Post-purchase attribution: which channels bring in the revenue
Post-purchase attribution means asking buyers where they heard about you, and adding up the answers. Most stores stop at counting them. The more useful step is to add up what each channel’s buyers spent.
Why counting attribution answers isn’t enough
Here are 30 days of answers from a demo store:
Instagram has the most answers: 30, against 22 for Google search. But Google search buyers spent $1,617 in total and Instagram buyers $1,468, because the average Google search order was $73 and the average Instagram order $49. TikTok had the biggest increase on the previous 30 days, +88%, and the smallest orders, at $35.
If you only counted answers, you would put Instagram first and might cut Google. Revenue shows that Google search brings fewer buyers who spend more.
Read answers, revenue and AOV for each channel
| Number | What it tells you |
|---|---|
| Answers | How many buyers name the channel. |
| Revenue | The total of their orders. |
| Average order value (AOV) | Whether the channel brings large or small orders. |
Then look at the change against the previous period. A small channel that doubles in a month deserves more attention than a large one that stays flat.
Survey attribution vs. your ad platforms
Your ad platforms count the clicks and views they can track. The survey records what buyers remember. The two will never match, and the way they differ is useful:
- The survey credits a channel far more than your analytics does. This is typical for word of mouth, podcasts and creator videos. People hear about you there, then search for your name, and your analytics credits search or “direct”.
- An ad platform claims far more sales than the survey shows. The platform may count buyers who saw an ad but bought for another reason, or who would have bought anyway.
A simple check: divide each channel’s survey revenue by what you spent on it in the same period. It is a rough number, because not every buyer answers. Still, if one channel returns $4 for each $1 and another returns $0.80, the difference is too large to be noise, and it tells you where to look first.
Common traps
- Too few answers. Wait for about 30 answers per channel before you compare. A change from 4 to 8 answers shows as +100% and means nothing. Post Purchase Surveys never shows an answer that fewer than 3 buyers gave, so tiny groups don’t appear on their own.
- Not everyone answers. You only see the buyers who chose to answer. If one channel’s customers answer less often, that channel looks smaller than it is. Trends over several months are more reliable than one month’s split.
- Order totals include shipping and tax. Revenue here is the order total after discounts, including shipping and tax. It is fine for comparing channels with each other. It is not your margin.
- Refunds aren’t deducted. The total is taken when the order is placed. If one channel brings many returns, its revenue looks better than it is. Check your returns by channel with the order tags.
- Multiple choice counts an order more than once. If a buyer picks two channels, the whole order counts for both, and the rows add up to more than your real sales.
- Returning customers. Someone on their fifth order may still answer “Instagram”, because that is how they first found you. Add an option such as “I’ve bought from you before” to keep repeat buyers out of your acquisition numbers.
- “Google search” is often the last step. Many buyers heard about you somewhere else and then searched for your name. If search looks very large, add an optional second question: “Where did you first hear about us?”.
Do your own analysis with a CSV export
The built-in view covers the questions above. To combine answers with ad spend, returns or customer data, export them. Post Purchase Surveys has two CSV exports: one with the individual answers it still keeps, with each order’s total and a link to the order, and one with daily totals for your whole history. The Insights docs list every column.
A monthly review
Once a month, compare the last 30 days with the 30 days before:
- Which channel grew or shrank in answers?
- Did any channel’s average order value change?
- What did buyers type under “Other”? Is there a new source to add as an option?
- How does each channel’s survey revenue compare with what you spent on it?
Post Purchase Surveys shows answers, revenue and average order value for every option, compared with the previous period. Install it from the Shopify App Store to see your own channels.