Classification Axis

Origin & evolution

The axis comes from positioning🎪PositioningGo-To-MarketProduct positioning statementView reference → theory and the perceptual map. Al Ries and Jack Trout's *Positioning: The Battle for Your Mind* (1981) made the case that a brand occupies a coordinate in the customer's mind, which presumes dimensions to be positioned along. Those dimensions are classification axes.

The rigorous version is older, drawn from psychometrics. Perceptual mapping borrows multidimensional scaling and factor analysis, developed in the 1950s and 1960s by researchers such as Warren Torgerson and Joseph Kruskal. The technique derives the axes statistically from how customers rate similarity between brands, rather than letting a strategist assert them. The map then shows distance as perceived interchangeability: brands close together are seen as substitutes.

The popular form collapsed all this into the 2x2: two axes, four quadrants, every competitor🥊CompetitorMarket IntelligenceA competing product or companyView reference → placed. Convenient, and easy to rig. The same companies rearrange depending on which two axes you pick, and the temptation is to choose the pair that puts you alone in the top-right. The honest practice treats axis selection as the real analytical work, choosing dimensions that buyers care about and that genuinely discriminate.

How it works in practice

A team mapping the project-management market starts with the lazy pair, price against features⭐FeatureProduct SpecificationA product capability or featureView reference →, and learns nothing: everyone clusters in a diagonal smear. They switch axes. The first becomes deployment model, from self-serve to sales-assisted. The second becomes target team size, from solo through enterprise.

Now the field separates. One cluster is self-serve and built for small teams; another is sales-led and enterprise-only; a sparse top-left, self-serve but enterprise-grade, holds two players and a gap. The axes did the work. The same competitors that looked identical on price-versus-features now sit in distinct regions, and the empty quadrant is a positioning opportunity🚪OpportunityDiscoveryA validated gap worth solvingView reference → the team can argue about with evidence🔎EvidenceValidationData supporting or refuting a hypothesisView reference →.

Classification axis vs. its neighbours

• Classification value📍Classification ValueMarket IntelligenceA position or value on a classification axis, representing where one option lands on that dimension.View reference → is one position on an axis; the axis is the whole dimension. "Enterprise" is a value; "company size" is the axis that contains enterprise, mid-market, and SMB. The axis defines the spectrum, the value marks a point on it.
• Market segment🎯Market SegmentMarket IntelligenceA distinct group of potential customersView reference → is a group of buyers defined by shared attributes; a classification axis is a measuring stick you can apply to competitors, segments, or products alike. Segments are often defined by their values along several axes.
• Competitive analysis📊Competitive AnalysisMarket IntelligenceA structured analysis of the competitive landscapeView reference → is the study that uses axes to structure comparison. The analysis is the activity; the axis is one of its instruments. A single analysis usually selects two or three axes from many candidates.

In the graph

In the Unified Product Graph, a classification axis is the dimension node of a taxonomy, and it hangs off whatever owns that taxonomy. A product_dimensioned_by_classification_axisProductdimensioned byClassification Axishierarchy edge makes it a product-wide axis; competitive_analysis_dimensioned_by_classification_axisCompetitive Analysisdimensioned byClassification Axishierarchy scopes it to a single analysis. Until 0.32.0 only the second existed, and an axis whose only possible parent was a competitive analysis read as a positioning device rather than the general grouping mechanism it is. The product parent is what lets one axis serve the whole graph. Below it, classification_axis_includes_classification_valueClassification AxisincludesClassification Valuehierarchy enumerates the discrete positions along it, and nodes take a position through the polymorphic node_classified_as_classification_value edge. Separating the axis from its values, as two node types joined by an edge, lets the graph reuse one axis everywhere and classify competitors, segments, and personas🎭PersonaUserAn archetype representing a user segmentView reference → against the same shared scale, so a "company size" axis means the same thing wherever it appears.