GTM Chemistry Framework: GTM Discovery for Early-Stage Startups

1. What the GTM Chemistry Framework is

Everyone on a go-to-market team has opinions about customers, competitors, partners and who to call. Those opinions live in different heads, in different words, and never get compared side by side. So the team argues about tactics without having agreed on the diagnosis.

This method fixes that with a constraint: each area of discovery gets exactly one question and one sentence. One sentence is short enough that a room can align on it, and specific enough that it can be combined with the other three.

The combination is the point. Knowing your ideal customer is a diagnosis. Knowing your competitor is a diagnosis. Knowing both, in one sentence each, gives you a headline, because a headline is a promise of the outcome your customer wants, aimed at the gap your competitor left open.

Elements are diagnoses. Compounds are decisions. This exists to get you from one to the other in a single session.

2. Who it is for: early-stage startups, hackathon teams and pre-seed founders

I built this for teams that have to write positioning before they have the numbers that would justify it. If you have thousands of customers and a working analytics stack, go read the data instead. This is for the stage before that, when the honest answer to "who is this for" is a guess that nobody has written down.

Pre-seed and seed startups You have a product, a rough sense of the buyer, and no usage data worth reading. You need a landing page headline this week and a story for the next investor call.

Hackathon and weekend project teams You have 48 hours. The demo matters less than whether the judges understand who it is for and why it beats what already exists. Two of these questions, answered honestly, will fix most hackathon pitches.

Solo founders and indie builders No team to argue with, which sounds easier and is actually harder. The templates give you the argument you cannot have with yourself.

Accelerator and incubator cohorts Every company in the room fills the same four templates in the same session, so mentors can compare answers instead of sitting through eight different deck formats.

When it will not help you

Skip it if you already have a positioning document the team agrees on, if you have enough customers that qualitative guessing is worse than reading your funnel, or if your problem is execution rather than direction. This method produces a diagnosis and copy. It will not fix a product nobody wants, and it will not tell you whether your pricing is right.

3. How to use it

Same four steps for every element. The steps matter more than the tooling. A whiteboard, a Miro board and a shared doc all work.

Step 1 Ask the question One question per element, read out loud, answered individually before anyone discusses.

Step 2 Write the sticky A short, messy, honest answer. Fill-in-the-blank so nobody writes an essay.

Step 3 Compress to the outcome The stickies collapse into one structured sentence the room agrees on.

Step 4 React the pair Two outcome sentences combine and produce a concrete artifact.

4. The four elements of GTM discovery

Four areas of discovery, four questions, four symbols. Everything downstream is built out of these.

Who is your ideal customer, and why wouldn't they use your product?

02

Cp

market

Competitors Discovery

Who are your strongest competitors, and what makes them strong?

Who are the partners you want building alongside you?

Who do you need to talk to in order to leverage your product?

Ic · who you sell to Cp · who you sell against Pt · who you sell with Bd · who you sell through

Read the legend as a sentence and you have the shape of a go-to-market story: for whom, against whom, with whom, through whom. The four elements are the four prepositions.

01 Ic

ICP Discovery customer

Who is your ideal customer, and why do you think your ideal customer wouldn't use your product?

The second half does the work. Teams can describe their ideal customer fluently and have never once written down the reason that customer says no.

Sticky note My ideal customers are who, and they wouldn't use my product because why not.

Outcome template The ICP think my product is negative perception because real cause; what they actually want is desired outcome; the proof I deliver it is feature / case / number; therefore my headlines must promise outcome and neutralize objection.

02 Cp

Competitors Discovery market

Who are your strongest competitors, and what makes them so strong?

Name the strength honestly before you look for the gap. A team that cannot articulate why the leader is winning will invent a weakness that is not there.

Sticky note My strongest competitors are who and they're strong because they what they do well, but they miss the gap.

Outcome template Competitor is ahead because they did action / move and today they sustain it with feature / moat / distribution; but they are weak at real gap; I differentiate because value proposition, which matters especially to the segment they serve poorly; so I don't compete on their turf, I compete on my turf.

03 Pt

Partners Discovery ecosystem

Who are the partners you want building alongside you?

