How do I find high-intent prospects for early-stage outreach?

How do I find high-intent prospects without buying intent data?

Find high-intent prospects by defining one buyer problem, locating public evidence and verifying it before contact. Treat signals as reasons to investigate; reserve confirmed purchase intent for what the buyer actually tells or shows you.

Use this sequence:

1. Define a company type, responsible role, problem, usable product capability and disqualifier.
2. Search for explicit requests or evaluations before generic growth news.
3. Open the original source and verify its date, context and author.
4. Check whether your product could help without unacceptable implementation work.
5. Choose a permitted, relevant way to ask one small question.
6. Record the reply, next step and what changed your hypothesis.

The bottleneck is a defensible reason to contact someone. If the segment is vague, start with validating an ICP before outreach. Place prospect selection inside a founder-led outbound loop.

AI illustration: a useful conversation starts with an invitation that the other person can choose to accept.

What is the difference between ICP fit, a buying signal and purchase intent?

ICP fit asks whether a company could benefit from your product; a buying signal supplies a reason to investigate now. Purchase intent requires evidence of evaluation or a buying action, while permission to contact remains a separate check.

Working definitions:

Dimension	Evidence to seek	What remains unknown
Fit	Relevant workflow, required integration, suitable operating context	Whether the problem is currently important
Trigger	A relevant hire, launch, expansion or funding announcement	Whether anything needs replacing or purchasing
Problem evidence	A person’s account of a broken process or workaround	Whether they want an outside solution
Evaluation evidence	A request for alternatives, requirements or a demonstration	Authority, constraints and final decision
Contact permission	Applicable channel rules and any actual invitation	Permission for unrelated follow-up

Crunchbase distinguishes unqualified leads from qualified prospects and inbound from seller-initiated outbound. Do not equate qualification with a purchase commitment. In this guide, customer outreach means initiating or continuing contact to understand a need and offer relevant help.

Which public signals are useful, and which are weak proxies?

Prioritize a public request connected to your product over a company characteristic alone. The following hierarchy is a research method, not a validated prediction of conversion.

Signal	Research priority	Check before acting
Explicit request for recommendations or proposals	Investigate first if fit is plausible	Is the request open, relevant and welcoming supplier input?
Named evaluation or replacement project	Investigate the requirements	Can you satisfy the stated constraints and timeline?
Specific complaint or manual workaround	Verify the problem	Does it remain unresolved, and who owns it?
Relevant job advert or new functional leader	Research the planned work	Is the company hiring someone to solve it internally?
Funding, launch, expansion or acquisition	Research the connection	Does the announced plan actually create your use case?
Installed technology or a technology change	Check compatibility	Does the evidence describe current use or replacement intent?
Likes, follows, downloads or category discussion	Keep as context	Is there any stated problem or evaluation?

An account-level signal should not become a claim that a particular employee wants to buy. Do not turn several weak clues into certainty by adding them together.

Where can I find prospects through posts, job listings, reviews and company news?

Find prospects by searching for a specific workflow and problem, then checking the original post, vacancy, review or announcement. Use the evidence to choose contact, verification, watching or exclusion.

Build a search from these fields:

Search site or domain:
Problem phrase:
Intent term, such as recommend, alternative or replacing:
Responsible role:
Workflow:

For a lead-routing product, a completed query is site:linkedin.com/posts "manual" "recommend" "sales operations" "lead routing". Here, manual describes the problem, recommend the possible intent, sales operations the role and lead routing the workflow. These are search examples, not reported findings. If results are too narrow, remove the role or intent term and check it in the original instead.

Google documents quotation marks for exact phrases and the site: operator for a specific domain. Use these Google queries as starting points:

Search location and query	Evidence to save from the original	Decision rule
Posts: site:linkedin.com/posts "lead routing" "recommend"	Author, role, request, date and later replies	Contact only if the request remains open, supplier input is welcome and your capability fits.
Communities: site:reddit.com "lead routing" "alternative"	Full thread, stated constraints and participation rules	Answer publicly where allowed. Do not identify an anonymous author for private outreach.
Job listings: "revenue operations" "lead routing" "manual" jobs	Employer’s live vacancy, responsibilities, tools and integrations	Verify whether the role will solve the workflow internally. Exclude a product mismatch; hiring alone is no buying evidence.
Reviews: site:g2.com "lead routing" "manual"	Review date, exact complaint and any openly identified company	Verify current need independently. Keep anonymous or outdated reviews as market research.
Funding and news: "seed funding" "sales operations"	Original announcement, event date and stated operating plan	Watch if funding is the only signal. Verify further only when the plan connects to your use case.
Company websites: "lead routing" "changelog"	Dated product update, current capabilities and relevant team page	Check whether the update resolves the problem. Exclude a resolved need; investigate an unresolved fit.

