A lifecycle programme that speaks to the right shopper at the right moment.

01

The brief

ASHGROVE already had a brand, a website, a design system, newsletters and a demo shop. What it lacked was the part that makes a brand behave like a business: the right message reaching the right person at the right moment, without anyone doing it by hand.

The goal was to design and build that in Klaviyo for a fictional fashion brand, document every decision, and be able to show it without ever emailing a real customer. I am a non-coder, so the whole programme was built by describing outcomes to Claude, checking each result myself in the Klaviyo dashboard, and approving every change before it happened.

The strategy in plain words

A lifecycle programme works out what a shopper has done and sends the one message that fits that moment. I started with four objectives: welcome new subscribers, recover abandoned carts, bring first-time buyers back for a second order, and win back lapsed customers. These four come first because each sits at a point where a customer is easiest to lose and a small, well-timed message is most likely to help. Everything else, such as a birthday email or a review request, can wait until these work.

These are the numbers a real programme would measure. None are claimed here, because nothing was sent to a real customer and the shoppers are invented.

02

Built safely: a demo that cannot hurt anyone

A marketing system that sends email is the one kind of project where a small mistake reaches a real person. So the safety rules were designed before anything else was built.

• Everything starts as a Draft. A Draft email or flow can never send. Nothing was switched on at any point.
• Only four inboxes may ever receive anything, and all four are mine. The other 82 shoppers have no email address at all.
• A seatbelt on every journey. A custom property, demo_send_allowed, must be true for a profile to get into a flow. Campaigns can't filter by property, so the campaign audience is a list containing only the four owned inboxes.
• Approval before every change. Claude showed me the exact change before writing anything through the Klaviyo connector: imports, events, test sends, new lists.
• Test sends only to my own inboxes, logged one by one in a send ledger (8 of a 100-send budget used, previews not counted).

03

The data foundation

A lifecycle programme needs customers with a history. I invented one, on purpose, and labelled it as invented everywhere.

A lesson that shaped the design: backfilled history does not trigger flows. Klaviyo only starts a journey from events that happen after it is live, which is why the shop (chapter 02 of the main case study) sends real events of its own.

Sending domain 3 screenshots

Importing the shoppers 10 screenshots

04

Five lifecycle stages

Segments sort shoppers by what they have done. Each is a live guest list: people move in and out as their behaviour changes.

Building the five stage segments 12 screenshots

The journey map

The same five stages, left to right, with the goal of each and the flow that speaks to it. The hatched columns are gaps: stages with no flow today. Counts are from the recount on QA day.

Never ordered is the biggest stage (29 of 85 staged shoppers) and nobody speaks to it as a stage. Welcome and Abandoned cart catch some of these shoppers at a moment, but nothing is aimed at the stage itself. Fading has no flow at all.

How I would close the two gaps

These two flows are a design only. Nothing above is built in Klaviyo and none of it has been tested. The four built flows are Welcome, Abandoned cart, Post-purchase and Win-back.

The top of the funnel is the concept social creative described in the Campaigns chapter: posts bring people to the shop, the shop's tracking tells Klaviyo what they did, and these flows take over.

05

Four journeys, built step by step

Each flow has a trigger (what starts it), a bouncer (profile filter), a re-entry rule (can the same person go round again) and emails with waits between. Triggers come in kinds, which is worth knowing before building: added to a segment, added to a list, or a metric such as Placed Order.

Decisions worth explaining

• Abandoned cart waits one hour, then checks again. Someone who bought in that hour should not get a reminder. The split looks for a Placed Order since the flow started; if found, the journey ends silently.
• Post-purchase re-entry is 90 days. A customer who buys again after three months deserves the care guide again; one who buys twice in a week does not.
• The Welcome list has no public sign-up form. A public form on a demo shop would collect strangers' addresses, which is exactly the risk this project avoids. People enter the list only when I add them.
• The trigger can't be changed after saving, but filters can. That made the trigger the first decision to get right on each flow.

Win-back flow, step by step 14 screenshots

Post-purchase flow, step by step 10 screenshots

A Draft campaign to the four-inbox list 5 screenshots

Abandoned cart flow, step by step 12 screenshots

06

Personalisation that survives missing data

The second Welcome and Post-purchase emails read the shopper's preferred_category and change the product, price, photograph and button. The important part is the fallback: a shopper with no name or no category still gets a complete email.

Hi {{ first_name|default:"there" }},

{% with affinity=person|lookup:'preferred_category'|default:'tees' %}
  {% if affinity == 'knitwear' %}   The Ribbed Crew, $128
  {% elif affinity == 'outerwear' %} The Work Jacket, $218
  {% elif affinity == 'accessories' %} The Belt, $68
  {% else %}                          The Heavy Tee, $38
  {% endif %}
{% endwith %}

Hugo · knitwear

Ribbed Crew, $128

"Hi Hugo,"

Sam · outerwear

Work Jacket, $218

"Hi Sam,"

Blank profile

Heavy Tee, $38

"Hi there,"

Two practical lessons: Klaviyo's template language uses elif, not elsif; and the template library and editor show the raw tags. Only the Preview window renders them, so a "broken-looking" template in the library is expected, not a bug.

Welcome flow, step by step 13 screenshots

07

Connecting the shop

Flows can only react to what they hear. The demo shop reports three events to Klaviyo (Viewed Product, Added to Cart, Started Checkout) and identifies the shopper by an email that already exists. Placed Order is not sent from the browser; it comes from my connector, since a real order would come from the store's back end.

Because the public shop records no anonymous visitors and has no sign-up form, I needed a safe way to be a customer. The picker lets me play Rajiv (new subscriber), Sam (cart abandoner), Leo (repeat buyer) or Hugo (lapsed) with one click, and everything I do lands on that person's profile.

The picker only shows with ?demo=1 in the address. Normal visitors see a clean store.

A finding on the way: some browser content blockers stop Klaviyo's tracking script. It worked in Safari and not in a blocker-heavy Chrome, so tracking can silently fail for real visitors too.

Shop events and the demo picker 2 screenshots

08

How I tested it

Nothing was trusted until it had been shown to work, in three layers.

Testing evidence 2 screenshots

10

What is real, and what comes next