Post Purchase Survey Guide: Questions That Reveal Influencer Sales

A post purchase survey is a short question shown to customers right after they buy, usually “How did you hear about us?”, that records where they first came across your brand. For influencer marketing it catches the sales that links and codes miss: people who watched a creator, searched your name days later and bought without clicking anything.

It is self-reported recall, not tracking. People forget, round off and pick the answer nearest the top. That does not make it useless. It makes it one of three signals, alongside tracked links and codes, that you compare rather than trust individually.

Used well, the survey answers two questions your analytics cannot: which channels start the journey, and whether a creator is driving demand that shows up as direct or search traffic.

Keep the first question short, neutral and about discovery, not persuasion. “What convinced you to buy?” mixes the first touch with the last one, and customers answer whichever is easier to remember.

Only show the follow up to people who pick a creator or social option. Everyone else sees a thank you and leaves. That keeps completion high and puts the extra effort only where it earns detail.

The option list shapes the answers more than the question does. Customers scan, recognise and click. If “Instagram” is the first option and “A friend” is the ninth, you will overcount Instagram.

Randomise the order of the channel options on every load, and keep “Other” and “Don’t remember” fixed at the bottom. Aim for a list short enough to read on a phone without scrolling, and separate the platform from the person: “Instagram ad” and “A creator I follow on Instagram” are different channels with different budgets behind them.

Keep a stable core of options for months, so you can compare periods. When a campaign goes live, add it as a follow up option rather than rewriting the main list. Read the “Other” text every week, because it tells you which option is missing.

The order confirmation or thank you page is the default because the customer is still there and still engaged. On Shopify, the help centre says “You can customize your Thank you and Order status pages with built-in Shopify features, apps, and custom apps,” and that “Most customizations are managed from the checkout and accounts editor.”

Add a second placement in the order confirmation or delivery email for customers who skip the page. In India, where many orders are cash on delivery, a delivery message catches buyers after they have actually received and paid for the product. Keep each response tied to an order ID so answers from two placements do not double count one sale.

Never put the survey before payment. A question in the way of checkout costs orders, and answers given in a hurry are worse data.

Build one table with a row per order and four columns: discount code used, tracked link or UTM source, survey answer, and creator named in the follow up. Then classify each order before you draw any conclusions.

For example, suppose a creator’s code shows 50 orders in a month, and another 30 orders with no code or link name that creator in the survey. You can report 50 tracked orders and 30 survey attributed orders as separate rows. Do not add them into one figure without saying so, and do not scale the survey count up to your whole order base unless your response base is large and similar to your buyers.

Recency bias is the biggest. Customers name the last thing they saw, which favours retargeting ads and search over the creator video that started the interest. The first touch wording helps, but it does not remove the effect.

Response bias comes next. The people who answer are not a random sample of buyers, and discount-driven buyers may behave differently from full price buyers. Check whether responders and non-responders differ on order value or new versus returning status before you generalise.

Name recall is weak. Many buyers remember “someone on Instagram” but not the handle. Treat “a creator I follow” as a category signal and creator names as a bonus, and match misspelled handles by hand. Finally, watch small numbers: a creator with four mentions this week and nine next week has not doubled anything.

Review survey data weekly during a campaign and monthly otherwise. Look for changes that line up with creator posting dates, such as a rise in “a creator I follow” answers in the week a set of videos went live.

When survey attributed sales for a creator are consistently large next to their tracked sales, that creator is probably driving demand your links miss. That is a reason to test more, not proof. Confirm it with a stronger test such as a geo holdout or a branded search check before you shift serious budget.

If you run campaigns on Limurse, keep each creator’s tracked and survey attributed results with their collaboration history in the creator CRM, so rebooking decisions do not depend on memory.