“We're more reliable than they are.”
The buyer says: “How do you know?”
Everything after that question determines whether you gained credibility or lost it. If the answer is a restatement — “our architecture is more robust” — the buyer has learned that your competitive claims are assertions, and will discount the rest accordingly.
The definition
A claim is a statement of superiority. A proof point is evidence a buyer could verify without you.
That last clause is the whole test. Not evidence you could produce if pressed — evidence they could check themselves.
- “We're more reliable” — claim.
- “We publish a status page with two years of incident history; they don't publish one” — proof point. Checkable in twenty seconds.
- “Our support is better” — claim.
- “Our published SLA is one hour on all paid tiers; theirs is one business day below enterprise” — proof point. Both pages are public.
The proof-point versions are also more specific, which is a side effect worth noting: evidence forces precision.
The four kinds worth having
1. Published artefacts. Something the competitor put in public: pricing page, SLA, documentation, status page, changelog. The strongest kind, because they cannot dispute their own material.
2. Third-party evidence. Review data, analyst coverage, independent benchmarks. Strong because the source has no stake, though it ages and must be dated.
3. Named customer outcomes. A specific customer, a specific situation, a specific result. Powerful when the buyer recognises the situation as their own; useless when generic.
4. Demonstrable behaviour. Something you can show live. The strongest of all when available, because it survives every objection except “that's a demo environment.”
What does not count
Frequently offered, uniformly discounted by experienced buyers:
- Unattributed testimonials. “— a Fortune 500 CTO” is read as invented.
- Internal benchmarks. Tests you ran on your own product against theirs. Buyers assume, usually correctly, that the conditions favoured you.
- Round numbers. “10x faster” without a method reads as marketing. “3.4x on this specific workload, method described here” reads as a measurement.
- Awards from organisations the buyer has not heard of.
- Feature presence. That you have a feature and they do not is a fact, not proof it matters. The proof is that customers use it and it changes an outcome.
Building them from public evidence
You do not need a research budget. Most durable proof points come from the competitor's own public material:
- Pricing pages. Tier boundaries, what costs extra, what is enterprise-only.
- Documentation. Written for existing customers, so it describes limits honestly.
- Published SLAs and status pages. Commitments and track record.
- Review free-text. Recurring, dated, independent complaints on a specific dimension.
- Changelogs. What they have and have not shipped.
Where to look in more depth: the public record.
Handling decay
Every proof point has a half-life. A pricing comparison is accurate until they change pricing. A feature gap closes the day they ship.
Three habits contain the damage:
- Date every proof point where it is stored, so a rep can judge staleness.
- Cite the source, not just the fact, so it can be re-checked in seconds rather than re-researched.
- Prefer structural to incidental. A gap arising from their business model — a support ratio implied by their pricing, a limit imposed by their architecture — lasts far longer than a missing feature.
This decay is the mechanism behind battlecards going stale in 90 days.
The cost of getting one wrong
Worth being blunt about. A rep who states a proof point that turns out to be false does not merely lose that point. They lose the deal's credibility, and they stop trusting the material that gave it to them.
One bad claim retires a battlecard more reliably than a year of neglect. Which means the correct standard for including a proof point is not “probably true” but “checked, sourced, and dated.”
If that standard leaves you with four proof points instead of twelve, you have four you can use with confidence, which is worth more than twelve you cannot.
Our Sales Battlecard report builds proof points from public, citable evidence, with the source attached to each so it can be re-verified when a competitor moves. The surrounding structure is in the anatomy of a battlecard.