Perceptual mapping

The study's commercial application used mid-1990s Dutch midsize-car data: 20 brands and 212 usable questionnaires. It capped comparisons at 20 pairs per respondent, randomly sampling when necessary; unrestricted comparisons would require 190 pairs. Those are design counts, not measured completion times or average workload. [Bijmolt, Wedel and DeSarbo, Section 5](https://link.springer.com/article/10.1007/s40547-020-00112-7)

The exhibit makes the measurement cost visible without pretending to recreate the map. The comparison cap reduces required questioning but creates incomplete data that the analysis must handle. A cap is not evidence that every respondent completed that many tasks. A study with many unfamiliar offers also needs a defensible way to distinguish lack of knowledge from perceived difference.

The selected model used four dimensions and two segments. A flat picture omits part of that structure. The paper also cautions that shared consideration need not imply similarity when purchases serve different occasions or household members. [Bijmolt, Wedel and DeSarbo, Sections 5–6](https://link.springer.com/article/10.1007/s40547-020-00112-7)

For a positioning decision, ask which interpretation would change the action. If customers see the brand as reliable but expensive, test whether price deters purchase, whether reliability is important and whether another offer supplies it at a lower delivered cost. A spatial position alone does not distinguish those possibilities. Research that connects perceptions with feasible choices is more useful than an appealing label on an axis.

An apparently vacant region needs a separate economic test. Customers may not want the combination, the attributes may conflict in production or the segment may be costly to reach. Investigate willingness to pay, delivery capability and customer response under an actual offer. The map can direct that investigation; it cannot establish demand, contribution or an attainable cost structure by itself.

Pooling distinct segments can produce a position that describes neither group well. Check whether similar-looking offers mean the same thing to different customers. A single average can conceal one segment valuing a feature that another dislikes. Segment results must still be supported by the sample and estimation rather than selected to fit a preferred strategy.

Awareness can reflect distribution and exposure rather than product similarity. Customers cannot make meaningful comparisons of offers they do not know. Excluding unfamiliar offers may improve judgment quality while limiting the competitive picture. Report that boundary and investigate whether unfamiliarity itself is commercially important.

Maps are conditional on occasion and alternatives. In an unnumbered hypothetical household purchase, different members may favor different products for different uses. A shared consideration set need not mean one common ideal product. Specify the use before interpreting proximity as substitution or distance as differentiation.

A perceptual change can follow advertising without a product improvement, or a product improvement can remain unnoticed. Competitive entry can alter relative positions without the focal firm changing. Validate the action with subsequent customer and economic evidence, and preserve the original measurement conditions so that a new map does not silently change the comparison.