Capture the Data at the First Purchase, Not the Third

Every personalisation project we've watched stall has stalled in the same place. Not the software, not the email platform, not the budget. The store simply doesn't know anything about its customers. Three orders in, the record says a name, an address and a purchase history, and the team is left inferring the important things — is this a puppy or a senior dog, a newborn or a toddler, dry skin or oily — from what someone happened to buy.

The frustrating part is that the customer would have told you. You just never asked, and by the time you want to, the easy moment has passed.

The willingness window

There's a short window where a customer is unusually happy to answer questions: right around the first purchase. They've just chosen you. The order is on its way. A question at that moment reads as service — of course the shop selling puppy food wants to know how old the puppy is. The same question six months later, cold, in a marketing email, reads as data collection, because that's what it is.

"The same question six months later, cold, in a marketing email, reads as data collection, because that's what it is."

The categories where this matters most are the ones where one fact predicts the entire future relationship. The pet's breed and age tell a pet store every reorder and life-stage transition for the next decade. The child's birth month does the same for a kids' brand, the vehicle for an auto-parts store, the skin type for skincare. This is context a customer gives you directly and deliberately, and it's worth more than a year of inferred behaviour, because it doesn't need interpreting.

Where to ask on Shopify

The good news is the plumbing for this is now standard Shopify, not a custom build.

The order status and thank-you pages take app blocks and checkout UI extensions, which is exactly where a one-question survey belongs. The sale is done, nothing you ask can cost you the conversion, and attention is still on the screen. One question, not a form.

Shopify Forms writes answers to customer metafields, and since early 2025 it syncs with merchant-owned metafield definitions, so a dropdown you define in custom data stays consistent everywhere the question is asked. That means the answer lands on the customer record as structured data you can act on, not as a note in a third-party app you'll export once and forget.

And the account itself: if you run customer accounts, registration and profile are natural places for the fields that don't fit at checkout.

Wherever you ask, attach the reason to the question. "Tell us your dog's birthday and we'll send something for it" gets answered. An unexplained field doesn't.

What the answer unlocks

Structured answers on the customer record are what turn generic lifecycle emails into ones that read like you're paying attention. Replenishment timing based on the actual animal rather than the average customer. A size-up prompt that arrives when the child would plausibly have grown, not on a fixed 90-day timer. Segments built on the metafield, so the winter-coat campaign goes to the huskies and not the whippets.

It's also the foundation everything more ambitious rests on. We've made this argument at length in "Why the AI personalisation project should be last, not first": a recommendation engine can only work with the data underneath it, and this is the data. Collecting it at the first purchase is the unglamorous project that makes the glamorous one possible.

Three rules so it doesn't rot

Only ask what you'll use inside ninety days. A question with no email or automation attached to the answer is friction spent on nothing, and a growing pile of unused personal data is a liability, not an asset.

"A growing pile of unused personal data is a liability, not an asset."

Store it structured. A metafield with defined options can drive a segment. The same fact typed into an order note cannot.

One question at a time. The willingness window is real but narrow, and a five-field survey on the thank-you page burns it. You can always ask the second question at the second purchase.

Go and measure it

Two checks. Pull up twenty recent customer records in your admin and count how many carry even one structured fact beyond name, address and order history — for most stores the honest answer is zero. Then place a test order on your own site and note every point between add-to-cart and the order confirmation email where you were asked anything about yourself. If both counts come back empty, the first question you add is probably the highest-leverage field on your site. Deciding what that question should be for your category is exactly the sort of thing we're happy to talk through.