#	The asset	New, or a change to something live	What we need from Design	Every version to write
1	The value-proof push PHASE 1 Its send is on the CRM brief. PUSH reaches them just after the trip ends	NEW Nothing fires at the reflection moment today.	COPY AND LAYOUT Title and body, two registers.	Returning winner: You saved {saved} on that trip · First Class on {route} for less than the walk-up price. That’s it: no ask, just a good trip First-trip winner: Your first upgrade paid off · {saved} under the walk-up price on {route}. That’s it: no ask, just a good trip ▸ The job, the copy and where it lives Where the copy lives BRAZE A Braze Canvas send on the journey's entry clock. The job to be done Convert post-trip emotion into proven value, ninety minutes after arrival. Never on a disrupted trip, never to a power user, and never an up-to figure: a missing saving holds the send. Copy direction, not final words Prove the win with the customer’s own figure, then stop. The absence of an ask is the design.
2	The day-after email PHASE 1 Its send is on the CRM brief. ▸ Today, live: the creative this reworks Today, live The full record ↗ EMAIL reaches them the day after the trip	NEW The journey's own trip-anchored email; the existing day-after survey stays untouched beside it.	COPY AND LAYOUT One master email, three job variants by Liquid, one subject line each.	The return-leg bid: Heading back on {route}? · upgrades live from {from_price} on your return leg · CTA See your return upgrades The earned referral ask: Give £5, get £5 · first-trip register leads, the {n}-trips line is the repeat variant · CTA Share your link. The reward-timing line is contingent on the reward-delivery fix The savings tracker default: Your Seatfrog savings so far · {total} saved · CTA Plan your next trip ▸ The job, the copy and where it lives Where the copy lives BRAZE One master campaign, Liquid-forked, in the journey Canvas. The job to be done Carry the slot's one ask, or its quiet default, without competing sends. The referral variant fires from the first completed first-class trip, positive-trip gate passed, once per reward window, never at the win itself; the return-leg variant stands down once the pre-travel journey owns the trip; an unsubstantiated figure holds the line, never “up to”. Copy direction, not final words One email owns the day-after slot and arbitration picks its job: the return-leg bid when one is live, else the referral when the gate passes, else the tracker. The three variants must read as one family. A fourth variant, the review ask, is a consider-testing mock only: it is not in the proposed strategy and nothing fires until a test says otherwise.
3	The re-booking nudge PHASE 1 Its send is on the CRM brief. ▸ Today, live: the creative this reworks Today, live The full record ↗ PUSH reaches them on their own purchase rhythm	PART NEW The push is new; the built post-trip nurture Canvas and the Route Watch email already exist and are re-anchored, not re-authored.	COPY ONLY Title and body; the second-knock email reuses the Route Watch email verbatim, no new design.	Default: Heading back on {route}? · Set a Route Watch and we’ll tell you the moment an upgrade opens on your line Return-pattern timing variant: the same copy, fired at 48 hours ▸ The job, the copy and where it lives Where the copy lives BRAZE The switched-off nurture Canvas is the built vehicle; its control machinery already works. The job to be done Seed the next booking on the customer’s own rhythm, before the month-one cliff. Floor at day three: an earlier send claims credit for the return leg. Never to power users; silent when the next trip is booked. Copy direction, not final words Point at the existing Route Watch mechanic on the customer’s own route; effortless, never salesy.

If it is not in these two tables, it is not in this round. The survey repair, the referral stand-downs and the reward fix are rule and engineering work with nothing to design.

One job per message, and the value proof carries no ask of any kind. Never an up-to claim: the saving is the customer's own, true and dynamic, or the message holds. No discounts anywhere in this journey. Push titles stay under forty characters and bodies under one hundred and twenty five; email subjects under forty five with a preheader and one button. Copy is timeless: no dates, seasons or offers. Persona sets tone only, never cadence.

The design question. We get one glance 90 minutes after arrival, one email the day after the trip, one nudge weeks later. What should the value proof feel like so the win lands as THEIR win, and the one ask that follows feels earned rather than extracted?

The split. We hold the frame (spine, arbitration, triggers, channels, measurement), you own the experience and creative, and this round can still challenge the spine before Gate 1.

The voice: inherited, not re-tested (one headline test per journey release). Version A (brand voice) and Version B (frog) run through the walkthrough as copy exploration; Journey 8’s test carries the evidence and the guardrail read.

Format rules, every asset:

• Two data states, with a harder fallback than usual: no substantiated figure means a HELD send, never generic copy (why D2 is launch-blocking).
• Push: title ≤40 ch, body ≤125, no CTA line (the tap deep-links). Email: subject ≤45, preheader, one button. The For-you card: headline + 2 lines + the opt-in action, kept clear of any rating ask.
• Timeless: no dates, seasons or offers. Savings recall may appear as supporting copy in the day-after email, never as a second CTA.

What to produce. The assets below, each under its brief. Variants differ by register and test arm, never by job.

Open design decisions: the opt-in framing on the For-you card (does the ask depress the proof? guardrail in the headline test) · the frog’s presence in a value message · per-persona tone (D7 routes copy at confidence, generic below threshold). What good looks like: the value proof needs no tap to succeed; every other asset has one CTA and one job.