Account Research

Account Research

Purpose

Walk into the call already knowing who they are, what changed recently, who's likely in the room, and what you'll bet is hurting them — so the call is spent testing hypotheses, not gathering facts you could have found beforehand.

Inputs

• Company name and/or domain
• Any prior relationship (past deals, contacts, meetings) if it exists
• Your product and the pains it removes (so research maps to value)
• The call type and goal (first outbound, inbound demo, expansion)

Method

Work the five blocks below, then collapse them into the one-page brief. Skip nothing — a missing block is a blind spot you'll discover live on the call.

1. Firmographics (the baseline).

• Industry / sub-vertical, what they actually sell, and to whom
• Headcount + growth trend (flat, hiring hard, RIF'd recently)
• Revenue band / funding stage, ownership (PE, VC, public, bootstrapped)
• HQ + key locations, and whether they're multi-entity
• Tech stack signals relevant to you (e.g. CRM, CI, comms tools)

2. Recent triggers (the "why now"). Hunt for events in the last ~6 months:

• Funding round, M&A, or new market entry → budget + change appetite
• Leadership change (new VP/C-level in your buyer function) → new mandate, fresh eval
• Aggressive hiring in the function you sell to → scaling pain
• Product launch, repricing, or strategy pivot → operational strain
• Press, earnings commentary, layoffs, or regulatory pressure → priorities shifting Each trigger becomes a line you can open the call with.

3. Org structure & likely buying committee. Map the roles, not just names:

4. Map likely pains to your product. For each plausible pain, name the evidence and the bridge to your value:

Pain hypothesis → Trigger/evidence that suggests it → Your capability that removes it → Proof point

Three strong mapped pains beats ten generic ones.

5. Hypotheses to test on the call. Turn research into questions, not assertions:

• "You raised a Series B in March — I'd guess the team's scaling faster than your current process. Is that landing on ?"
• "Saw you hired three s last quarter — usually that means . True here?" Each hypothesis is falsifiable and earns you a discovery answer either way.

One-page account brief output format:

ACCOUNT: <name> (<domain>)
Snapshot: <industry> · <headcount> · <funding/ownership> · <HQ>

WHY NOW (triggers):
- <trigger> → <implication>
- <trigger> → <implication>

LIKELY BUYING COMMITTEE:
- Economic buyer: <name/title or UNKNOWN>
- Champion: <name/title or UNKNOWN>
- Other: <names/roles>

MAPPED PAINS → VALUE:
- <pain> → <your capability> (<proof point>)
- <pain> → <your capability> (<proof point>)

HYPOTHESES TO TEST:
1. <falsifiable question>
2. <falsifiable question>

OPEN QUESTIONS / GAPS:
- <the most important unknown to close on the call>

PRIOR RELATIONSHIP:
- <past deals / contacts / meetings, or "net new">

Tool binding

This skill works from a company name alone. It gets sharper when connected to your stack — strongest with Doris, the reference integration, when there's prior history.

With Doris (recommended)

If the Doris MCP (mcp.meetdoris.com) is connected and the account already exists in your world, pull its real history instead of starting cold:

• ontology_resolve("account", id, expand=["deals","contacts","meetings","activity","sentiment"]) — deals gives prior/open deals and outcomes, contacts seeds the buying-committee map with people you already know, meetings surfaces what's been discussed, activity shows the recent timeline (the real "why now"), and sentiment tells you whether the relationship is warm or cooling.
• ontology_search / ontology_list to find the account by name or domain and list accounts in the same segment for pattern reference.
• search_transcripts(...) to quote what people at this account have actually said in prior calls — gold for hypothesis framing. Prefer Doris-derived contacts, meetings, and sentiment over re-deriving from public data — they reflect what's true in your relationship, not just the open web.

With a CRM / CI / email MCP

• CRM MCP (Salesforce/HubSpot) → pull the account + contact records, open/past opportunities, and owner/activity history.
• Apollo / LinkedIn (or a data-enrichment MCP) → firmographics, headcount/growth, and the org chart for the buying-committee map.
• Web search MCP → recent triggers (funding, hiring, leadership change, news, product launches).
• CI MCP (Gong/Chorus/Fireflies) or an email MCP → any prior conversation content.

With nothing connected

Fully manual path — needs only a company name or domain and the user's notes:

1. Ask the user for the company name/domain, their product, and the call goal.
2. Use web search (or have the user paste): the company website/about page, LinkedIn company page (headcount + recent hires), Crunchbase/PitchBook (funding/ownership), recent news/press, and the leadership/team page for the org map.
3. Work the five Method blocks from that public info plus anything the user already knows about the account.
4. Output the one-page brief, marking every unverified field as a hypothesis and every missing role as UNKNOWN so the gaps are explicit going into the call.

Works without Doris

Fully functional from a company name and public sources — Doris only removes the manual gathering and adds your real relationship history (prior deals, known contacts, past-call quotes, and sentiment) that the open web can't give you.

Common mistakes

• A brief that's all firmographics and no "why now" — facts without a reason to call.
• Listing names without mapping them to buying-committee roles and stances.
• Generic pains not tied to a trigger or to your specific capability.
• Hypotheses written as assertions you'll defend instead of questions you'll test.
• Ignoring prior relationship — pitching an account you've already lost or are already working as if it were net new.