Quick answer: Big brands win most brand image scores because more people know them, so more people tick them against every statement, a pattern often called the halo or brand size effect. To see real strengths, compare each brand's score on an attribute with what you would expect given its overall size, for example by indexing against its average or using normalised scores. This shows what a brand stands for, rather than simply how well known it is.
Why does the market leader top every list?
Look at a typical brand image chart and the market leader is usually first on almost every statement: tastes good, good value, trusted, modern, for families. It looks like the leader is winning everywhere, and smaller brands look weak across the board. Clients often ask us whether that can really be true.
Usually it is not. In a standard brand image question, respondents tick the brands they associate with each statement. People mostly tick brands they know and use. A brand known by most of the market collects more ticks on everything, simply because more people have an opinion of it. The chart is measuring size as much as image.
How does normalising scores fix this?
Normalising removes the size effect so you can compare profiles. The idea is simple: work out what score each brand would get on each attribute if it were associated with everything equally, then compare its actual score with that expectation. A positive gap is a relative strength. A negative gap is a relative weakness.
| Method | How it works | Best for |
|---|---|---|
| Index against brand average | Each attribute score divided by the brand's average score across attributes | Quick, easy to explain |
| Expected versus actual | Compare actual scores with values expected from brand size and attribute popularity | Fairer comparison across many brands |
| Correspondence analysis | A statistical map placing brands near the attributes they over-index on | Visual positioning maps |
| Users only | Base each brand's image on its own users | Removing awareness gaps entirely |
What does a normalised view reveal?
Once the size effect is removed, brand profiles become much clearer. A smaller brand may turn out to over-index strongly on health or authenticity, giving it a real positioning to build on. A leader may turn out to under-index on modernity or excitement, even while scoring highest in absolute terms.
This view is particularly useful in categories with one or two dominant brands, which is common across Southeast Asian FMCG. It helps challengers find the space they genuinely own, and it helps leaders spot erosion in an attribute before absolute scores fall.
What does a worked example look like?
Picture an instant noodle category with one dominant brand known by nearly everyone and a smaller challenger known by about half the market. On raw scores, the leader wins every attribute: taste, value, modern, convenient and trusted. The challenger's team is told its brand image is weak.
After indexing each brand against its own average, the picture flips on two attributes. The challenger over-indexes strongly on being modern and on bold, new flavours, while the leader under-indexes on modernity relative to its size. That gives the challenger a clear, defensible positioning to build on, and it gives the leader an early warning that modernity is a gap a competitor can exploit, especially with younger buyers. Neither insight was visible in the raw chart, even though the underlying data was identical. The challenger's next step was to check whether 'modern' actually drives purchase in the category, so it would not build a positioning that consumers notice but do not reward.
How should you set up the survey for a fair comparison?
- Ask image questions of brands people know. Show only brands the respondent is aware of, so you are not collecting random guesses.
- Use a consistent attribute list. Keep wording stable across waves so profiles can be tracked.
- Include enough brands. Normalising works best with several brands and attributes in the grid.
- Read by segment. Profiles can differ by age, region or ethnic group, so check the key splits.
- Pair image with outcomes. Link attributes to consideration or purchase to see which strengths actually matter.
If your tracker has drifted or the attribute list feels dated, our checklist of signs your Malaysia brand tracker uses the wrong metrics is a helpful starting point.
What are the common traps when reading brand image?
- Presenting raw association scores as a brand's positioning.
- Concluding that a small brand is weak on everything when it is simply less known.
- Comparing image scores between brands with very different awareness without adjusting.
- Changing the attribute list mid-tracker and losing the trend.
- Reading profiles from the total only, when segments tell a different story.
For a wider list of tracker questions worth asking, see our brand tracking questions for Southeast Asia.
Where does Vase.ai fit?
We build Vase.ai for quantitative consumer surveys, including brand health and image studies, on a verified panel of Southeast Asian consumers, with AI-validated responses and results in a real-time dashboard. Teams can run them DIY or with our research experts, who can add a full research report with normalised brand profiles.
If your brand image results never seem to change, or always favour the leader, tell us about the decision you are trying to make. We are happy to talk through how to read your data more fairly.
Frequently asked questions
What is the halo effect in brand image research?
It is the tendency for well-known or well-liked brands to be associated with almost every positive statement, which inflates their scores across the board and hides their specific strengths.
What is a normalised brand image score?
It is a score adjusted for brand size and attribute popularity, so you can see whether a brand is more or less associated with an attribute than you would expect, rather than simply how many people ticked it.
Should brand image be asked only among users?
Asking among users removes awareness differences and shows how each brand's own customers see it. Asking among all aware consumers shows broader market perception. Many trackers report both.
