Guide

How to do account research for sales

Account research for sales is the disciplined process of learning what matters about a named company before deciding whether, why, and how to engage it. Good research connects current company evidence, relevant people, relationship history, and the team's qualification criteria. The output is not a dossier. It is a concise, dated brief that makes the next decision easier to review.

See how Account Intelligence supports named-company research

What account research for sales is

Sales account research starts with a company the team already cares about and a decision it needs to make. That decision might be whether to prioritize the account this week, how to prepare for a discovery call, which stakeholder needs further research, or whether new evidence changes an existing account plan.

The work combines four kinds of context:

Company context: what the business does, how it operates, and what has changed.
People context: which roles may be relevant to the sales thesis and what those people have said or done publicly.
Relationship context: what the CRM, account owner, and prior conversations already establish.
Decision context: the team's fit criteria, exclusions, current hypothesis, and proportionate next move.

This is narrower than general market research and deeper than contact enrichment. A contact database can help identify a person. Account research explains why the company and person may deserve attention now, which parts of that conclusion are supported, and which parts still need validation.

Nearby activity	Primary question	Typical output	What it does not establish
Contact enrichment	Who is this person and how can we reach them?	Profile and contact fields	Why the account matters now
Lead scoring	Which records match a model?	Rank or score	The complete reasoning behind the rank
Account planning	How should we develop an important account?	Goals, stakeholders, risks, and actions	Whether the underlying research is current
Account research	What is true now, why might it matter, and what should we verify next?	Dated evidence brief	A private buying plan or guaranteed opportunity

Start with the sales decision

Research expands to fill the time available unless the team names the decision first. “Research this account” is too broad. A usable question identifies the account, the sales context, and the decision the research must support.

Examples of bounded questions include:

Has anything changed since this account was last reviewed that could affect its priority?
Which two or three roles are most relevant to our current thesis, and why?
What public evidence should the account owner understand before the next meeting?
Does recent evidence support the current account hypothesis, contradict it, or leave it unchanged?
What must be verified in the CRM or directly with the buyer before action?

The question determines the depth. A weekly prioritization review needs a short change brief. A strategic-account meeting may justify a deeper operating model, stakeholder, and relationship analysis. More pages of research do not automatically create a better sales decision.

A seven-step account-research workflow

1. Write the account thesis

State why this kind of company could be relevant before looking for supporting evidence. Record the fit criteria, disqualifiers, problem hypothesis, likely operating trigger, and roles that would normally own or influence the problem.

This prevents the researcher from treating every company announcement as a reason to sell. It also makes contradictions useful: evidence can change the thesis instead of being forced to confirm it.

2. Establish the company baseline

Capture the stable facts needed to interpret later changes: business model, customers, geography, company structure, relevant products or services, and operating constraints. Prefer the company's own site, documentation, filings, careers pages, and official announcements for claims about the company.

Keep the baseline concise. Only include facts that affect qualification, stakeholder relevance, timing, or the proposed next move.

3. Look for dated changes

Search for material changes since the last review: leadership appointments, hiring patterns, acquisitions, product or pricing changes, market expansion, public contracts, partnerships, or shifts in executive language. Save both the event date and the date checked.

A change is not automatically a buying signal. Ask how it relates to the team's written thesis, how recent it is, and whether a second source or later event strengthens or contradicts the reading.

4. Map relevant people cautiously

Identify roles that own the problem, experience its consequences, control budget or procurement, influence technical evaluation, or already have a relationship with the team. Confirm current role and company before using a person's activity as account evidence.

Do not turn a large organization chart into a fictional buying committee. Public information rarely proves who is involved in a specific unpublished purchase. Mark people as relevant to the thesis, not as confirmed decision-makers, unless a reliable source establishes that role.

5. Add CRM and relationship context

Check ownership, open opportunities, prior conversations, customer status, exclusions, and known relationships. A public trigger can be real while the proposed action is still wrong because an AE is already in an active conversation or the person moved companies.

When a CRM match is absent or uncertain, verify manually. “No match found” is not the same as “no relationship exists.”

6. Separate observation from interpretation

Write the brief using four explicit layers:

Observed: what the source shows, who or what it concerns, and when.
Inferred: why that evidence may matter under the team's account thesis.
Unknown: private priorities, ownership, budget, timing, or relationships the evidence cannot confirm.
Next move: the smallest sensible action that tests or uses the inference.

