Perceptual positioning map
Display relative positions on two defined dimensions, retaining data nature, sample size, calculation rules and unplaced objects.
Use a positioning map to investigate perceived differences between offers. Define objects, a customer task and two dimensions before placing points. This example describes three synthetic respondents; it is not representative research or a positioning recommendation.
Objects, respondent scope, dimensions, raw ratings or attributes, scales, calculations, missingness and data status.
A labeled map, traceable coordinates, unplaced objects and questions for perception or demand research.
On this page
How to use it
- Define the task, objects, group, period and perception or attribute status.
- Specify endpoints, direction, units and scales, and investigate their relevance.
- Retain raw records, sample sizes, missingness and calculation rules.
- Plot supplied coordinates with a data table; retain unknowns and do not shift overlapping points.
- Turn clusters or gaps into research questions before choosing a position.
Keep its limits in mind
Perceptions are not objective attributes; listed price is not perceived value. Keep populations, periods and scales compatible. Do not place unknowns at zero or the center. Empty space is not demand, and a mean does not describe everyone. This figure plots supplied 1–5 coordinates under an explicit teaching interval assumption; it does not run correspondence analysis, statistical inference or ideal-position selection.
Mapping three synthetic respondents' ratings
Fictional teaching example: Alder Lunch is reviewing its direction. S3 contains three provider cards (E1, Oct 5–9, 2026), Alder's 100-order ledger (E2), ratings by three fictional respondents (E3), capability records (E4), and supply/entry questions (E5). These are synthetic records, not market research. Proposed actions have not been carried out.
On a small screen, scroll across the diagram to read it.
Keep data types separate: E1's CNY26/22/30 list prices do not supply perceived-value coordinates. E3 supplies three complete synthetic rating pairs per brand. X is pickup convenience and Y is perceived value for price, both from 1 to 5.
Reproduce the positions: Under the stated equal-interval teaching assumption, means are A(4,3), B(2,4), C(3,2), each n=3. D has no ratings and remains unplaced. These are direct rating means, not correspondence-analysis coordinates.
Result: The map describes only the synthetic records. A top-right gap establishes neither demand nor an ideal position. The next step is to investigate relevant dimensions and customer tasks, not announce a market opportunity.
Three fictional respondents' 1–5 ratings are averaged under a teaching interval assumption; D stays unplaced. Perceived value is not listed price, and a gap is not demand.
Next step: The research lead should check whether these dimensions matter to actual customers before designing a study; sampling, budget and action remain undecided.
Three situations where a dot should not be plotted
First, pause if the dimensions describe different things across offers. Pickup convenience needs the same task, endpoints, direction, and respondent scope. Listed price and perceived value for price are different variables. A cheaper price does not supply a higher perceived-value coordinate, and a brand claim is not a customer rating.
Second, leave an object unplaced when comparable evidence is missing. Zero is outside this example’s one-to-five scale, while the center would imply an actual rating. Neither is an honest representation of unknown data. Retain the missing object and the evidence needed to place it.
Third, stop before interpreting a gap as a recommendation. A blank area can result from the selected dimensions, offers, or respondents. It does not show that customers want an offer there or that the business can deliver profitably. A mean also hides disagreement; inspect individual records before telling a market story.
Choose between perceptions, offer descriptions, target hypotheses, and options
This map displays supplied rating means under an explicit teaching assumption. It does not estimate correspondence-analysis coordinates. Select the method that matches the data and question rather than copying the appearance of another perceptual chart.
| Question and available records | Method | Output and boundary |
|---|---|---|
| Comparable ratings on two defined dimensions | Perceptual positioning map | Traceable positions; no inferred ideal point or demand. |
| Provider cards describe target, value and marketing mix | 4P + offering and target analysis | Offer comparison; claims remain separate from perceptions. |
| Customer tasks and groups need investigation | STP | Target and positioning hypotheses; not representative findings by default. |
| Concrete feasible alternatives need a reasoned choice | Decision matrix | Explicit criteria and sensitivity; no customer ratings invented. |
Reproduce three positions without manufacturing a market gap
The complete fictional E3 packet has the same three respondents U1–U3 rating Alder, QuickBox, and Lane for October 5–9, 2026. X is pickup convenience, one least to five most, increasing rightward. Y is perceived value for price, one lowest to five highest, increasing upward. The raw records below are paired in U1, U2, U3 order. Treating these categories as equally spaced is a teaching assumption, not a property demonstrated by the data.
Alder’s X mean is (3+4+5)/3=4 and Y mean is (2+3+4)/3=3. The same arithmetic gives QuickBox (2,4) and Lane (3,2). Each position summarizes three synthetic people, not three independent samples per axis or a representative population. Offer D has no comparable ratings and stays outside the map.
The CNY26, CNY22, and CNY30 provider-card prices are not used to calculate Y. The top-right area contains no mean among these three plotted offers, but that observation does not select a new position. The next question is whether the dimensions matter in actual customer tasks and whether a suitable study can be designed.
| Offer and sample | U1 / U2 / U3: X,Y | Mean coordinate and limit |
|---|---|---|
| A Alder, n=3 | (3,2) / (4,3) / (5,4) | (4,3); variation retained, synthetic only. |
| B QuickBox, n=3 | (1,3) / (2,4) / (3,5) | (2,4); no population superiority claim. |
| C Lane, n=3 | (2,1) / (3,2) / (4,3) | (3,2); same dimensions and period. |
| D, no comparable ratings | Unknown; no fabricated respondent values | Unplaced; not zero, center, or a guessed coordinate. |
Review relevance before treating positions as strategy
The research lead should first check how customers understand each dimension and whether both relate to the intended task. A participant may value convenience differently for a rushed pickup and an unhurried meal. Define the use situation before combining answers. A revised dimension requires new compatible evidence, not a relabeled old axis.
If real research is later authorized, document recruitment, eligible respondents, timing, rating wording, missing answers, and the analysis plan. Keep respondent counts separate from the number of ratings. Inspect variation and explain any numerical treatment of ordered categories. Do not borrow a software product’s sample requirement or statistical interpretation for this different calculation.
At review, reproduce the coordinates from the retained records and list which objects remain unplaced. Preserve overlapping coordinates instead of moving dots to make the picture attractive. The example supplies no real survey, statistical inference, confidence interval, or test of an ideal offering.
The outcome may be a revised research question rather than a repositioning decision. Demand, operational capability, cost, and willingness to pay need their own evidence. Keep any proposed market action separate from its authorization and from later observed results; a persuasive map is still only as useful as its defined inputs.
Check your work
- Axes, direction, scales, sample, period and data nature are explicit.
- Coordinates are traceable; unknowns and prices have not become perceptions.
- Gaps and means are not demand or representative findings, and no ideal position is selected.
Framework references
References explain the framework and its scope; they do not validate the fictional example's results.
Put the framework to work
Copy the outline and add your question, evidence, and open questions.