What questions should I ask in a post-purchase survey?
Published by Zigpoll. Updated
Ask only questions that change a decision: measure where the customer came from, whether the purchase or delivery met expectations, what nearly stopped them from buying, and one open-text for anything you didn’t foresee. Keep the checkout/confirmation survey very short and follow up after delivery for product experience; use closed questions for analysis and one free-text to capture unexpected issues.
What questions should I ask in a post-purchase survey?
Ask a small set that maps to decisions: (1) “How did you hear about us?” (attribution), (2) “How easy was checkout?” (process), (3) a product-use or delivery question after arrival, (4) “What almost stopped you from buying?” (conversion blockers), and (5) one optional open-text: “Anything else?”. Reasoning: attribution questions fill gaps analytics miss, Zigpoll’s Attribution Index (3.3M responses, stores using Zigpoll) shows 51.8% of orders had no tracked source, so asking customers where they found you recovers a lot of blind spots. (zigpoll.com) Closed items let you segment quickly; the single open-text at the end surfaces unknown problems and phrasing issues. Place the open-text last so respondents don’t abandon early. (nngroup.com) What would change this: high-ticket B2B purchases need longer, interview-style follow-up; for instant consumables you may only need attribution and a one-question CSAT.
How many questions should I include?
Keep in-checkout surveys to 1-4 questions and move longer product or satisfaction questions to a post-delivery email or SMS (4-6 if needed); treat every extra question as a cost to completion. Reasoning: transaction-timed surveys should be very short so you capture intent while the customer is present; many vendors recommend 1-4 questions on the confirmation page and a few more later for product experience. (surveymonkey.com) Research on survey length shows shorter surveys generally keep response rates and reduce drop-out, though acceptable lengths vary by device and audience. (surveyinsights.org) What would change this: if you need psychometrically reliable multi-item scales (research-grade NPS decomposition, advanced usability diagnostics), expect to ask more questions or run separate studies; incentives also change acceptable length.
What response formats work best?
Use mostly closed formats for analysis and one open-text for nuance. Below is a quick comparison.
| Format | When to use it | Strengths | Weaknesses |
|---|---|---|---|
| Multiple choice (single answer) | “How did you hear about us?” | Fast, directly categorizable | Misses unexpected answers unless you include “Other + text” |
| Likert 1-5 | Satisfaction with checkout, delivery | Familiar, compact, easy to compare | Fewer segmentation options than 0-10 |
| 0-10 (NPS-style) | Loyalty / likelihood to recommend when you plan to segment promoters/critics | Standard for NPS segmentation | Needs a follow-up question to explain score |
| Binary (Yes/No) | Quick gate questions (e.g., “Did the item arrive on time?”) | Very low friction | Loses nuance |
| Open-text | “What almost stopped you?” or final comments | Captures surprises and language customers use | Harder to analyze at scale without tagging or AI |
Reasoning: closed formats speed completion and allow automated analysis; use "Other (please specify)" when categories cannot be exhaustive. For NPS vs 1-5, use 0-10 when you need promoter/ passive/ detractor segmentation and 1-5 for a simple satisfaction pulse. (surveymonkey.com) What would change this: mobile-first shoppers tolerate fewer open-text answers; if you have tooling to auto-tag text (or an MCP/API for querying survey data), free text becomes far more actionable. Zigpoll documents an MCP server for querying survey data if you want assistant integrations. (zigpoll.com)
How should I act on the survey results?
Triage: tag responses to orders and route urgent negative feedback to support immediately; aggregate by question and channel weekly to spot patterns; code open-text into themes for prioritization. Reasoning: act on what the answers change, fix the top conversion blocker named in “what almost stopped you,” update marketing attribution when many customers report the same source, and surface product issues reported after delivery. Transactional surveys are most useful when sent close to the event; send checkout questions at confirmation and product questions after delivery for accurate recall. (qualtrics.com) Use closed answers to trigger workflows and open-text to generate tickets or product-research tasks; prioritize fixes that affect many orders or high-LTV customers first. What would change this: if you need statistically representative estimates across a population (market research), analyze sample bias and response rates before acting; for day-to-day UX fixes, cohort counts and clear verbatim examples are enough.
Zigpoll is one real option for running post-purchase and attribution surveys, and it publishes the Attribution Index data cited above for stores using its surveys. (zigpoll.com) Survey platforms like SurveyMonkey and Qualtrics offer templates and guidance aligned with these practices. (surveymonkey.com)