This structure makes weak reasoning visible. It also gives the account owner a way to disagree with the interpretation without losing the underlying evidence.

7. Assign a disposition and review date

End with a decision: advance for human review, verify a specific unknown, watch for a defined change, keep at its current priority, or park the account. Name the owner and the date when the reading should be revisited.

Research without a disposition becomes an archive. A disposition without evidence becomes an opaque score.

Which sources should you use?

Use the source closest to the fact being claimed. Search results, aggregators, and generated summaries can help discovery, but the brief should link to the original material whenever practical.

Source	Useful for	Common limitation
Company website and documentation	Business model, products, positioning, locations, current language	Tells the company's preferred story
Official news, filings, and public contracts	Dated company events and formal commitments	Coverage varies by company type and geography
Careers pages and job descriptions	Team investments, responsibilities, systems, and operating priorities	An open role does not prove budget for your category
Executive and employee public activity	Language, stated problems, initiatives, and professional attention	One person's activity may not represent the account
Trusted reporting and industry publications	External context and event confirmation	May lag or simplify the operating detail
CRM and conversation history	Ownership, prior objections, relationships, and deal state	Internal records may be stale or incomplete

Record the source URL, publication or event date, date checked, and a short excerpt or faithful paraphrase. If freshness matters and no reliable date is available, label it as an unknown rather than silently treating it as current.

A decision framework for account priority

Do not collapse the full account reading into one unexplained score. Review the dimensions separately, then choose a disposition that matches the evidence.

Dimension	Question to answer	Stronger evidence	Reason to slow down
Fit	Does the company meet the written account criteria?	Several relevant, current facts	Qualification depends on an assumption
Change	Has something material changed?	Dated primary source tied to the thesis	Generic news with no operating consequence
People	Are relevant current roles identifiable?	Role responsibility is explicit and verified	Title is stale or relevance is guessed
Evidence	Can a reviewer inspect the basis?	Original sources and independent corroboration	Summary without provenance
Freshness	Is the evidence current enough for the decision?	Event and checked dates are recorded	Old activity is presented as new
Relationship	What does the team already know?	Clear owner and current CRM context	Missing or conflicting records
Contradiction	What weakens the thesis?	Conflicting facts are included	Only confirming evidence was collected

Use those dimensions to choose one of five outcomes:

Advance: fit is clear, the evidence is timely and inspectable, and a human owner can review a specific next move.
Verify: the thesis may be useful, but one material fact—such as role, ownership, or timing—needs confirmation first.
Watch: the account fits, but no current change justifies a different priority. Define what change would trigger review.
Keep steady: new research adds context but does not change the existing account plan.
Park: the account fails a qualification gate, the evidence is too weak, or a known relationship makes the proposed motion inappropriate.

What should an account brief contain?

A practical account brief should be short enough to review before a call or weekly account meeting. Include:

account name, owner, segment, and research as-of date;
the decision this research supports;
current fit and disqualifiers;
two or three material observations with source and date;
relevant people, current roles, and why each role may matter;
CRM and relationship context;
the account-level inference and confidence-lowering evidence;
important unknowns;
one proposed next move, owner, and review date.

Leave out company-history trivia that does not change qualification, timing, stakeholder relevance, or action. The brief should let a reviewer trace every material conclusion back to evidence without forcing them to repeat the research.

A synthetic account-research example

The following example is fictional. The company, people, and events are synthetic and do not describe a Lumnis customer or a real buying process.

Sales context: LedgerNorth sells purchasing software to multi-site service businesses. Its team is reviewing Harbor Field Services, a fictional named account that was closed-lost 11 months ago because standardization was not a priority.

Observed as of August 27, 2026:

Harbor Field Services published an August 18 job description for a director of procurement systems. The role is responsible for standardizing purchasing across 12 operating locations.
An August 20 company announcement says Harbor acquired two regional operators.
On August 24, the fictional finance leader wrote publicly that post-acquisition systems integration was underway.
The CRM shows no open opportunity. The prior contact has left, and the previous account owner is still assigned.

Inferred: The acquisition and role description may make purchasing standardization more relevant than it was during the prior evaluation. The finance leader and incoming procurement-systems role are relevant to the thesis, but neither is confirmed as an active buyer.

Unknown: The team cannot confirm whether Harbor has approved a project, selected a vendor, allocated budget, or assigned ownership. It also does not know whether another internal relationship exists outside the CRM record.

