When battlecard claims are tested with buyers who chose a competitor, some are confirmed and some contradicted — and the largest single category is claims that are accurate but played no part in the decision.
What is competitive intelligence for?
Informing positioning and equipping the people who have to hold a conversation with a buyer who is also talking to someone else. That is the operational test, and a lot of CI output does not meet it.
Most programs collect what is public, because that is what can be collected systematically. What is public is what a competitor has chosen to say — which is why the real choice is between talking to your own customers or to the ones you lost.
The bias toward public sources is not laziness; it is a response to what can be resourced. Public monitoring can be assigned, scheduled and automated, while buyer interviews require commissioning, recruitment and someone to read the transcripts. A program judged on coverage will always drift toward the first, which is why the operational test has to be written into how the function is measured.
What do public signals actually show?
Positioning, pricing where it is published, hiring direction, and product announcements. That is genuinely useful for understanding what a competitor is trying to do.
It says nothing about whether it is working, and nothing at all about why a specific buyer chose them over you.
Public signals are also strategically shaped. A competitor's pricing page describes the price they want compared, their job postings describe the direction they want believed, and their announcements describe the roadmap they are willing to commit to publicly. Reading them as evidence of behavior rather than of intent is where most public-source analysis goes wrong.
What do lost buyers say instead?
Something more specific and frequently more mundane. Implementation confidence, a reference in their sector, a perception that the other vendor had done this before — reasons that surface on a call with the buyer and rarely appear on a battlecard, because they are not product claims.
The proportion of battlecard content that is accurate and simply irrelevant to the decision is the finding worth acting on.
That category is the useful output because it is actionable in a way a contradiction is not. A contradicted claim gets corrected and the card carries on; a large body of accurate-but-irrelevant content means the card is answering the wrong question entirely, and correcting individual claims will not fix it. The finding is about structure rather than accuracy.
How should a battlecard be built?
From loss interviews first and public signals second. The card should lead with the two or three things that actually decided recent deals, whatever they are, rather than with a feature comparison.
Cards built the other way round are complete, defensible and ignored by the sales team, which is the pattern most enablement leaders will recognize.
The other structural change is length. A card that leads with two or three decision factors can be held in a rep's head before a meeting, which is the only place it is any use. A comprehensive card cannot, and a rep who cannot recall it will not open it mid-conversation, so the completeness that made it defensible is what makes it unused.
How often does a battlecard decay?
Faster than a calendar suggests, and unevenly. Product claims decay on the competitor's release cycle; decision criteria decay more slowly but shift when a market's buying pattern changes, which a periodic buyer survey will catch.
Tying refresh to losses rather than to quarters keeps the card anchored to what is actually happening in deals.
Decay is also uneven across the sales team. A card out of date on a competitor's newest product costs a rep in that segment immediately and costs everyone else nothing, which means the signal that a refresh is needed reaches the enablement team late and through complaint rather than through review. Tying the review to losses catches it earlier, because losses are already reported.