Quick answer: Define lapsed users by recency, verify the exact product they used, then ask a ranked set of reasons for using less, plus where they went instead. Test price alongside marketing, product change, availability and competitors, rather than assuming price is the cause. Keep recent and long-ago lapsers apart in the analysis, because their memories are not equally reliable.
Who counts as a lapsed user?
A lapsed user once bought your brand and no longer does, or buys it much less often. That sounds simple, but the definition drives everything that follows. Someone who stopped buying last month is in a very different position from someone who stopped years ago.
Set a recency window and write it down. Many teams use bands such as used in the past few months, used once but not recently, and used long ago. Report each band separately, because the reasons and the quality of recall change as time passes.
What should the screener check?
Screening is where lapsed-user studies go wrong. Respondents often confuse brands, or confuse a brand with its sub-products, and the sample ends up full of people who never actually used what you meant. Screen on the specific product, not only the brand, and show pack images where possible.
- Ask which products in the range they have ever consumed, using images alongside names.
- Ask when they last consumed each product.
- Ask how often they consumed it at their peak and how often now.
- Separate those who stopped completely from those who reduced use.
- Include a trap or consistency check to remove inattentive respondents.
Allow a broad definition of past use if the aim is to learn why people left. Excluding long-ago users removes some of your most useful evidence, as long as you analyse them separately.
Which questions reveal the real reasons?
Open with behaviour, then move to reasons, then to alternatives. Ask respondents to rank their reasons for using less, rather than tick everything that applies, so you can see which reason matters most. Add a clear exclusive option for respondents who would not use more under any circumstances, so they do not distort the other answers.
| Question block | Purpose | Practical tip |
|---|---|---|
| Usage history | Establish when and how use changed | Use recency bands, not open dates |
| Ranked reasons | Show which reason outweighs the others | Randomise the order of reasons |
| Brand imagery | Measure how the brand is perceived now | Keep the list short and balanced |
| Switching | Show where users went instead | Use a list of named alternatives plus other |
| Win-back | Test what might bring them back | Offer an exclusive no-change option |
How do you avoid blaming price too early?
When sales fall, price is the obvious suspect, and it is often part of the story. But it can also be a convenient label for something else: a change in the product, less visible marketing, weaker availability, a new competitor, or a life-stage change such as a child leaving home.
Test all of these on the same list so respondents can weigh them against each other. If price is truly the leading reason, the ranking will show it. If it is not, you will have saved the business from a discount that does not solve the problem.
A useful closing block asks what, if anything, would bring them back. Keep it practical: a different pack size, a clearer benefit, a change in availability or a new variant. Avoid leading wording such as asking whether a discount would win them back, because most people will say yes to a discount and the answer tells you little.
How do you handle recall bias?
People forget, and they rationalise. A user who stopped long ago may give a tidy reason that is not the original one. Reduce this by using recency bands, showing images of the product and old and current packs, and asking what changed in their life or routine rather than only what changed in the brand.
Treat very distant lapsers as directional evidence. They help you understand the brand image people carry, even if their exact reasons are fuzzy. Recent lapsers give you the sharper diagnosis.
How many lapsed users do you need?
A common norm is around one hundred respondents per group you want to compare, and more if you plan to cut the data many ways. Lapsed users can be a low-incidence audience, so check feasibility before fixing the sample size. Our guide to the survey sample size covers the rules of thumb, and the usage and attitude study guide shows how to build the surrounding questionnaire.
In Malaysia, remember to decide which languages the survey will run in, and plan quotas for the ethnic and regional groups the brand actually sells to, so the findings reflect your real customer base.
Where does Vase.ai fit?
We build Vase.ai for brand and insight teams who need fast, structured answers from verified Southeast Asian consumers, and lapsed-user and switching studies are a strong fit. Built-in screening, quotas, logic and a real-time dashboard let you see who lapsed, why, and where they went.
Other methods cover different jobs: in-depth interviews are the better tool when you need rich personal stories, and a local field agency suits audiences that cannot be reached online. For the question of how many people left and why, a well-built survey is usually the quickest route.
Frequently asked questions
What is the difference between lapsed users and low-frequency users?
Lapsed users have stopped buying. Low-frequency users still buy but rarely. Their reasons often differ, so screen for both and report them separately.
Should I include current users in a lapsed-user study?
Often yes, as a comparison group. Seeing how current users answer the same questions shows which perceptions truly separate those who stayed from those who left.
Can I find out why sales fell with a survey alone?
A survey shows what lapsed users say and how common each reason is. For a full diagnosis, combine it with sales data, distribution data and, where useful, a few interviews to explore the stories behind the numbers.
