Short answer
When new customers buy once and do not return, the usual suspects are price and product quality. Often it is something else that happened around the first order: delivery, set-up, a size that did not fit, a subscription they did not expect, or simply no reason to think of you again.
A satisfaction score will not tell you which. Ask ten recent first-time buyers to describe what happened after they ordered, and ask five repeat buyers the same thing. Compare the two groups.
Want to ask your own customers? Start with your question.
Why a good satisfaction score does not predict the second order
Post-purchase surveys are sent at the best possible moment for the business: the parcel has just arrived and the product is new. A customer can honestly answer 5 out of 5 and never order again.
Three things explain the gap:
- The survey measures the product; the decision to return depends on the whole experience. Delivery windows, returns and reordering effort rarely appear in the questions.
- The people who did not come back are the least likely to answer. The score describes your happiest customers.
- A rating has no room for "it was fine, but". The second half of that sentence is what you need.
Repeat purchase rate tells you that customers leave. Satisfaction tells you they were content. Neither tells you what stood between the first order and the second.
What to ask one-time customers
Ask about events, in order. People remember what happened far better than how they felt on a scale.
| Moment | What it can reveal | Example opening question |
|---|---|---|
| Before the order | What they were trying to solve, and what else they considered | "What made you look for this in the first place?" |
| Ordering | Friction, doubts, surprises at checkout | "Was there a point where you nearly did not complete the order?" |
| Delivery or set-up | Timing, missed slots, effort | "Tell me about the day it arrived." |
| First use | Whether it did the job they bought it for | "What did you use it for, and how did that go?" |
| Afterwards | Whether the need came back, and where they went | "The next time you needed this, what did you do?" |
| Contact from you | Emails, offers, reminders | "What have you heard from us since? What did you do with it?" |
The last two rows matter most. "The next time you needed this, what did you do?" often reveals a competitor, a habit or a small obstacle that no rating would surface.
Who to invite
- Eight to ten customers who ordered once, two to four months ago. Recent enough to remember, long enough for a second order to have been likely.
- Four or five who ordered again, as a comparison. What they describe differently is your clearest evidence.
- A mix of products or regions, if your range is wide.
Do not invite only people who complained. Most customers who leave never complain. They just stop.
A small thank-you, such as a voucher, is reasonable. Say clearly who will read the answers and why you are asking.
A fictional example: a meal-kit company
A fictional meal-kit company sees 62 of every 100 new customers stop after the first box. Satisfaction with the first order is 4.4 out of 5. The team is preparing a discount on the second box.
Before spending on the discount, they ask ten one-time customers and five regulars one question: "What happened between your first box and deciding about the second?"
- Six of the ten one-time customers say no delivery slot fits a working day. Two had a box left outside for hours.
- Four did not realise they had subscribed and cancelled as soon as the second charge appeared.
- Three mention price, each time alongside one of the points above.
- Four of the five regulars work from home or have someone in during the day.
The discount would have made the second box cheaper for people who still could not receive it. The findings suggest two things to test first: an evening delivery slot, and a plain confirmation of what the customer has signed up for.
Fifteen people do not prove that delivery causes the drop. They give the team a specific, cheap test to run before a costly one.
From findings to something you can test
Keep the output short:
| Finding | People | Test | Measure |
|---|---|---|---|
| No delivery time fits a working day | 6 of 10 | Evening slot in two cities | Second-order rate in those cities |
| Subscription was a surprise | 4 of 10 | Clear confirmation screen and email | Cancellations within 48 hours of second charge |
Here the conversation and the numbers work together. The conversations suggest the cause. Your order data confirms or rejects it at scale.
Mistakes to avoid
- Asking "why didn't you order again?" as the first question. It invites a polite, vague answer. Ask what happened.
- Offering a list of reasons. The reason you did not think of will not be on it.
- Waiting a year. People forget, and the email address may have changed.
- Treating one vivid story as the finding. Count how many people described the same thing.
- Arguing with the answer. If someone says delivery failed, the task is to understand it.
Where Lontra fits
With Lontra, you write the question in one sentence, such as "Why do new customers not come back after their first order?". You review the topics it proposes and send each customer a private link. It works on a phone, needs no account and takes a short written conversation. Lontra follows up on what each person says, then shows you how many people described each point, where one-time and repeat customers differ, and the quotes behind it. It is free to try with 10 conversations. A conversational AI survey like this suits a small group where each answer counts.
It does not replace your order data, and ten customers are not a sample of all of them. Lontra reaches the people you invite; it does not find customers for you. Participants are told before they start that you will read their answers.
Ask before you discount
A discount is the most expensive way to find out that price was not the problem. Ten conversations cost less and tell you more. Ask your own customers what happened after the first order.
Frequently asked questions
- Why do customers not come back after their first order?
- Often because of something around the product: delivery, set-up, an unexpected subscription, or no reminder when the need returned. The cause differs by business, so ask your own one-time customers what happened.
- How do you ask customers why they stopped buying?
- Ask about events in order: what they were trying to solve, the day the order arrived, the first use, and what they did the next time they had the need. Avoid offering a list of reasons.
- How many customers should I ask?
- Eight to ten who ordered once and four or five who ordered again is enough to see what repeats. Confirm the pattern afterwards in your order data.