South African retail brands are collecting more information than ever. In 2023, the country’s big data market was valued at $5.13bn and is projected to reach $9.27bn by 2028. However, more data does not automatically mean better decisions. Without a clear commercial question, even the most comprehensive research risks becoming an expensive exercise that changes nothing on the ground.
According to Carmen Brown, general manager of Field IQ, research is often commissioned before the business has clearly defined what it needs to know and the results can be lengthy store audits, fatigued fieldworkers and mountains of data without a clear commercial direction.
A brand or retailer may have a nagging sense that something is wrong for a promotion to underperform, for a competitor’s products to appear in more fridges or if a product becomes unavailable. These are familiar commercial frustrations but the instinct is to commission a broad survey covering everything from product availability and pricing to share of shelf, point-of-sale compliance, staff knowledge, shopper demographics and promotional visibility.
“What you end up with is a clean and complete dataset that answers 40 questions nobody will act on, and gives a thin answer to the two that would really matter. The process can be lengthy and costly and instead of solving a commercial problem, the research creates several more, and a brand or retailer ends up with an abundance of information without a clear direction,” explains Brown.
The hidden cost of asking too much
Collecting additional information can appear relatively inexpensive, particularly when a surveyor is already visiting an outlet, but the costs of adding another ten questions to an existing questionnaire can quickly accumulate.
Longer audits consume more time in-store, test respondents’ patience, place additional demands on surveyors and generate more information to analyse. They can also compromise the quality of the data brands are paying to collect. Consider a 45-minute audit at a busy forecourt or bottle store.
By the sixth outlet, a fatigued surveyor may start skipping questions or guessing answers. Store managers become less willing to cooperate, and respondents, especially those in the informal trade sector, have little incentive to entertain a lengthy questionnaire.
When a research report attempts to answer too many questions, important findings can become buried among dozens of observations that appear equally important. Decisions that should be straightforward can then take weeks to make.
A lengthy report can create the impression that a project has been successful, even when it has not led to meaningful change. Brown says Field IQ has encountered briefs in which a large proportion of the questionnaire was never referenced in the final presentation. Surveyor hours, travel, data capture and quality assurance had all been paid for, yet much of the information generated was ultimately unused.
“Having too much data, and not enough direction are two big problems but the root cause of both is sequencing. The clients who get direction from data can tell me in one sentence what they will do differently depending on the answer, and that distinction is fundamental. A lot of the fieldwork at Field IQ for example gets scoped long before the first survey is conducted. Our research is not measured by the volume of information collected or the number of slides in a presentation, but by whether it enables a business to make a better decision,” adds Brown.
Better questions equal better decisions
Moving away from asking descriptive questions to ones that inform a decision is crucial. Asking what our share of the shelf is provides a snapshot of the situation but whether our share of shelf is under 30% in the retailers where sales are dropping is more commercially useful.
The second question establishes a connection between a known problem and a potential intervention and helps a brand determine whether changes to make.
“We recently worked on a liquor launch, where the brand wanted to know if their new packaging was physically available and visible in targeted stores in the second week after launch and if not, where was the breakdown occurring. It was a very narrowly defined question which meant surveyors could cover more outlets per day, and we were able to sum up the findings and use them to develop a corrective plan within the same week. Same team, same tools, but the question was the difference,” explains Brown.
As well as getting the question right, research must also reflect the realities of the trade. A forecourt, a tavern, a spaza shop and a hawker operating at a taxi rank have different operating conditions, time pressures, and levels of trust, and require different approaches to gathering information.
“Surveys must be designed around the people answering them, and applying the same methodology uniformly across formal retail and informal trade could obscure important differences. The results may look useful in a report but will fail to accurately represent the conditions a particular retailer faces,” adds Brown.
For brands and retailers looking to get more value from their research investment, Brown suggests working backwards from the decision they need to make. She identifies four stages that help turn information gathering into a commercially useful exercise.
1. Define the commercial question first
The question should be expressed in one sentence, with a decision attached to it. For example, a brand might need to establish whether it should invest in additional merchandiser visits or improve products ranging across Gauteng. If the business cannot explain what it would do differently depending on the answer, the question is not yet ready for research.
2. Collect only information that could result in change
Every question should justify its inclusion. Having the right amount of relevant questions can cut a survey length by half or more which makes fieldwork more manageable and improves the quality of the information collected.
3. Design research for the trade environment
The right research design takes account of how an outlet operates, when it can be visited and what is required to establish the trust needed for honest answers. A one-size-fits-all approach to formal and informal retail risks producing findings that fail to capture the realities of either.
4. Turn findings into an action plan
A research report should conclude with three to five specific actions, each assigned to a person and given a deadline. Those actions must then be implemented and measured through follow-up research. Without follow-up, brands cannot establish whether the intervention worked or whether further changes are needed.
More data may promise greater insight, but volume alone does not guarantee understanding. Brown’s test for whether a survey has done its job is whether the client would still know what to do if their research report was lost and only the action list survived. If the answer is yes, the research has done its job.
