Hightouch Capstone Walkthrough

Solstice Skin is invented for this build, along with every row of its data, which you generate in Phase 1. Nothing here touches a real company or a real customer. But the shape of the business is deliberately ordinary, because the interesting problems in a composable CDP deployment come from ordinary businesses.

Direct-to-consumer skincare, founded 2023, based in Austin, eleven people. They sell a 40-SKU line — cleansers, serums, moisturizers, SPF — through their own Shopify storefront and an iOS app, with a small wholesale line into regional spas. About 5,000 customers and 14,000 orders since launch, average order value in the mid-eighties, and roughly 900 active replenishment subscriptions. Acquisition is mostly paid social and influencer, with email and SMS doing the retention work.

Why they called. Every segment they can build lives inside Klaviyo, so it can only see email behavior and whatever Shopify syncs through an app. Subscription state, wholesale orders, app activity and web browsing are all invisible to marketing. When the lifecycle manager wants a list, she files a request, waits a week, and the campaign ships late or not at all. Meanwhile paid is still serving ads to people who ordered on Tuesday, because the suppression list is a CSV someone uploads when they remember.

What they asked for. Marketers building and activating their own audiences on warehouse data. One definition of "VIP" and "lapsed" that holds in every channel. Suppression that runs on a schedule instead of on someone's memory. And a winback program with a real holdout, so they can tell whether it works.

What constrains you. No budget for an enterprise CDP, which is the whole reason composable is on the table. Legal is cautious about consent, so merge rules need writing down and defending. And you are part-time, so anything you build has to survive you not being there on a Thursday.

Fifteen phases, aligned against the course transcripts rather than guessed at. Each phase is labelled with the course it puts into practice, so you can do a phase as you finish the course it belongs to.

Phases 1 and 2 have no course equivalent. Academy assumes you arrive with modeled data already in the warehouse. That's fine for the exam and wrong for real engagements, where the modeling is most of the work and the reason projects stall. They're marked in amber so it's clear they're yours, not the curriculum's — though Phase 2's lean-parent design comes straight from Mastering Customer Studio.

Two phases are amber for the opposite reason. Phase 6b and Phase 14 aren't grounded in any transcript you have — 6b belongs to Customer Studio Use Cases, which isn't in your notes, and 14 is a portfolio step of my own. Treat both as provisional until you've taken the remaining courses.