The competitive field, in the same graph as the work it informs

A portfolio usually holds what a company owns. It also has to account for what it does not. UPG lets a competitor🥊competitorA competing product or company sit inside the same portfolio as the products it competes with, tagged with a member_kind of watched.

The watched tag changes how the graph is scored. A watched graph is held to a different standard than a product the company owns: an empty discovery region in Adyen’s graph is not a gap in the company’s own discovery, and a half-mapped rival does not drag the portfolio health score down. portfolio_validate and the coverage scorers read member_kind and adjust what they expect.

Once a competitor🥊competitorA competing product or company is in the portfolio, it is modelled like any product the company owns. It offers a competitor feature🔬competitor_featureA feature offered by a competitor, competes for the same persona🎭personaAn archetype representing a user segment, addresses the same need🫀needA user need, pain, desire, or constraint, is classified as a classification value📍classification_valueA position or value on a classification axis, representing where one option lands on that dimension., and emits the competitor signal📡competitor_signalA competitor signal entity nodes the team tracks.

The comparison stays current with little manual upkeep. The same query that walks an owned product walks the rival, so a question like “how do the two compare on fraud tooling?” is one traversal across two graphs.

The unit of competitive change is the competitor signal📡competitor_signalA competitor signal entity: one dated move a rival makes. A feature launch, a pricing change, an acquisition, a partnership, a market entry. It is emitted by the competitor inside the watched graph, with the date and the evidence attached.

What makes it more than a feed is that each signal connects back to the team’s own work. A move that lands on a shipped feature becomes a signal_maps_to_feature edge to a feature⭐featureA product capability or feature; a move that exposes a gap worth chasing becomes signal_surfaces_opportunity to an opportunity🚪opportunityA validated gap worth solving. Each new signal updates the picture in place, ready to query.

The competitive question teams ask most is also the hardest to keep current: where does a product lead a rival, and where does it trail? UPG models it as the parity edge feature_rivals_competitor_feature, connecting a feature⭐featureA product capability or feature node to a competitor feature🔬competitor_featureA feature offered by a competitor and carrying the assessment on the edge itself.

Parity status, relative quality, whether it counts as a gap, when it was last assessed, the supporting evidence, a confidence score: all of it rides the relationship rather than a paragraph. The question becomes a filter over the competitive lens.

Parity covers where a product leads or trails on a given feature. Classification answers a different question: what kind of competitor is this? A classification axis↔️classification_axisA single dimension in a competitive classification matrix that defines one axis of comparison across alternatives. is defined once in the shared registry (AI maturity, say, with the values agentic, integrated, bolt-on) and every rival is placed against the same classification value📍classification_valueA position or value on a classification axis, representing where one option lands on that dimension. set, so the map holds across the whole organisation instead of being reinvented by each analyst.

Each placement carries its own assessment: a typed confidence on a named scale, the date, the evidence, who made the call. A stale machine-polled guess is easy to tell apart from a fresh hand-verified one, and re-classifying merges new evidence onto the existing edge rather than spawning a duplicate. portfolio_digest reports the distribution across each axis, so the shape of the field is available as a single read.

A market trend📉market_trendAn emerging trend in the market creates an opportunity🚪opportunityA validated gap worth solving, and where that opportunity meets a rival weakness, there is a move. A competitive battle card⚔️competitive_battle_cardA competitive battle card for sales enablement turns it into a position a rep can win with: it references the competitor🥊competitorA competing product or company and addresses the objection🚧objectionA common sales objection a buyer actually raises.

Because the card sits in the same graph as the parity edges and signals underneath it, it stays tied to the evidence. When the rival closes the gap, the card built on it is one traversal away, and can be updated in place.