Every contact decision also requires a suitable recipient, applicable contact checks and no exclusion. For a post-seed target, verify the funding stage separately; do not treat it as evidence of a broken workflow. If public evidence leaves ownership unclear, ask a colleague, adviser or investor whether a mutually useful introduction is appropriate.

For monitoring, Google Alerts emails new matching Search results. Save problem and account queries; check the event date separately.

Consider Sales Navigator saved searches for new matching people and companies and saved-lead alerts for job changes and updates when tracking roles is the constraint. Neither establishes buying readiness.

For relevant US filers, SEC EDGAR provides public filings with date, company, person and category filters. Check disclosed changes; do not expect coverage of every private startup. Keep manual research where tools lack coverage.

How do I verify that a buying signal is recent, real and relevant?

Verify a buying signal against the original dated statement, the current company situation and the person responsible for the work. Keep the observed fact separate from the problem you infer.

Before writing:

1. Open the original post, vacancy, announcement or filing. A search snippet or AI summary is a pointer.
2. Record publication date, event date if different and your verification date.
3. Check the author, company and current role. Do not guess the identity behind an anonymous account.
4. Look for a later update: completed hire, closed request, selected vendor or changed plan.
5. Explain why this event could affect a workflow your product handles.
6. Mark missing evidence as unknown and choose a review date.

Set freshness by the situation: a closed request has expired; an older problem may need checking. The verified sources establish no universal expiry window. Deduplicate reports of the same event.

How many accounts should I research before sending the first messages?

Start with a research batch of five accounts in one segment around one problem. This is a suggested working limit, not a validated optimum or a requirement to send five messages; send the first reviewed message as soon as one account clears the contact gates.

1. Reserve forty-five minutes for research and fifteen for drafting and recording as an initial capacity limit.
2. Complete an evidence record for each account reviewed and assign contact, verify, watch or exclude with a reason.
3. Stop adding accounts when five have been reviewed or the research time ends, whichever comes first. Leave unresolved checks in verify; do not fill gaps with assumptions.
4. Before starting another batch, handle replies and next steps from this one. Reduce the batch if you cannot do that personally.

If no account qualifies, record whether fit, evidence, ownership or contact eligibility blocked progress. Revise that constraint before expanding the list. Five reviewed accounts can produce no messages; the batch is for learning, not a sending quota.

What should I record for each prospect?

A prospect record should preserve the evidence behind contact and the outcome that can improve the next round. Start in a spreadsheet or CRM with the same fields, keeping unknown information visible.

Account name, domain and responsible role:
Fit evidence; product capability; implementation constraint:
Signal type; primary signal if several apply:
Original source URL and observed wording:
Publication date; event date; last checked:
Known fact; inferred problem; unknown information:
Contact identity, role and appropriate route:
Jurisdiction, recipient type and channel requirements:
Permission evidence, scope and privacy notice where required:
Existing relationship, open opportunity and suppression check:
Decision: contact, verify, watch or exclude; reason:
Message version, contact date and research time:
Reply classification; buyer-confirmed problem and workaround:
Next step, owner and agreed date:
Outcome, loss reason and next review date:

Keep account-level observations separate from each person’s replies. Use an ICP evidence record to retain what buyers confirm rather than overwriting it with enrichment.

How do I prioritize accounts without inventing a misleading lead score?

Prioritize prospects with explicit gates and a stated reason instead of untested point weights. Use four queues so uncertainty determines the next action.

• Contact: fit established, relevant current evidence, suitable recipient and contact requirements met.
• Verify: plausible fit, but the problem, date, ownership or contact basis needs checking.
• Watch: suitable company with no current problem evidence; name the event that would justify reopening it.
• Exclude or route elsewhere: poor fit, unavailable capability, unworkable implementation, resolved problem, objection, relationship conflict, existing customer or active opportunity.

Route customers and active opportunities to their existing owner. Prioritize explicit requests within their deadline, then confirmed evaluations, then problem hypotheses. A trigger cannot override an exclusion. Store the reason for each decision.

Should I comment publicly, ask for an introduction or send an email?

Choose the contact route that fits the person’s invitation, the context and applicable rules. A useful public answer, an introduction and a direct email have different purposes; none should be an automatic escalation after silence.

• Public question: answer the question in place, disclose your commercial connection and ask before sending a resource privately.
• Shared relationship: ask the intermediary whether an introduction is appropriate. Let both people choose.
• Relevant direct contact: after the contact checks, state the verified observation, explain your actual capability and ask about the uncertain problem.
• Weak or sensitive evidence: keep researching or leave the account alone.

