How to Find Qualified Sales Leads: Sources, Checks, and a Worked Example

To find qualified sales leads, define which companies your offer can help, search sources that reveal those companies and relevant buyer roles, verify each person and contact route, then test the need in a conversation. A company that fits your ideal customer profile is a prospect worth investigating. A booked first call is a discovery step; call the lead sales-qualified only after it meets your team's written buyer-evidence and handoff criteria.

In short: Start with one offer and a small account segment. Write down fit rules and exclusions. Find companies through customer referrals, first-party inquiries, public company information and professional directories. Record the source and date for each account and person, check the role and contact route, and make one of four decisions: contact, research, hold or stop. Count an SQL only after buyer evidence satisfies your written acceptance rule.

What makes a sales lead qualified?

"Qualified" is often used for two different judgments. Fit qualification asks whether this company and person plausibly match your offer. Sales qualification asks whether a buyer has shown a relevant problem, interest or agreed action that merits deeper sales work. Your team should define both stages, because the evidence available before and after contact differs.

For example, a 150-person US services firm could fit a workflow software vendor's target market. A new operations job posting might give the seller a reason to investigate. Neither fact proves that the firm has a broken process, budget, a current project or interest in that vendor. If its operations leader later describes a handoff problem and agrees to a scoping call, the record supports a different decision. Salesforce's SQL guidance connects sales qualification to fit, a relevant need and readiness to progress; it does not make a database match a buyer conversation.

Use distinct CRM states:

• Candidate: Found, but not checked.
• Fit-approved prospect: Meets your team's documented account and role criteria.
• Engaged lead: Responded or took a meaningful first-party action.
• Sales-accepted for discovery: Described a relevant issue and agreed to a specific exploratory next step.

A sales-qualified lead (SQL) needs the stronger buyer evidence in your team's written SQL rule. In the example below, discovery must confirm an active need, relevant stakeholders, a workable path through constraints and budget, a decision path and an accepted next action. A booked first call alone is not an SQL. These labels are illustrative; agree on your actual handoff rule before reporting SQLs.

Set a fit rule before searching

Start with a specific offer and use case. Write an ideal customer profile from customers you can serve and recent wins you understand, then let sales and delivery teams challenge it. Salesforce's lead qualification model guide begins with a shared definition of the ideal or qualified prospect. That is more useful than one rep's private idea of a "good lead."

For a first segment, document:

• Company fit: geography, industry or business model, operational size and the condition that makes your offer useful. Give ranges only when your own service capacity or evidence supports them.
• Exclusions: circumstances that make the offer unsuitable, such as a required integration you cannot provide or a customer type you do not serve.
• Buyer jobs: who may feel the problem, own the workflow, approve a change and evaluate risk. A job title is a search clue, not proof of authority.
• What would justify contact: a first-party inquiry, referral, specific public event or other attributable reason to ask a relevant question. A generic "intent score" is not an observed buyer statement.

An example fit rule for a fictional workflow vendor could read: "US B2B service firms with separate, multi-person sales and delivery teams that our tool can serve; exclude sole operators and firms outside the service model we support." This is an illustrative test, not a recommended industry or employee cutoff. A public team and services page may support the preliminary fit check. Whether handoffs are difficult, an integration is feasible or a project exists must be learned from the buyer.

Keep fit separate from timing. A strong-fit company without a known current project may enter a research or nurture queue. A company outside your service scope does not become a good target because one employee downloaded an e-book. To get started, compare five recent customers you served well, write the shared fit conditions and one exclusion, then collect ten candidate accounts from two distinct sources. Save the source and check date for each. Only after that sample should you expand a database search.

Find candidate accounts in sources you can check

The best source depends on what you sell and whether you need account fit, a current buyer question or a route to a named person. Try two sources first and compare the records they produce.

1. Your own conversations and referrals. Review recent customers, lost deals, support questions and partner introductions for a repeated use case. Ask the referring person whether an introduction is appropriate; do not treat a named contact as having opted into a pitch. Record the referring relationship and what was actually said.
2. First-party inquiries. Demo requests, event questions and replies reveal more about the person's stated interest than an anonymous page view. Verify whether the inquiry concerns your offer, whether the person fits the buyer role and what follow-up was requested. A download alone is a weak readiness signal.
3. Public company and role information. Company sites, job posts, annual reports and professional profiles can identify organizations, initiatives and plausible buyer jobs. Save the URL and date. A funding round or new hire may make a question timely; it does not prove the company has the problem you sell against.
4. Directories and data providers. Filter for your fit criteria and treat results as candidates. Check the company, role, current employment and address or contact route against a reliable source before use. A provider's "verified" label does not independently establish an individual's current job or permission to receive a message.

For a specific LinkedIn workflow, use the LinkedIn prospecting guide. LinkedIn's account-page documentation describes company summaries, saved accounts, people and updates; these help research an account, while a conversation must establish private needs. A broader account map helps when several stakeholders may influence the purchase.

Find the right person and verify a contact route

Once an account passes the fit gate, look for the job that likely owns the workflow or the consequence of the problem. Record the person's current company, role and source date. If ownership is unclear, add a question to resolve it; do not silently promote a likely title to "decision-maker."