Partners are distribution you do not have to buy. The test is not whether the logo is impressive, it is whether their builders overlap with your ideal customer.

Sticky note The partners I want building alongside me are who because they bring what, and together we can do what.

Outcome template Ecosystem is growth signal; its builders are ICP fit; expected pipeline $X; entry via warm path; the incumbent is weak / absent / locked in, so the play is campaign / partnership / day-one docs listing.

04 Bd

BD Discovery network

Who are the people you need to talk to in order to leverage your product?

This element is the only one that ends in a calendar entry. If the sentence does not name a person and a date, it is not finished.

Sticky note The people I need to talk to are who, I can reach them through how, and they can help me with what.

Outcome template The people I need to reach are role / profile and they're concentrated in community / event / network; the warm path is person / channel / content; what I offer first is value I give before asking; what they unlock is intro / pipeline / credibility / feedback; the next step is action within 7 days.

5. The structural formula

Drawn the way a skeletal formula is drawn. The four elements sit in a row along the top, and every bond drops to the compound that pair produces. Depth carries meaning: two elements standing next to each other bond in the shallow tier, and the two at opposite ends of the row bond at the bottom of the sheet.

6 bonds · 6 compounds Deeper bond · less obvious combination All four together · the GTM narrative

The bottom tier is the one teams never reach on their own. Ic and Bd sit at opposite ends of the row, so nobody thinks to put the ideal customer and the personal network in the same sentence, and that combination is the warm intro map: the named people who can get you in front of the buyer you just described. The two middle-tier bonds have the same property, which is why the drawing puts them below the obvious ones rather than beside them.

6. The seven compounds

Each reaction reads as a formula: two outcome sentences in, one artifact out. The first three are the ones teams find on their own. The next three are the ones the drawing above puts in the lower tiers.

Ic + Cp → Marketing headline & UI copyPromise the outcome your customer actually wants, aimed at the gap the competitor left open.

Cp + Pt → Ecosystem positioningOwn the turf the incumbent left open, inside the ecosystem where it is still unclaimed.

Pt + Bd → Alliance pipelineWarm paths into partnership deals, ranked by who can actually sign.

Ic + Pt → Co-marketing & integration roadmapBuild where your users already are instead of asking them to come to you.

Cp + Bd → Competitive intel networkThe people who already know why customers switch, and will tell you if you ask well.

Ic + Bd → Warm intro mapThe named people who open the door to your buyers, and the path to each one.

Ic + Cp + Pt + Bd → The GTM narrativeFor whom, against whom, with whom, through whom, in one paragraph anyone on the team can repeat.

7. A worked case

CatTrade is a fictional trading platform whose ideal customer already uses the market leader, DogTrade. The whole session, filled in.

Open the CatTrade case · four elements filled, three compounds produced

Ic · ICP discovery

Sticky note My ideal customers are active retail traders, and they wouldn't use my product because DogTrade is already their trusted favourite and they think CatTrade isn't safe, they've never heard of us.

Outcome The retail traders think my product is unknown and unsafe because they've never heard of CatTrade and switching away from DogTrade feels risky; what they actually want is to trade with confidence, without fearing for their money; the proof I deliver it is regulated custody, audits, 40k traders onboard; therefore my headlines must promise total security and neutralize "is my money safe here?".

Cp · competitors discovery

Outcome DogTrade is ahead because they did a three-year brand campaign while everyone else fought on features and today they sustain it with name recognition and a default-app habit; but they are weak at fees, which they have never cut because they don't have to; I differentiate because the same regulated custody at a third of the cost per trade, which matters especially to high-frequency retail traders whose fees eat their margin; so I don't compete on brand trust, I compete on cost per trade for people who trade often.

Pt · partners discovery

Outcome Trading-bot and portfolio-tracker tools is growing 3x year on year while exchanges stay flat; its builders are exactly our high-frequency traders, wearing a developer hat; expected pipeline $400k; entry via the two bot authors already in our Discord; the incumbent is absent, DogTrade has no public API tier, so the play is a day-one docs listing plus a revenue share for integrations.