An opening question might be: “Is that workflow still under review, or does the opening cover a different need?” Use it only when a checked vacancy actually supports that question. Avoid claiming internal pain you cannot know.

Offer a small resource or discussion relevant to the answer. For the writing step, see personalizing outbound from verified research. Apply outbound designed for mutual fit: a decline ends the relevant contact, and a poor-fit opportunity should be refused.

How do I measure outreach results by signal type?

Measure each signal type by positive replies, qualified conversations and agreed evaluation steps, alongside research effort and negative feedback. Define outcomes before sending so a polite response does not become evidence of demand.

Use these recording rules:

Measure	Definition for this workflow
Positive reply rate	Unique contacted accounts accepting relevant help or a next step divided by unique accounts contacted
Qualified conversation rate	Contacted accounts confirming fit and a current problem in a two-way exchange divided by unique accounts contacted
Evaluation progression	Qualified accounts completing an agreed evaluation step, with an owner and date
Research efficiency	Research and follow-up time per qualified conversation
Failure signals	Bounces, wrong roles, objections, resolved problems and missing implementation fit, recorded separately

Keep counts beside rates and discovery source separate from buying signal. Deduplicate accounts, retain secondary signals and compare equivalent response windows. Hold segment, channel and message stable where possible; label other comparisons directional rather than causal.

Do not prioritize by opens: Apple’s Protect Mail Activity downloads remote content regardless of engagement. Use measuring early GTM tests to set review decisions. Reduce priority when checked rounds produce irrelevant conversations; retain buyers’ explanations.

What privacy and opt-out checks should I make before B2B outreach?

Check the applicable country, channel, recipient type and personal-data requirements before contacting a business prospect. Public availability and relevance do not establish unrestricted permission.

The ICO’s UK B2B guidance distinguishes corporate subscribers from sole traders and certain partnerships. Named business contacts can involve personal data, with lawful-basis, transparency and objection requirements. The page is under review following legislative changes, so recheck it before activation. Do not apply its corporate-email distinction globally.

The EDPB’s consultation-version guidance, paragraphs 110 to 118, explains that legitimate interest is not automatic for direct marketing; specific European and national rules may require consent. Establish both the basis for processing data and permission for the chosen communication.

Before sending, record:

• Which rules apply and whether required permission exists.
• Where the data came from, why it is needed and what notice is required.
• Clear sender identity and an easy way to object.
• A shared suppression check across CRM, inbox and queued follow-ups.

The ICO recommends a suppression record rather than simply deleting an objection, so it can be screened against future lists. Pause unresolved eligibility checks; an unsubscribe cannot repair missing permission.

What does the 3-3-3 rule in sales mean?

Do not use an undefined 3-3-3 rule to qualify purchase intent. Ask which version is intended.

One verified example is the American Gem Society’s sales material, page 61: its after-sale section proposes follow-up three days, three weeks and three months after the first interaction. That is a retail follow-up heuristic, not evidence for selecting B2B prospects. No standard qualification meaning was established by the sources checked here.

How do I find new prospects and reach my target audience without a sales team?

Without a sales team, give the founder or first GTM hire ownership of research, replies and learning. Repeat one problem-focused search and review loop before adding channels or automation.

Keep first messages and substantive replies with a person. Let AI prepare source-linked summaries and drafts; reject invented context and check against originals. Automate alerts, deduplication, review reminders and CRM updates only after the fields and stop rules are clear.

When buyers leave little public evidence, consider introductions, specialist events, search-led content or product-led discovery. Do not infer that quiet companies have no need. Use a permission-based research conversation to learn how they recognize and evaluate the problem.

When should I stop looking for high-intent prospects?

Stop signal-based outreach when you cannot verify relevance, contact appropriately or serve the buyer successfully. A high-intent label should never override a product limitation or an objection.

Pause when nobody can answer replies, conversations contradict the problem hypothesis, or research stops improving decisions. Bring implementation barriers into product planning instead of escalating the pitch.

Refuse the deal that requires capabilities you do not have. Record why it did not fit and what evidence would justify reconsideration.

How would I approach prospect research with you?

I would start with the buyer problem and the evidence needed to justify contact, then build a research and follow-up process you can inspect. I am Piet Baudoin at Poldermarketing, working remotely in Dutch and English with post-seed startups building their go-to-market.

The aim is the first customer, the next, and a repeatable process that reaches people genuinely happy with the product. Alongside positioning and messaging and visibility in Google and AI answers, I build and run AI automations for lead follow-up, enrichment, qualification, CRM workflows, reporting, content distribution and tool integrations. I can set up a CRM and read Search Console queries. Google Ads and Meta Ads are relatively new to me.

A free growth scan is a starting point for identifying where the path from visibility to a useful buyer conversation needs work.