What “How did you hear about us?” is really for (hint: not reconciliation)
Chinmay Raibagkar · reads survey answers the way pollsters do — skeptically🗳️ with survey rules, not a lecturePublished August 28, 2026
The 60-second version
- Surveys measure memory, pixels measure clicks. They will never reconcile — and they shouldn’t.
- Memory leans in predictable ways: last-touch recall, big brand names, whatever the customer heard most recently.
- The fix: use surveys to spot untracked channels and pixels to count — route each to its own decision.
Somewhere in your post-purchase flow there is a dropdown that asks: "How did you hear about us?" Roughly 40% of customers answer it. Of those, a large share pick "Google" when they came from Instagram, "a friend" when they came from a podcast, and "Instagram" when what they actually mean is "somewhere on my phone".
Then you compare survey answers to your pixel data and the two disagree wildly. Someone concludes the survey is useless and suggests deleting it.
The survey is not broken. It is simply answering a completely different question from your pixel, and the mistake is expecting them to reconcile. They cannot, and the reason is more interesting than the disagreement.
Two instruments answering two different questions
What Each Instrument Actually Measures
Data JourneyThe Pixel
Records the last technically observable digital touchpoint before checkout. Precise about what it sees, blind to everything it cannot.
The Survey
Records what the customer remembers as the reason they know you. Imprecise, unreliable on detail, and covering channels no pixel touches.
The Gap Between
Not error. The systematic difference between the two is itself the most useful output — it locates demand your pixel structurally cannot see.
Your tracking pixel answers a question about mechanism: which link or ad campaign loaded immediately before this purchase. It is extremely accurate about technical events and completely blind to everything else.
Your survey answers a question about memory: what does this customer believe introduced them to your brand. It is unreliable on technical specifics, but uniquely capable of surfacing touchpoints no software can track — a podcast episode, a friend's recommendation, a billboard, or a WhatsApp group chat.
Neither tool is a broken version of the other.
The reconciliation trap: Teams often try to score the survey against the pixel: "63% of people who selected Google actually came from Meta, so the survey is inaccurate." That framing assumes the pixel is ground truth for a question it does not answer. A customer who discovered you on a podcast might later search your brand on Google and click an ad. All three touchpoints happened. Only the last click lives in your pixel.
Why survey answers lean in predictable directions
Customer survey responses are not random noise; they carry known, predictable biases.
- Recency bias: People recall the most recent touchpoint they consciously remember — which is rarely the ad that first created their desire. Someone who saw twelve Instagram ads over three weeks and finally searched your brand on Google reports "Google Search".
- Brand-name substitution: Customers use familiar names as generic categories. "Google" often means "the internet." "Instagram" frequently means "my phone."
- Options bias and social desirability: Whatever options you list get selected. Adding "TikTok" immediately generates TikTok responses, even from buyers who have never used the app. Meanwhile, "A friend recommended you" is a flattering answer that costs nothing to select. "Other" and "I don't remember" absorb uncertainty; removing them just forces shoppers to guess.
- Non-response bias: The 40% of buyers who take the time to answer are more deliberate and engaged than average. Fast impulse buyers, who are heavily driven by paid social ads, skip the question.
Where the two instruments genuinely disagree
Your survey says 14% of buyers heard about you on a podcast, but your pixel attributes only 0.4% to podcast links. What should you do?
What a post-purchase survey is genuinely good at
When you stop treating the survey as an ad attribution model, it becomes powerful for four specific tasks:
- Spotting invisible marketing channels: Podcasts, sponsorships, offline events, PR, and word of mouth produce very few direct link clicks. If 14% of your buyers name a podcast that your pixel rates at 0.4%, you have located an ad channel that works without generating clicks.
- Tracking trends over time: While absolute percentages are skewed by bias, the trend line over quarters is highly reliable because the bias stays constant. If "a friend recommended you" climbs from 8% to 19% over six months, word of mouth genuinely expanded.
- Sanity-checking suspicious performance: When a channel reports an unbelievable 14x ROAS on branded search, survey responses act as a reality check. If almost nobody names search unprompted, your ads are harvesting existing demand rather than generating new buyers.
