Post-Purchase Attribution After iOS 14: Surveys and Analytics

A post-purchase attribution survey adds the buyer's account of discovery to your existing analytics. It can surface a podcast, a friend's recommendation or an earlier creator mention that is missing from a tracked visit. It does not reconstruct every touchpoint, replace consent requirements or prove which channel caused an order.

The useful question is how to compare these signals well enough to decide what to investigate next.

What App Tracking Transparency changed

Apple introduced App Tracking Transparency requirements with iOS 14.5. Apps need permission for tracking that links data across other companies' apps, websites or offline properties for advertising purposes, including access to the advertising identifier. Apple's privacy and data-use documentation explains the boundary.

That is a specific limitation on tracking. It does not mean every iOS order is invisible, every analytics event is wrong, or all browsers follow the same cookie policy. Your observed data also depends on implementation, consent choices, attribution windows, device changes and the path the buyer took.

Before attributing a discrepancy to a privacy change, check that your store and advertising reports use comparable dates, time zones, conversion definitions and attribution settings. A report about sessions is not directly comparable to a report about orders or to answers about first discovery.

Ask about the part of the journey you mean

“Where did you first hear about us?” and “What most influenced today's purchase?” ask different questions. Choose the one that supports your decision and keep the wording stable during the comparison period.

Pew Research Center's question-writing guidance explains why wording, answer options and order matter. Use understandable choices, an “Other” option and, where appropriate, “I don't remember.” A broad “Instagram” answer may need a follow-up to distinguish an ad, a creator and a friend's message.

The discovery-question guide covers setup; the question change log helps preserve comparisons when you revise the survey.

A disagreement can describe two true observations

Consider this illustrative example, not customer results:

1. A buyer hears your brand on a podcast.
2. Days later, they search for your store and place an order.
3. The recorded visit comes from search, while their survey answer says “A podcast.”

The survey describes recalled discovery and analytics describes a recorded interaction. Neither observation requires the other to be wrong. The buyer might also have forgotten an earlier exposure, and a different attribution model might assign credit differently.

Use the disagreement to ask a narrower question: is the podcast bringing awareness that appears later as branded search? Then look for supporting evidence, such as timing, relevant landing-page visits, code usage or a properly designed experiment. The survey alone cannot establish the incremental revenue.

Use a consistent weekly comparison

Keep a small record rather than trying to force every source into one number:

• Survey: question, audience, page placement, date range, number of responses and channel mentions.
• Store analytics: report name, date range, time zone and whether the denominator is sessions, orders or another measure.
• Advertising report: attributed conversions, spend, attribution window and the event being counted.
• Decision: the discrepancy to investigate, other evidence needed and the next review date.

Suppose 18 of 100 answering orders mention a podcast. That is 18% of the answering orders in this illustrative example. It does not mean the podcast caused 18% of all orders, and it does not justify allocating 18% of store revenue to that channel.

People who answer voluntarily may differ from those who skip the question. More responses can make a pattern worth investigating, but do not automatically make the group representative. Keep counts alongside percentages. If you change the question, audience or placement, mark that change before comparing periods.

For channel-level calculations and their limits, use the attribution and ROI guide. For early results, see what to do with your first survey answers.

Put the question on the right Shopify pages

OrderSurvey adds blocks to Thank you and Order status after an order. Shopify supports these extension surfaces on all plans except Starter; information, shipping and payment-step extensions have different Plus requirements. See Shopify's checkout customization overview.

Enable the survey for the pages you want, add the block to each page in the checkout editor, save, and verify both placements. Keep the question optional and the order information easy to use. A native extension still loads content and makes network requests; it is not a guarantee of zero latency. The performance guide gives a practical check.

Start with All orders when it suits your research. First- or repeat-order audiences are available, but restricted Order status surveys rely on a classification previously captured on Thank you; unknown historical orders are excluded. Other rules depend on the order values available to the block. Customer-tag targeting is unavailable on these two pages. Check the targeting guide before comparing audience segments.

Review results before deciding to upgrade

OrderSurvey's Free plan includes analytics, response review, unlimited surveys, conditional branching and targeting rules, with 100 responses across the store per rolling 30 days. The dashboard remains visible before the first answer, with setup guidance above it.

Unlimited is $24 USD/month and adds unlimited responses, CSV export, configured low-NPS webhook alerts and priority support. It is useful when you need the additional capacity or a recurring export workflow; collecting attribution answers does not itself require Unlimited.

The CSV contains question-and-answer rows and order references, with no dedicated customer-email column. A multi-question response produces multiple rows. It does not automatically connect survey answers to an ad platform, create email segments or calculate causal return on ad spend. See pricing and the permissions and data guide.

Survey answers also do not override a buyer's privacy choices. Keep participation and marketing opt-in separate, and use collected information for the stated purpose; Shopify's app privacy requirements explain the relevant responsibilities.

Install OrderSurvey, follow the Thank you page setup guide, or book setup help. Start with one question you can use, then compare the answers with the evidence you already have.

Frequently asked questions