Your team says the brand is “premium but accessible.” So do two competitors. A perceptual map tests that claim against what customers actually think, by placing every brand on the same chart on the attributes buyers use to choose. It is one of the most practical tools in competitive intelligence, as long as it is built from survey data rather than a whiteboard session. This guide shows how to build one, read it and keep it current.
Key takeaways
- A perceptual map plots brands on the attributes customers use to choose, so you can see crowded positions and open gaps.
- Build it from survey data: ratings for a two-attribute map, association counts and correspondence analysis for a multi-attribute map.
- Pick axes from what drives customer choice, not from what the team wants to be known for.
- Map key segments separately, add their ideal points, and track the map across tracker waves.
What is a perceptual map used for?
A perceptual map answers three positioning questions. Where do customers place us? Which combination of attributes does no brand own? Is a relaunch, price change or campaign shifting perception between waves? It shows perception, not sales, so a brand can hold a distinctive position that few people want.
What is the difference between a two-attribute and a multi-attribute perceptual map?
A two-attribute map plots each brand’s mean score on two rated attributes, one per axis. It is easy to explain and redraw each wave, but you choose the axes, so everything else customers care about is left off.
A multi-attribute map summarizes many attributes at once. A technique such as correspondence analysis finds the two dimensions that separate the brands most. The axes come from the data, which makes them more honest and harder to label.
How do you make a perceptual map?
- Choose the brands. Include every brand customers actually consider. A competitor analysis framework helps you settle the set.
- Choose the attributes. Draw them from qualitative work and from driver analysis of what predicts choice.
- Ask the questions. A 1–10 rating per attribute per brand for a two-attribute map, or “Which of these describe each brand?” for a multi-attribute map. Ask only about brands each respondent knows.
- Build the brand-by-attribute table. Mean scores for ratings, mention counts for associations, weighted as in your other reporting.
- Plot two key attributes, or reduce with correspondence analysis.
- Label the quadrants. Split each axis at the scale midpoint or the brand average and name each quadrant in plain words.
- Read the gaps. Find empty quadrants, then check whether customers want what is missing.
How do you read a perceptual map from survey data?
The numbers below are illustrative. 600 respondents rated five brands on a 1–10 scale for price (1 = affordable, 10 = premium) and quality (1 = basic, 10 = high).
| Brand | Price (mean) | Quality (mean) | Quality minus price | Quadrant |
|---|---|---|---|---|
| Brand A | 8.2 | 8.0 | −0.2 | Premium, high quality |
| Brand B | 7.4 | 5.0 | −2.4 | Premium, basic |
| Brand C | 5.9 | 6.1 | +0.2 | Premium, high quality |
| Brand D | 2.8 | 3.3 | +0.5 | Affordable, basic |
| Brand E | 3.7 | 4.6 | +0.9 | Affordable, basic |
Put price on the horizontal axis and quality on the vertical, and split both at the scale midpoint of 5.5. Splitting at the brand averages (5.6 for price, 5.4 for quality) gives the same quadrants. Three readings follow:
- The open gap is affordable, high quality. No brand sits there. Brand E is closest, but its quality score is 0.9 below the midpoint.
- Brand B has a problem. It is seen as nearly as premium as Brand A (7.4 vs 8.2) with quality of only 5.0. Its quality-minus-price score of −2.4 is the worst of the five: customers think it charges for quality it does not deliver.
- Brand C is undifferentiated. At 5.9 and 6.1 it is neither the premium choice nor the value choice.
Before acting, check that the differences are real. If a rating’s standard deviation is around 2.0, the standard error of a mean from 600 respondents is 2.0 ÷ √600 ≈ 0.08, so gaps of a few tenths hold up. Then check whether anyone wants the empty position, which is what ideal points answer.
How do you build a multi-attribute map with correspondence analysis?
With ten or more attributes, start from an association grid and count the mentions into a brand-by-attribute table. Correspondence analysis turns that table into a map where brands with similar profiles sit close together, and each brand points toward the attributes it is linked with more than average. It needs counts, not mean ratings. Read it by direction from the center, report the share of inertia on each axis, and label the axes only after seeing which attributes sit at each end.
What should a perceptual map template include?
Use this perceptual map template before you draw anything: one row per brand, one copy per segment and wave. Note the question wording, scale anchors, split points and weighting above it.
| Brand | Attribute 1 | Attribute 2 | Base (n) | Quadrant | Change vs last wave |
|---|---|---|---|---|---|
| Your brand | |||||
| Competitor 1 | |||||
| Competitor 2 | |||||
| Ideal point (segment) | n/a |
How do ideal points and segments change the map?
An ideal point is where a group of customers would place their perfect brand. Ask respondents to rate “your ideal brand” on the same attributes and plot the mean for each segment.
Suppose price-focused buyers put their ideal at 3.5 on price and 6.5 on quality, and quality-focused buyers at 7.0 and 8.5. Brand A sits 1.3 points from the quality-focused ideal, while the nearest brand to the price-focused ideal, Brand E, is 1.9 points away. The empty quadrant has a segment waiting in it. Positions also differ by segment: a brand newer customers see as premium may look ordinary to long-standing ones, so map key segments separately.
How do you track position shifts across waves?
Ask the same questions about the same brands in every wave of a brand tracker and plot each brand’s path. In the example, Brand E moves toward the gap:
| Wave | Price | Quality |
|---|---|---|
| Wave 1 | 3.7 | 4.6 |
| Wave 2 | 3.9 | 5.1 |
| Wave 3 | 4.0 | 5.7 |
Quality rose 1.1 points while price rose 0.3, so Brand E has crossed into the affordable, high-quality quadrant. With 600 respondents a wave and a standard deviation of 2.0, a wave-on-wave change needs about 0.23 points to be significant at 95%, so each step is real. A market trend analysis tells you whether the whole category moved or just one brand. On a running tracker you can track brand health, consideration and competitors continuously and see perception shifts as they happen instead of redrawing the map once a year.
What are some perceptual map examples in marketing?
- Launch planning: a challenger maps price against convenience, finds an empty quadrant and tests a concept aimed at it.
- Repositioning checks: a brand that improved quality maps price against quality each wave to see whether customers noticed.
- Message choice: a multi-attribute map shows which attributes a competitor already owns, so the campaign claims one it can win.
What are the common mistakes?
- Axes chosen by the team, not customers. If the attributes do not drive choice, the gap may be empty for a reason.
- Opinion instead of data. A map drawn in a workshop shows what the team believes, not what customers think.
- Ignoring segment differences. A total-market map can hide two segments that see the brand in opposite ways.
- Treating every gap as an opportunity. Check ideal points and demand before moving into empty space.
- Changing brands or questions between waves. Positions then move for reasons unrelated to the market.
What should you do next?
Take your latest tracker wave and fill in the template for your brand and three closest competitors on the two attributes that best predict choice. If the map shows a gap or a drift, put it with its evidence in your next competitive intelligence report, and redraw the same map next wave to see whether it held.