Targets: who each target is in the world
A positioning names its targets: what each one needs, what blocks them, what to lead with. Targets adds who that person is in the world. A message can then be written to them. A contact list can be built from them.
The card
One card per target of the positioning chosen at the top. The top of a card comes from the positioning and is edited there. The fields under it are the card's own. Separate entries with commas.
- Job titles: as they read on a profile. "Chief Financial Officer, VP Finance, Controller", not "finance leader".
- Seniority: Owner, Founder, C-suite, VP, Director, Manager, Individual contributor.
- Department: the function they sit in. Leave empty for a private individual.
- Company size: employee bands, 1-10 up to 5001+.
- Industries: narrow beats broad.
- Geography: only where the business is limited.
- What makes them buy now: events you could observe.
- Who to rule out: people who match on title but are not a fit, and why.
- Role in the decision: economic buyer, champion, influencer, blocker, end user.
- What they do today instead: the workaround or the provider they already use.
- How they say it: their own words. You fill this in; AI Assist never does.
- In one sentence: the card at a glance.
What HubSpot shows
A positioning is an inference. Your portal is a witness. The panel at the top is the portal's record. It is read every night and when you press Check against HubSpot.
Four measures give the portal one of three postures. The posture decides how far the AI leans on it.
- Rich: fifty or more customers, key fields filled, customers different from everyone else. The AI leans on the portal.
- Messy: data, but thin or uneven. The AI uses only the filled fields. The positioning decides.
- Blank: under fifty contacts. The positioning decides. Normal for a company new to market.
Where a card says one thing and your customers show another, the panel says so in one line. Nothing changes by itself.
Inferred and confirmed
A card starts as an inference from the positioning. Check against HubSpot adds evidence. It counts how many contacts match each saved card.
- Confirmed: ten or more matches, and one in a hundred of the sample.
- A few: some matches, too few to confirm.
- Inferred: none yet. Not a fault. It is the list to go and build.
The panel also shows common job titles and industries no card covers. Add them where they belong. The check reads titles, industry, size, country, stage, persona and two dates. No names or email addresses leave HubSpot. Nothing is written back.
AI Assist
AI Assist at the top drafts a card for every target that has none. AI Assist on a card redrafts that one. The draft lands in the fields. Nothing is saved until Save.
Where a card is used
In a pool, Draft rows with AI has a Target choice. Pick one and the rows are written to that person. A card changes nothing in the positioning. Renaming a target keeps its card.
The ICP dossier
Open on a card shows one ICP on one page. Left: the name and the summary. Right: the trait wheel. Five axes in buying words: Novelty, Rigor, Pace, Consensus, Risk comfort, 0 to 100. The dashed shape is inferred from the card. A solid shape appears when results earn it.
Draft mentality fills the rest in one AI call. How they decide. What moves them. Say this. Not this. The proof they trust. The objections you will hear. Every value is from a fixed list. The DISC label comes from the five scores, in code. Edit the say and not-this lines and press Save lines. The Composer and the Architect use them when they write for this ICP.
Where they gather lists the ICP's confirmed events. What worked for them shows the earned numbers once 30 contacts were reached. Example first line is one line the model wrote from the say and not-this lines. It carries no number and no name. Profiles are for ICPs only. Nothing is scored or stored on a contact.
The family
Each target is mapped to a shared family in code: role family, industry group, size band. Fill the titles, industries and sizes and the family follows. Your agency's Portfolio and the benchmarks read the family, never your words.
Results by target
Each send is matched to one target in code, from the contact's title, industry, company size and country. Fill the card's titles, industries, sizes and disqualifiers and the match improves. Radar, Claims then shows which claims worked for each target.