- Segmenting customers by lifetime value: When you connect survey responses to backend customer orders, you can see which discovery channels produce your most valuable long-term buyers:
Five rules for designing a reliable survey
Five Design Rules
Process FlowAsk after the purchase, never during
A field in the checkout flow costs you conversions. Put it on the confirmation page or in the first post-purchase email, where abandoning has no cost.
Keep the options stable for at least two quarters
Every change to the option list breaks the time series. Since trend is the reliable output, changing options destroys the main value.
Make it optional and single-question
Required fields produce compliance answers — whichever option is first. A skippable question yields a smaller, more honest sample.
Include 'Other' and 'I don't remember'
Without an escape hatch, that uncertainty gets redistributed into your real channels and quietly corrupts them.
Store the response against the customer, not the order
Joined to the customer record, it becomes a segmentation dimension for life. Stored on the order, it is a one-off statistic.
Do not put the survey in your checkout. Every extra field in a checkout form lowers conversion rate. A post-purchase survey with a 40% completion rate and zero sales friction beats a mandatory checkout survey that costs you 2% of total sales every day.
How to route each instrument to the right decision
The golden rule of attribution is simple: never blend survey data and pixel data into a single hybrid score. Any blended formula produces an artificial number that helps nobody make a clear decision.
Instead, route each instrument to the specific decision it is qualified to inform:
Which Instrument Answers Which Decision
Reporting HierarchyBid and creative decisions
Pixel only. The platform's optimisation needs high-frequency mechanical signal, and survey data is neither fast enough nor granular enough to inform a bid.
Channel budget envelope
Blended CAC and MER, with the survey used to decide whether an untracked channel deserves a line in the budget at all.
Untracked channel investment
Survey only, because no pixel exists. A podcast or PR line is justified by survey mentions and by blended CAC moving in the right direction while it runs.
Causal truth
Neither. If the answer materially changes your budget, run a holdout — geo, audience, or a spend pause. Both instruments are correlational.
When a channel's survey mentions are climbing, its pixel-credited orders are rising, and your blended CAC is dropping, all three instruments agree. If you need proof before committing major budget, verify the lift with a pause-spend incrementality test.
Frequently asked questions
What response rate should I expect?
Expect roughly 30% to 50% for an optional single question on your order confirmation page, or 15% to 25% in a post-purchase email. A response rate above 80% usually indicates the field was made mandatory, meaning customers are selecting whichever option appears first just to finish checkout.
Should I offer an incentive to answer?
No. Incentives attract customers who answer randomly just to get the reward, diluting data quality. They also skew results toward price-sensitive shoppers.
Free text or dropdown?
A dropdown menu paired with an optional free-text field for "Other". Free text alone is difficult to categorize in daily reporting. Use the free-text entries to discover new marketing channels that deserve their own dropdown option next year.
Can I use survey data to set channel budgets directly?
Not on its own. Use survey responses to decide whether an untracked channel deserves budget, and monitor overall company blended CAC to determine how much to spend.
My survey says 30% "word of mouth". Should I stop paying for ads?
No. Word of mouth is almost always powered by paid advertising. Customers acquired through ads tell their friends and family. The survey reveals the final word-of-mouth step in a chain that started with an ad you paid for.
Takeaways for your next report
- Pixels record the last digital click; surveys record remembered brand awareness. They cannot and should not reconcile.
- Survey biases like recency and brand substitution are consistent, meaning trend lines over time are much more reliable than absolute numbers.
- Surveys excel at spotting offline channels like podcasts and segmenting customers by long-term repeat purchase value.
- Never blend pixel and survey numbers into one metric — use pixels for ad bidding and surveys for untracked budget allocations.
Blended CAC Calculator
Total spend across every channel, divided by total new customers — the acquisition cost number that reconciles with what you actually spent.
Chinmay Raibagkar
About author →Founder of DataLens AI. He helps non-technical teams read their ad and database numbers with confidence — which number to trust, what to do next, and what to ignore.