# Competitive battlecards — what each competitor claims, with sources Recipe No. 37, Work and teams. From The know.sh Cookbook: https://know.sh/cookbook/battlecards - For: a product marketer who keeps the battlecards the sales team reads before a call - You bring: the competitors that come up in deals, their public websites, documentation and changelogs, your own release notes, and what sales hears on calls - You get: a battlecard per competitor: each public claim with its source and the date you read it, your answer, and the landmines to avoid, read by sales from a password link - Time: a morning per competitor, then an hour a month per card - Keep it: private while you work; share one document by a read-only link Battlecards go stale faster than anything else in marketing. A rep repeats that a competitor "has no returns portal" on a call, the prospect says it shipped in June, and the deal and the rep's confidence go with it. The card was right when someone wrote it, but nobody wrote down where the claim came from or when. This recipe makes every line traceable. Each competitor gets a document. Each thing they say about themselves is an *Evidence* finding that quotes their public page, links it and dates it; your answer sits beside it as an *Opportunity* that claims only what your product ships; and the things reps must not say are *Risk* findings called landmines. Your assistant does the first round of research from public sources, you check every line in the editor, and **Focus** and Revisions keep each card honest from month to month. Sales reads the cards from password-protected links. Win–loss interviews, pricing you learned in a deal and anything a prospect told you in confidence stay out of these cards. ## What you will use - **Shelf**: One shelf, *Battlecards*, with a line on how often each card is checked and who to tell when something is wrong. - **Research document**: A document per competitor; the overview says who they are, where you meet them and how deals are won. - **Finding**: Claims as *Evidence*, your answers as *Opportunity*, landmines as *Risk*, the ones every rep must know marked **Key**. - **Your AI assistant**: Claude or ChatGPT, with web search, researches each competitor’s own public pages and files a finding per claim with its links; it also checks your answers against your release notes. - **The editor**: Where you turn each claim into an exact, dated quotation, write the overview and the answers, and mark the landmines **Key**. - **Star and Focus**: Put a card in Focus for the week you are rechecking it, or when that competitor has shipped something. - **Revisions**: What a card said last quarter, who changed it, and Restore if a change was wrong. - **Public link**: A password-protected link per card for the sales team, replaced when the team changes. ## Method ### 1. Make the shelf and choose the competitors Make a shelf called *Battlecards* and give it a line: "Checked monthly. Wrong or out of date? Tell product marketing." Then list the competitors that actually show up in deals, say *Shipnest*, *Routewell* and *Carton*. Three good cards beat ten thin ones. Each will be one document, and its overview will be three short paragraphs: who they are and who buys them, where you meet them (company size, region, the moment in the deal), and in a sentence how deals against them are won. Reps read the overview in the minute before a call; keep it to what fits in that minute. ### 2. Have your assistant research from public sources only Ask your assistant to research on the web what the competitor says publicly about the areas your deals turn on: carriers, pricing model, returns, integrations. In the guidance, name the sources allowed: their website, documentation, changelog and press releases. Rule out review sites, forums, leaked material and anything behind a sign-in. This needs an assistant with web search; Claude and ChatGPT have it, and many local set-ups do not. It searches, reads the pages and files a document of findings on the *Battlecards* shelf, one per claim, each with its links. That document becomes the card once you have checked it. It is a first pass. Some pages will be out of date, some claims will be paraphrased too strongly, and some sources will not say what the finding says. ### 3. Check every claim, quote it and date it Open each source. Where the page supports the claim, rewrite the finding as a quotation of their exact words, add "Read on 12 September 2026", and file it as *Evidence*. Where it does not, delete the finding. Never paraphrase a competitor's claim into something stronger or weaker than they said. Then press **Edit**: retitle the document with the competitor's name, write the overview, and title each claim by its subject: "Carrier coverage", "Returns portal". Titles lead in Look up (⌘K), so when you search for *returns* before a monthly check, you land on every card that mentions it. ### 4. Write your answer beside each claim After each claim, add an *Opportunity* finding with your answer, in the editor or by asking your assistant to draft one from your release notes and then rewriting it: what your product does in the same area, in a sentence a rep can say aloud, with a link to your own public documentation or release notes. Link it back to the claim it answers by