How to Run a Customer Survey That Produces Useful Answers
The customers worth surveying are the ones who stopped buying, and they are the ones nobody asks. A short survey to lapsed customers produces more actionable information than a long one to satisfied ones.
Who to Ask
This guide is for supplement, Ayurvedic and herbal brand owners in India.
Surveys usually go to recent buyers or newsletter subscribers, who are the people most favourably disposed toward the brand. Their answers confirm what already works and rarely identify what to change.
The group with the most useful information is customers who bought once and did not return. They encountered whatever the problem is, decided against reordering, and almost never tell you why unless asked directly.
A short, polite survey to that group, sent a reasonable period after their expected reorder date, produces the most actionable information available to a supplement brand.
Questions That Produce Action
Short, specific and answerable.
- Did you finish the pack? A simple yes or no that separates abandonment from completion.
- If not, what made you stop? Offered as a short list including taste, texture, size, inconvenience, cost and no longer needed.
- What did you think of the taste or texture? The most common silent reason for abandonment.
- Was the product what you expected from the listing? Identifies expectation mismatch created by your own marketing.
- Did it arrive in good condition? Separates logistics problems from product ones.
- What would have made you reorder? An open question, kept optional.
What Not to Ask
Four categories to avoid entirely.
| Avoid | Why |
|---|---|
| Did it work for your condition? | Invites health information you should not collect and implies a claim |
| Any question about a health condition or symptoms | Sensitive personal data; avoid collecting it |
| Did you see results? | Implies an outcome claim and produces testimonials you cannot use |
| Would you recommend it for X condition? | Both of the above problems together |
| Long demographic batteries | Reduces completion and rarely changes a decision |
If a respondent volunteers health information anyway, do not use it in marketing and do not build it into a testimonial. An outcome testimonial is a claim regardless of who wrote it.
Running It Practically
Four points that improve response and usefulness.
Send It Through the Channel They Used
A message on the channel where they bought gets better response than an email to an address they rarely check.
Explain Why Briefly
Saying you are trying to improve the product, in one line, improves completion noticeably.
Do Not Incentivise Heavily
A large incentive attracts responses from people who want the incentive, which biases the answers.
Ask Once
One polite request, no follow-up chasing. Lapsed customers who do not respond have told you something too.
Follow applicable requirements for consent and messaging on whatever channel you use, and obtain advice if you are unsure what applies.
Acting on the Results
The part that is usually skipped.
Group the answers by cause, rank them by frequency, and pick the single largest actionable category. Then make one change and check whether the next cohort’s reorder rate moves.
Resist the temptation to act on everything at once, which makes it impossible to know what worked. And resist the temptation to discount the most common answer because it is inconvenient: taste is frequently the top answer and frequently the one brands least want to address.
Record what you found and what you changed, so the next survey can be compared against it rather than starting from nothing.
Checklist
Run it properly.
- Define the lapsed customer group by expected reorder date.
- Write three or four questions, no more.
- Include a question on whether the pack was finished.
- Include taste and texture explicitly.
- Include whether the product matched expectations from the listing.
- Include condition on arrival.
- Avoid all health condition and outcome questions.
- Send through the channel the customer used.
- Explain the purpose in one line.
- Keep incentives small or absent.
- Ask once, without chasing.
- Group answers by cause and rank by frequency.
- Act on one finding and measure the effect.
- Record findings and changes for comparison next time.
Found a Product Problem Worth Fixing?
Share what your customers told you and our team will suggest specification changes worth testing.
Frequently Asked Questions
Who should I survey?
Customers who bought once and did not return. They encountered whatever the problem is and almost never say why unless asked. Recent buyers and subscribers are favourably disposed and mostly confirm what already works.
How long should the survey be?
Three or four questions. A longer survey sent to a lapsed customer gets abandoned, and the few questions that matter are whether they finished the pack, what stopped them, what they thought of the taste, and whether the product matched the listing.
What should I not ask?
Anything about health conditions, symptoms or whether the product worked for a condition. Those collect sensitive personal information and imply outcome claims, and they produce testimonials you could not lawfully use anyway.
What if someone volunteers health information?
Do not use it in marketing and do not turn it into a testimonial. An outcome testimonial functions as a claim regardless of who wrote it, so it carries the same exposure as a claim you made yourself.
What is the most common finding?
Taste, frequently, and it is also the finding brands least want to act on. Group answers by cause, rank by frequency, act on the largest actionable category, and resist discounting the top answer because addressing it is inconvenient.