Rules
Part of A closer look at competitive intelligence
Competitive intelligence practice: the habits that separate good from bad
Good competitive intelligence practice runs on specific habits like win/loss interviews, pricing watch, claim audits and dated source records.
What to take away
- Six competitive intelligence habits separate good practice from bad. They are win/loss interviews, pricing watch, claim audits, and battlecards. Source dating and a kill rule complete the set.
- Each habit needs one owner, a cadence, and a written acceptance test. A habit nobody can point to is not running.
- Date every competitor fact and record where it came from. An undated claim is a guess wearing a suit.
- Run one habit through a controlled failure before you expand it, and retire any habit that never changes a decision.
The six habits worth running
Most teams claim they run competitive intelligence. Few can name the habit, the owner, or the evidence behind a call.
Six habits worth running
- Win/loss interviewsbuyer's words, outside caller
- Pricing watchdated competitor price points
- Claim auditsevery claim traced to source
- Battlecardsexpiry date visible on card
- Source datingeach fact carries its date
- Kill ruleretire habits with no decision change
Win/loss interviews. Someone outside the deal calls the buyer after the decision and asks what tipped it. Clozd and Anova Consulting Group sell that service; buyers often get a $50 to $100 gift card. The habit fails when the rep runs the call, because buyers soften answers to the person who lost.
Pricing watch. A named person checks competitor pricing pages, quote sheets, and public tenders on a set cadence and records the date. Visualping emails when a watched page changes; its free tier allows five checks a day, and paid plans start near $10 a month.
SpyFu lists Basic at $39 a month for tracking rival ad copy and paid keywords. Semrush lists Pro at $139.95 a month on monthly billing. The habit fails when the number lives in someone's memory from last spring.
Claim audits. Every public claim gets checked against a source before it ships. The FTC requires a reasonable basis for objective claims before they run.
G2 review pages carry dated buyer comments you can cite. The habit fails when marketing copies a competitor's feature list and nobody notices the feature does not exist yet.
Battlecards with an expiry date. A one-page sheet per competitor, with their strengths listed honestly and two questions that expose the weak spot. Klue and Crayon sell battlecard modules that push cards into Salesforce, and both quote pricing on request. It fails when the card is two years old and sales stopped opening it.
Source dating. Each fact carries the date it was true and the source behind it. The Internet Archive's Wayback Machine returns a dated URL for a page as it looked on a given day. archive.today saves the same snapshot on demand. It fails when a 2024 competitor roadmap is quoted as current.
A kill rule. Written in advance: if a habit produces no decision change in two quarters, it stops. Without this, the program grows reports nobody reads.
Give each habit an owner and a test
A habit is real when a reviewer can watch it happen. Write the owner, the cadence, the input, the output, and the acceptance test in one line each.
Habit owners and acceptance tests
Habit
- Win/loss interviews
- Within two weeks of lost deal
- Pricing watch
- Monthly
- Claim audits
- Before every launch asset ships
- Battlecards
- Quarterly refresh
- Source dating
- Continuous
- Kill rule
- Twice a year
Cadence
- Win/loss interviews
- Buyer's words, outside interviewer
- Pricing watch
- Dated price points with source links
- Claim audits
- Every claim traced to a source
- Battlecards
- Expiry date visible on the card
- Source dating
- No undated competitor fact in file
- Kill rule
- Habits with no decision change retired
Acceptance test
- Win/loss interviews
- Pricing watch
- Claim audits
- Battlecards
- Source dating
- Kill rule
| Habit | Cadence | Acceptance test |
|---|---|---|
| Win/loss interviews (Clozd, Anova Consulting Group) | Within two weeks of each lost deal | Buyer's own words on record, interviewer not on the deal |
| Pricing watch (Visualping, SpyFu, Semrush) | Monthly | Dated price points with source links |
| Claim audits (G2 reviews, FTC substantiation rule) | Before every launch asset ships | Every claim traced to a source |
| Battlecards (Klue, Crayon) | Quarterly refresh | Expiry date visible on the card |
| Source dating (Wayback Machine, archive.today) | Continuous | No undated competitor fact in the file |
| Kill rule | Twice a year | Habits with no decision change retired |
For any measurement question, the NIST experimental design selection guidance starts with the objective and the practical constraints.
That split matters here: a win/loss finding describes what happened, while a controlled test estimates what a change would cause. Keep the two apart before you act on either.
Where the evidence file lives, and what it holds
Keep one file per competitor in a shared repository such as Notion, Confluence, or SharePoint. It holds the approved scope, the version, the owner, the decision date, and the facts with their dates. The file answers who said it, when, and what would change the conclusion.
Accessibility applies to the file itself. The W3C explanation of information and relationships covers making visual structure available programmatically. If your battlecard or dashboard is a table or a form, that structure has to reach a screen reader, not just the eye.
Retention has a second purpose. CISA's logging guidance for business systems explains why organizations keep and review event records. Keep access, change, and correction records for your intelligence file the same way. A log shows who touched the file and when; it never proves a business result.
Run one habit through a controlled failure
Before you expand, test one habit on a normal week and one bad week.
- Pick a defined competitor set and a comparison basis.
- Track decisions the habit changed, with cost and quality noted.
- Watch one ordinary path and one failure path.
- Preserve inputs, versions, and the decision record.
- Write down what you still do not know.
- Set a stop rule and the next review date.
The failure path is where habits earn their keep. A stale battlecard, a copied claim, or a pricing number with no date should trigger a pause, a correction, and a named owner for the fix.
When AI enters the workflow, the NIST AI RMF Playbook offers voluntary actions grouped under govern, map, measure, and manage. Use it to structure the change. It is not a certification and it does not rank tools.
Common questions
What is the first habit to build?
Start with win/loss interviews, because they produce evidence no public source can. Have someone outside the deal call the buyer within two weeks of the decision. Record the buyer's words, not a summary written by the losing rep. Clozd is one firm that specializes in these calls.
How often should battlecards be refreshed?
Quarterly, with an expiry date printed on the card so a stale one is obvious. Refresh sooner when a competitor ships a major release or changes pricing. A card nobody opens is a signal to change the format or drop it under the kill rule.
What should a team avoid?
Avoid copying competitor claims into your own materials, and avoid quoting a competitor roadmap as fact. Date every figure and name the source.
If a claim cannot be traced, it does not go in the file or the deck. The FTC has brought substantiation cases against advertisers that could not back an objective claim.