pasting the claim's address. Claim only what ships today. If your answer depends on something on the roadmap, it is not an answer yet; leave it out of the card until it is released. ### 5. List the landmines File the things reps must not say as *Risk* findings, titled plainly: "Do not say they have no returns portal". Explain why in two lines: it shipped in June, or you cannot source it, or it disparages them. Mark the ones that have already lost a deal as **Key**. Landmines are where most battlecards do their damage, because an out-of-date weakness sounds confident. This list is the part of the card to check first each month. ### 6. Share each card with a password Press **Share** on each card, copy the link the moment it appears, and set a password. Post the links and the password where sales already looks, such as the sales channel's pinned message. Leave the links with no expiry (the default), and when someone leaves the team, **replace** each link, change its password too, and post the new ones. Readers see the overview and every finding with its type and importance: claims, answers and landmines. They do not see your notes, so internal doubts ("check this with the carrier team") can sit in the margin safely. ### 7. Keep the cards fresh with Focus and Revisions Each month, put one card in **Focus** and ask your assistant a new research question: "What has Shipnest changed on its changelog and product pages since 1 September?" Compare it with the card and update the claims, answers and landmines. When the week ends, *Revisit* says so, which is your prompt to finish or move on. When sales asks what the card said before the last change, **Revisions** show every version with who changed it. The link's read count tells you whether reps opened the card before the quarter's big deals. ## Prompts to try Your assistant, connected to know.sh (Claude, ChatGPT or a local model): > Research on the web what Shipnest says publicly about its carrier coverage, pricing model, returns handling and integrations? Use only their website, documentation, changelog and press releases; no review sites, forums or anything behind a sign-in. Quote each claim in their words, link the page, say when a page is undated, and file it in know.sh as a new document on my Battlecards shelf, one finding per claim. Your assistant, connected to know.sh (Claude, ChatGPT or a local model): > Research on the web what Shipnest has announced or changed on its changelog, product pages and press page since 1 September 2026? Public pages only. One finding per change, quoting it, linking it and giving the date on the page, filed as a new document on my Battlecards shelf. Your assistant, connected to know.sh (Claude, ChatGPT or a local model): > Using know.sh, read the Shipnest battlecard on my Battlecards shelf. List every Evidence finding whose “Read on” date is more than 90 days old or which has no source link, and every Opportunity finding that does not link to our own documentation or release notes. Do not change anything. Your assistant, connected to know.sh (Claude, ChatGPT or a local model): > Using know.sh, read the Carton battlecard on my Battlecards shelf and the “Parcelwise release notes” document. For each Opportunity finding in the battlecard, check that what it says we do appears in the release notes as shipped. List any that do not, quoting the line. ## Variations - Keep a private *Objections* document of the questions prospects ask, each a *Question* finding with the answer, and copy the answer onto a competitor’s card when the question is about them. - Before a sales kick-off, ask your assistant for a quiz on a card and take it in know.sh; quizzes are for you, not the team, and each question cites the findings it came from. - When a landmine turns out to be a real gap, write it up as a requirement in the next [product requirements document](/cookbook/product-spec). - For interview evidence about why customers chose you, a [user research repository](/cookbook/research-repository) keeps it anonymised and apart from the cards. ## Where it falls short - Research needs an assistant with web search, and it reads public pages only. It cannot see what competitors say on calls or behind a sign-in, and it can misread a page, so every claim still needs checking. - Nothing watches competitors’ sites for you. The monthly check is yours to schedule; know.sh sends no alerts. - Reps cannot add what they hear in the field to the cards. It comes to you by message and you file it. - Links are per card, so there is no single link to every battlecard. Pin the list of links where sales can find it. ## A note on fairness and sources Never misrepresent a competitor. Quote their words, link the page and date it; if you cannot source a claim publicly, it does not go on the card. Comparative claims can carry legal risk, so have cards reviewed under your company's policy for marketing claims before they reach sales. Use public sources only: no leaked decks, nothing learned under a prospect's NDA, and no signing up for a competitor's product under false pretences. Your own answers claim only what ships. Assistants can make mistakes; check every claim against its source. Indexed under: Battlecards, Competitive intelligence, Competitor claims, Landmines, Sales enablement, Public sources.