# Versus > Decide which competitor to take on, and whether naming them helps or hurts, using battle rap matchup logic - the ladder, the crowd, the style matchup, and their response. Use this whenever the user is choosing a competitor to position against, asking whether to name a rival in their copy, planning a comparison or alternative page, building a battlecard, reacting to a competitor's launch, pricing move, or funding round, or saying they keep losing deals to someone. Also use when the user asks who their real competitors are, whether to respond to a competitor's attack, or whether a fight is worth taking. This picks the fight. It does not write the copy. - Skill: `michaeldoesmarketing/versus` (Agent Skill, multi-file: 3 files) - Install (CLI): `npx skillmds@latest add michaeldoesmarketing/versus` - Raw SKILL.md: https://api.skillmd.com/api/skills/michaeldoesmarketing/versus/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: MichaelDoesMarketing (https://skillmd.com/u/michaeldoesmarketing) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/michaeldoesmarketing/versus --- # Versus Marketing picks its opponents badly. Brands attack whoever is nearest, whoever annoyed them last quarter, or whoever the sales team complains about, and then act surprised when the comparison page does nothing or, worse, sends traffic to a competitor. Battle rap solved opponent selection a long time ago, because the cost of a bad matchup is public and immediate. You do not pick a fight because you dislike someone. You pick it because of what the matchup does to your standing when it is over, win or lose. The load-bearing idea: a matchup is decided before anyone speaks. Position on the ladder, who owns the crowd, whether your strengths reach their weaknesses, and what they can say back. Execution matters less than selection. Use the logic. Do not perform the frame. The output is a strategic call in marketing language. Never announce that this came from battle rap unless the user asks. ## What this skill does and does not do This decides whether to take a fight, which one, and on what terms. It stops at the verdict and the terms. It does not write the comparison page, the ad, or the headline. When the verdict is TAKE IT, hand the angle and the terms to whatever writes the copy. ## Modes Pick by what the user brought. If it is ambiguous, pick one and say which. **Scout the division** - no target yet. The user knows they need to position against something but not what. Run all six steps with the field wide open. **Size up** - the user already has a target in mind. Run all six steps against that one matchup and return a verdict. This is the most common entry. **Ducking** - find the opponent the user keeps avoiding. Usually the real leader in the category, sometimes the status quo, sometimes an uncomfortable truth about who they actually lose to. Run steps 1, 2, and 5 on that opponent specifically. **Fight card** - the full division ranked, the user placed honestly inside it, every plausible matchup called. Use `assets/fight-card.md`. This is the one people screenshot, and it only works if the user's own placement is honest. Do not flatter the placement. ## Step 1: Fill the card The user's list of competitors is usually wrong. It is built from who they see at conferences, not who takes their money. Establish, inferring what you reasonably can and researching where you can: - Who the buyer actually evaluated before choosing, which is rarely the same as the market map - What the buyer does today if they do nothing, because that option is on the card whether or not anyone lists it - Who the user loses to, and for what stated reason - Who the user beats, and why Four types show up on every card. They are fought differently. Read `references/ladder.md` for how each one behaves: - **The giant** - bigger, better known, more budget - **The peer** - roughly your size, same buyers, real overlap - **The ghost** - the status quo, the spreadsheet, the in-house build, doing nothing at all - **The crowd** - the buyer's own skepticism, which is the opponent when the category has burned people before Most brands pick a peer when their actual opponent is the ghost. That is the most common misidentification and it wastes entire campaigns. ## Step 2: Place yourself on the ladder Punching up transfers their attention to you. Punching level is a real fight with real risk. Punching down transfers your attention to them, for free, and makes you look worried in the process. Be honest about placement. Awareness, not quality. A better product with a tenth of the recognition is punching up, and that is fine, because up is where the gain is. `references/ladder.md` has the mechanics of each direction and what it costs. ## Step 3: Read the crowd The audience decides, and some rooms cannot be won at any level of execution. Work out whose audience is watching, and whether that audience is loyal or shopping. A crowd that loves the opponent will hear an accurate criticism as an attack on themselves. A crowd that is already unhappy with the opponent will finish your sentences for you, which is the best room to fight in. Read `references/crowd.md`. ## Step 4: Style matchup Ranking is not the whole story. Some opponents are above the user and still bad matchups for themselves, because their strength is exactly the thing the user's buyers have stopped caring about. Find the pairing. The user's real strength against the opponent's structural weakness, where structural means the weakness comes from how the opponent is built and cannot be patched in a quarter. A weakness they can fix by Friday is not a matchup, it is a news cycle. Read `references/style-matchup.md`. ## Step 5: Write their response This step is mandatory and it is where most fights die. Before writing a single line of attack, write the round they would drop on the user. Their strongest true answer, in their voice, aimed at the user's softest real spot. Not a strawman. The version their best marketer would ship. Then compare. If their answer is stronger than the user's opening, the fight is a loss on paper and the verdict is LEAVE IT, no matter how much the user wants it. Read `references/the-response.md` for the standard comebacks and how to write theirs. ## Step 6: Score and call it Score the matchup on six dimensions, apply the kill rules, and return a verdict with terms. Dimensions, thresholds, and verdict definitions are in `references/scoring.md`. The template is `assets/matchup-scorecard.md`. The four verdicts: - **TAKE IT** - name them, position directly, commit - **TAKE IT SIDEWAYS** - the matchup is real but naming them is not worth it. Fight the category habit or the behavior instead of the brand - **LEAVE IT** - the fight loses. Say why, and say what the user should do instead - **DON'T SAY THE NAME** - taking this fight actively helps them. Usually punching down, sometimes a beloved incumbent Always give the verdict before the reasoning. The user needs the call, not the buildup. ## Output format Deliver in this shape unless the user asks for something else. Prose, no long preamble. ``` THE VERDICT One line. Take it, take it sideways, leave it, or don't say the name. THE CARD Who is actually in the division, including the ghost. One line each. THE MATCHUP Ladder position, crowd, and style pairing, in three or four sentences. THEIR RESPONSE The round they drop back, written straight. Then whether it beats the opening. THE TERMS If the verdict is take it: the angle, what can be backed, what has to stay unsaid. If the verdict is anything else: what to do instead. THE SCORECARD Six dimensions scored, kill rules flagged. ``` ## Rules of engagement **Write their round first.** Any fight recommended without their response written is not a recommendation, it is a guess. If their best answer beats the user's best opening, the verdict is LEAVE IT regardless of appetite. **Every comparative claim needs a receipt that is public and dated.** A source anyone can check, with a date attached, because competitor facts expire. No receipt means the claim gets cut, however true it feels. An uncheckable comparative claim is a legal problem and a credibility problem in the same week. **Never punch down.** Not on manners, on effectiveness. Naming someone smaller hands them your audience, your credibility by association, and a permanent place in your story. If the user insists, tell them what it costs and let them choose. **Never say a name for free.** Every mention of a competitor is impressions you gave them. If the mention does not earn more than it gives, cut the name and keep the argument. **Swing at positioning, claims, pricing behavior, and category habits.** Never at founders, employees, or anyone's identity, personal life, or protected characteristics. Never invent a weakness. A fabricated flaw is the fastest way to lose a fight you were winning. **Do not confuse a fight with a strategy.** Some categories are won by being the only one in the room talking about something else. If the honest answer is that no fight helps, say that.