in-app.
+£14k net a year: seasonal reactivation, sent by life stage. Capability in place (the Senior Data Lifecycle Strategist, joined August).
SHIPS Q4 2026
Stage 3: Behavioural
“You understand my journey”
Right context: journeys triggered by live behaviour, a lost bid, an abandoned search, the first win, the first slip in a routine.
+£73k net a year: four journeys triggered by behaviour (the first win, the abandoned upgrade, losing bidders, the first transaction). Capability in place: every send on one platform, Braze, in real time.
SHIPS Q4 2026 · CHURN PREVENTION LIVE
Stage 4: Predictive
“You're one step ahead”
Right moment: the churn-risk and propensity-to-return scores decide who enters a journey and what they are offered. The first of these journeys, churn prevention, is live.
+£393k net a year: churn prevention and win-back, where the scores decide who enters and what they are offered. Needed: the propensity-to-return score (Analytics) and the live customer-data feed into Braze (Engineering).
Q1 2027
Stage 5: AI 1:1
“My personal travel agent”
Next best action: one decision per customer across every channel, learning from every result.
Not yet sized. Needed: the AI decisioning engine and machine-learning engineering capability.
Stage values: the Q4 2026 Marketing Model's planned case (a 30% lift in conversion), as presented to the board on 23 Sep 2026, net a year at run rate, upgrades only, with 20% of every journey's audience held back as a control; each journey counted under the stage it needs; together £480k (£2.09m gross), and £697k with the acquisition tests (£217k, sized on assumptions); Q4 launches drive £130k net, £60k of it landing inside the quarter and the rest phasing into 2027 with the launch dates. The behavioural and predictive stages ship inside the Q4 decisioning engine rather than as separate phases: every journey ships with a control group and a pre-registered metric so the score-driven versions have a baseline to beat, and the AI one-to-one engine launches in Q1 2027 on that evidence. Stage framing after McKinsey and BCG personalisation research.