Next move: The account owner reviews the old opportunity notes, confirms current ownership, and decides whether to ask a narrow discovery question about the public standardization initiative. If role verification fails or the acquisition integration is unrelated, the account returns to watch status.

The example is useful because the evidence changes the research question, not because it proves intent.

An account-research checklist

Before handing a brief to sales, confirm that:

the decision and account thesis are written;
stable facts are separated from recent changes;
material observations link to original sources when practical;
event date and date checked are recorded;
each relevant person's current role is verified;
person-level activity is not presented as coordinated account intent;
CRM ownership, relationships, exclusions, and active deal state were checked;
observed, inferred, unknown, and next move are visibly distinct;
contradictory evidence is included;
the proposed action is proportionate to the evidence;
a human owner and next review date are explicit.

Limitations of sales account research

Public research is incomplete by design. It cannot reveal unpublished budgets, internal politics, project status, procurement constraints, or the buyer's private priorities. Even a primary source can become stale after it is published.

Research can also introduce bias. A visible executive may not own the problem. A prolific poster may not represent the company. A researcher looking for urgency can overvalue exciting events and ignore quiet contradictions. Tool coverage varies by source, geography, industry, language, and privacy setting.

The answer is not endless collection. Use dates, provenance, explicit unknowns, human review, and a proportionate next step. For important accounts, validate public evidence through direct conversation and current relationship context.

How Lumnis fits into account research

Lumnis Account Intelligence begins with a company the team already cares about and turns the recurring research job into one reviewable account reading. It can examine company changes, relevant people, and attributable public activity against the team's written criteria; preserve the source, event date, and research date; and explain what was observed, inferred, and still unknown.

When supported CRM context is connected, the review can also account for whether the person or company is already known, available lifecycle or deal context, and whether an active deal exists. That context helps a reviewer avoid treating every new public observation as a new relationship. Availability varies by connected CRM and record; Lumnis does not claim to ingest private conversation history or replace the account owner's knowledge.

The operational difference from manual account research is continuity. Instead of rebuilding a brief from browser tabs each time, a team can preserve the previous thesis, check multiple signal types for material change, and revise the reading without separating the conclusion from its evidence. The same structured reading can be inspected by a seller or supplied as grounded context to an AI agent, while qualification and consequential action remain reviewable decisions.

For research that must be revisited, continuous account research explains how to compare the current reading with the previous one. For downstream handoff, use the evidence-preserving process in operationalizing buyer signals. Lumnis should be treated as a research and judgment layer, not as proof of private intent or an autonomous sales system.

Frequently asked questions

How long should sales account research take?

Depth should match the decision. A weekly priority check may need only a short change review, while a strategic meeting may require deeper company, stakeholder, and relationship work. Stop when the material fit, timing, people, evidence, unknowns, and next move are clear enough for a reviewer to decide.

What are the best sources for account research?

Use sources closest to the fact: the company's site and documentation, official announcements or filings, current job descriptions, attributable public activity, trusted reporting, and the team's own CRM history. Use aggregators for discovery, then verify material claims at the original source when possible.

Is account research the same as account planning?

No. Account research establishes a current evidence base. Account planning uses research plus relationship knowledge, goals, stakeholder strategy, risks, and coordinated actions to develop an account over time. Research should inform the plan and be refreshed when the underlying facts change.

Does a company change prove buying intent?

No. A change may create a relevant hypothesis, but it does not confirm an active project, budget, vendor evaluation, or private plan. Label the observation, explain the inference, state the unknowns, and choose a next move that tests the hypothesis.

Should sales research identify the full buying committee?

Only when evidence supports it. Public research can identify roles likely to own, influence, use, approve, or procure a solution. It usually cannot confirm the exact committee for an unpublished purchase. Describe relevant people cautiously and verify roles through CRM context or direct conversation.

Can account research be automated?

Collection, change detection, summarization, and brief preparation can be assisted. Qualification, relationship checks, interpretation, and consequential action still need a visible review boundary. Automation should make the evidence easier to inspect, not hide uncertainty behind a score or trigger an unsupported action.

Build a reviewable account-research habit

Start with one named sales decision, then let Lumnis keep the company evidence, relevant people, CRM-aware context, recency, and reasoning together in a brief the account owner can challenge. If the account remains important, preserve the thesis, define what to watch, and review only the changes that could strengthen, weaken, or replace the current reading.

See how Account Intelligence keeps the reading reviewable · Learn the continuous-research method · Browse all resources