Bd · BD discovery

Outcome The people I need to reach are trading-community moderators and bot authors and they're concentrated in three Discord servers and one annual conference; the warm path is the bot author who already reviewed our API; what I offer first is free API access and a fee report they can publish; what they unlock is credibility with an audience that distrusts new exchanges; the next step is send the fee comparison to two moderators by Friday.

The compounds that came out

Ic + Cp → headline Everything you love about trading. Fees you won't find on DogTrade. Fully regulated.

Cp + Pt → Ecosystem positioning"The exchange that trading bots are built on." DogTrade has no API tier, so the developer surface is unclaimed and it reaches the same high-frequency traders from a direction the incumbent cannot follow.

Ic + Bd → Warm intro mapBot author in the Discord, then the three moderators he vouches for, then their audiences. Each step is a person who already has the trust CatTrade is missing, and the fee report is what gets handed over first.

What the reaction surfaced: "nobody knows us" is a distribution problem, because headlines do not reach people who never land on your site. "They already trust the incumbent" is a messaging problem, because headlines have to prove safety and give one concrete reason to switch. Separate the two before writing a single line of copy.

8. The launch compound

The final reaction. Every slot in it is already answered by one of the four elements, so a blank tells you which element to go back to.

Launch template We're launching product / feature to ICP where they already are: channel; the hook is message tied to a pain or trigger; success in X days looks like one metric; if it hits, we scale via next channel / motion.

One product, one audience, one channel, one hook, one metric, one next move. Anything that does not fit the sentence is next quarter's problem.

9. Prompts to run the reactions

Once the four elements are answered, the sheet is a prompt. Paste the two outcome sentences, name the reaction, ask for the artifact.

The six reactions, written out. Copy the one you need.

Ic + Cp → Headline & UI copy

Hey Claude, here are two outcomes from my GTM Chemistry sheet. ICP: [paste]. Competitors: [paste]. React them and write five landing page headlines that promise the outcome my ICP actually wants while attacking the gap the competitor left open. No claim I cannot back with the proof point in the ICP sentence.

Cp + Pt → Ecosystem positioning

Hey Claude, react my Competitors outcome with my Partners outcome: [paste both]. Give me one positioning line that claims the turf the incumbent left open inside that ecosystem, and the three proof points I would need to defend it.

Ic + Bd → Warm intro map

Hey Claude, react my ICP outcome with my BD outcome: [paste both]. Build the warm intro map: who to reach, through whom, what I give before asking, and a first message for each path that does not read like a sales email.

Bd + Pt → Alliance pipeline

Hey Claude, based on the reaction Bd + Pt from my GTM Chemistry sheet: [paste both]. Rank the partnership targets by who can actually sign, then draft the first outreach for the top three and tell me what I am offering each one before I ask for anything.

Cp + Bd → Competitive intel network

Hey Claude, react Cp + Bd: [paste both]. Give me eight questions for the people who already know why customers switch, phrased so they answer honestly instead of hearing a pitch.

Ic + Cp + Pt + Bd → The GTM narrative

Hey Claude, here are all four outcomes from my GTM Chemistry sheet: [paste]. Write the GTM narrative in one paragraph, for whom, against whom, with whom and through whom, short enough that anyone on the team can repeat it from memory.

The templates do the hard part. A model given two vague sentences returns vague copy, which is the whole reason the outcome templates are rigid.

10. Where it came from, and what is actually new

The ingredients are not new. ICP discovery, competitive analysis, partner mapping and business development are standard go-to-market blocks, and crossing quadrants to generate strategy has an ancestor in the TOWS matrix. What I have not found elsewhere is this combination: four discovery areas compressed into one rigid sentence each, with the pairwise combinations mapped to named artifacts rather than general findings. The classic methods stop at diagnosis. This one is built to force output.

It started on a Miro board, four quadrants along the top and lines converging into the artifacts underneath. The funnel made the logic obvious and the chemistry metaphor made it teachable. It is a living document: as I run it with more teams, some of these templates will get thrown out.

Want to run this with your team? Find me on Twitter / X. If you run it yourself, I would genuinely like to see what came out.

← Back to Growth