Members are never invited to a study without being told, before they decide, what it is about, what taking part involves, roughly how long it takes, and what happens to the answers. Nothing below changes that. The purpose of this page is to let an applicant see, in advance of applying, the full range of what could later be asked of them.
Study types
1. Product feedback surveys
A structured questionnaire about products a member already buys and uses in a specific category. Questions cover which products are in the home now, what drove the last purchase, what works and what does not, what would make a member switch, and how a described change to a product would land.
What taking part involves
One sitting at a computer or phone. A mix of multiple choice, rating scales, and two or three open text boxes where a member writes in their own words. The open text is usually the part the research team reads most carefully, so the instrument is kept short enough that a member still has the patience to write a real sentence at the end of it.
Question it answers
What people currently do in a category and why, measured across a defined composition of households rather than across whoever volunteered.
2. Purchase diaries
A member records purchases in an agreed category as they happen: what was bought, roughly when, where the decision was made, and whether it was planned or not. Entries are short by design, usually under two minutes each, because a diary that takes ten minutes an entry stops being filled in accurately by the second week.
What taking part involves
A commitment across the field period rather than a single sitting. Members receive a reminder on a fixed rhythm and log entries through a simple form. A member who misses a few days logs what they remember and marks it as reconstructed rather than guessing precisely; the instrument is built to accept that, because a diary pretending to perfect recall is worse than one that admits its gaps.
Question it answers
What actually gets bought over time, as opposed to what people remember buying when asked in a single survey. Recall over a month is unreliable in nearly every category, and a diary is the standard correction for it.
3. Packaging and label reviews
A member is shown packaging designs, label layouts, or on-pack claims on screen and asked what they understand from them. Typical tasks: say what the product is and what it does after a few seconds of exposure, find a specific piece of information on the label, say which of two layouts is clearer, and describe anything that reads as confusing or misleading.
What taking part involves
Looking at images and answering questions about them. Some tasks are timed to a few seconds to simulate a shelf glance; a member is told when a task is timed before it begins. Material shown in these studies is usually unreleased and is disclosed in confidence, so members agree not to photograph, post, or describe it outside the study.
Question it answers
Whether a design communicates what its makers believe it communicates. It very often does not, and this is the cheapest point at which to find that out.
4. Category habit studies
A broader read of how a household organizes an entire category: who decides, how often stock is replenished, what triggers a restock, which products are treated as interchangeable and which are not, and what would have to happen for the household to change its routine.
What taking part involves
A longer questionnaire with more open text than a product feedback survey, sometimes accompanied by a short set of household composition questions. Where a study is designed to measure change, the same members are invited again after an interval of several months and asked a substantially similar set of questions. The second invitation states plainly that it is a follow-up wave and that the member took part in the first.
Question it answers
How a category behaves in a real household over time, including the parts of the routine that a member would never think to mention if asked about a single product.
5. Follow-up interviews
A conversation with a researcher, arranged in advance at a time the member chooses, to go deeper on answers a member has already given in a study. Interviews are invited only from members who have completed the study they follow, and only where the written answers left a question that structured research cannot resolve.
What taking part involves
Scheduling a slot, then talking through a small number of topics. There are no trick questions and no test to pass; the researcher is trying to understand a reasoning process, not check an answer. Sessions may be recorded only where the member is asked in advance and agrees; a member can take part without being recorded, and can end the session at any point without giving a reason.
Question it answers
Why a pattern seen in the numbers exists. Quantitative studies are good at establishing that something is true and poor at explaining it, and an interview is how that gap gets closed.
How participants are selected and matched
Selection runs off the screening answers already held in a member's record, and off nothing else.
Each study starts with a sampling plan that names the profiles it needs and how many completed responses are required in each. A plan for a household-cleaning study might call for a specified split across household sizes and across purchase frequency bands, with a minimum number of households containing children. Those cells are defined before a single invitation is sent.
We then match the plan against member records and invite only the members who fit an unfilled cell. A member whose profile does not appear anywhere in the plan is not contacted about that study at all, which is the main reason members go through periods with no invitations. Because quotas are set in advance, they are also not adjusted afterwards: if a cell cannot be filled with genuinely matching members, the study is delivered with that cell reported as short rather than filled with the nearest approximation.
Two rules apply on top of the plan. Members are not invited to two studies in the same category close together, because a member who has just spent twenty minutes thinking about a category no longer answers questions about it like an ordinary shopper. And where a study measures change over time, the follow-up wave goes to the original participants specifically, which is the one case where a member's history, rather than their profile, decides the invitation.
What is asked of a participant, and what never is
Across every study type, participation is voluntary at the point of invitation and at every point after it. A member can decline, can start and stop, and can withdraw an answer set before the field period closes by writing to us.
- We ask for honest answers, completion within the stated field window, and confidentiality about unreleased material shown inside a study.
- We never ask for payment card or bank details, Social Security numbers, driver's license numbers, passwords, or account credentials of any kind. No study will ever require them, and any message claiming otherwise is not from us.
- We never require a member to buy anything, to keep anything, or to install software in order to take part.
Members are removed from a study, and may be removed from the panel, for answering at random, straight-lining rating grids, giving contradictory answers between screening and study, or submitting open text that appears to be machine-generated. This is a research-integrity rule rather than a disciplinary one: a single set of invented answers can distort a small quota cell badly.
How findings are used
Study responses are combined into a findings set for the client that commissioned the study. In the ordinary case that set is aggregated or de-identified: results are reported at the level of a group, and individual members are not named. Where a research design genuinely requires responses to be identified, the invitation says so before a member decides whether to take part, and a member who is not comfortable with it declines and takes no further part.
Findings are delivered together with the sampling plan, the sample composition actually achieved, the field dates, and the number of responses removed at quality review. Open text answers may be quoted in a report to illustrate a pattern; quotations are attributed to a profile, such as a household size and a purchase-frequency band, and not to a person.
What clients typically do with the result is decide something: whether to change a formulation, redraw a label, drop a line, or leave a product alone. Research that cannot change a decision is not worth a member's time, and we decline studies designed to confirm a conclusion the client has already reached.