Then check your route. A warm introduction may be best when a relevant mutual contact agrees. A public business contact address, professional network message or first-party reply may be appropriate in other cases. Verify the address and check the organization's outreach policy, opt-outs and applicable law before sending. For US commercial email, the FTC says CAN-SPAM also covers B2B messages. Federal CAN-SPAM does not impose a blanket prior opt-in requirement, but a public address is not permission to ignore opt-outs or applicable channel and state rules. Our unsubscribe guide explains the practical US routes.

Keep the record lean: account, fit reason, exclusion check, named person and role, source link/date, contact route, observed signal, unresolved question, owner and next action. A contact that cannot be verified goes back to research; it should not be added to a sending list by guesswork. Use a lead tracker or CRM to preserve the evidence and owner.

Route each lead by evidence, not a guessed score

Use four decisions for every candidate:

Contact. The account passes the fit rule, the person and route are verified, and you can state a relevant reason to ask a question without inventing a pain point. The first message's job is to learn, not to declare the account sales-ready.

Research. The fit or buyer job looks plausible but a required fact, source or contact route is missing. Assign one owner and a dated question, such as "Does this firm actually sell through a multi-person team?" Avoid an indefinite "needs more research" status.

Hold. Fit is confirmed, but the timing or permission for contact is unclear, the buyer asked to revisit later, or the team lacks capacity to follow up. Record the reason and a review date. A hold is not a qualified opportunity.

Stop. The account fails an exclusion, the person is irrelevant or has opted out, or the evidence establishes that this offer does not fit. Record the reason. An unanswered email alone does not prove the account is a bad fit; it can justify an outreach stop under your policy.

If you later use a scoring model, keep the underlying evidence visible. Separate account fit from observed engagement and buyer-stated need. A high total generated by job title, funding and page views can still have no buyer conversation behind it. A score may prioritize review; the sales qualification decision should use your written acceptance rule.

A worked example: from company signal to buyer conversation

The following people and companies are fictional. The vendor sells a workflow tool for US service companies with multi-person sales-to-delivery handoffs. Its pilot rule accepts a lead for a sales scoping call when a relevant buyer describes a current handoff problem and agrees to explore a specific next step. It does not require a confirmed budget before that call. Its separate SQL rule requires discovery to confirm an active relevant issue, the people who evaluate and approve a purchase, a preliminary feasibility check on buyer-described technical and data requirements with no known blocker, a budget and decision path with no known blocker, and a buyer-agreed next evaluation step. The IT lead's detailed validation may still be pending. These are this fictional team's operating rules, not a universal SQL definition.

1. Candidate found. The seller searches a regional business directory for US B2B service firms, then opens each candidate's own site rather than trusting the listing alone. Northstar Services' team pages show a sales team of four and a separate delivery team of six. Its services page shows the supported B2B project model; it is neither a sole operator nor outside the vendor's stated service scope. A current job post seeks a revenue operations manager. The seller saves the page names and October 22 check date. The public pages support fit under the written rule; the job post does not say handoffs are broken. Status: Research until person and route are checked.

2. Person and route checked. The company site names Maya Chen as VP of Sales, and a public professional profile confirms her current role. A trusted partner knows Maya but has not agreed to introduce the seller. The seller asks the partner if an introduction would be welcome and records the answer. Status stays Research until a permitted route is clear.

3. Contact with a bounded question. The partner agrees to make an introduction. The seller asks Maya whether consistent sales-to-delivery handoffs are a current priority as the teams grow. This is a hypothesis, not a claim that a job posting proves a broken process. Status: Contacted fit-approved prospect.

4. Buyer evidence changes the state. Maya replies: "We lose the implementation notes when deals pass to delivery, and I want to fix that before our next onboarding wave. Could you show how your team would map our handoff?" The seller copies her exact words and date, and Maya accepts a scoping call.

That is enough for sales-accepted for discovery under the fictional pilot rule: Maya described a relevant current issue and agreed to a defined next step. It is not yet an SQL. Budget, purchasing authority, decision process and technical fit remain open. The sales owner assigns those questions to discovery.

5. Discovery tests the SQL rule. On the fictional October 27 call, Maya confirms that lost implementation notes affect the next onboarding wave. She owns the sales handoff; the delivery director will evaluate the workflow; the IT lead will check the integration and data handling. Maya says the COO approves this kind of purchase after those checks, and a budget range has been reserved for an approved handoff project. She and the delivery director agree to review a handoff map on October 30. From Maya's description, the seller makes a preliminary scope check: the stated integration and data requirements appear to fit the vendor's supported range, with no known blocker. IT validation has not happened yet. These are fictional buyer statements and a seller assessment, not real-company evidence.

The five conditions in this team's SQL rule are now met: active issue, known evaluator and approver, a preliminary scope check with no known blocker, budget and decision path without a known blocker, and an accepted next evaluation step. The sales owner marks the lead SQL on October 27 and records the October 30 map review as the next action. IT validation, approval, contract and purchase remain unconfirmed; an SQL is not a won deal. If Maya had declined the map review, record Hold with a follow-up date when appropriate. If the required integration were impossible, record Stop for this offer.

Here is what a compact record of that decision would contain. All names, source labels and dates in this example are fictional:

• Account and fit: Northstar Services; US B2B service firm; company pages show four sales and six delivery staff and a supported services model. The stated fit exclusions do not apply. Maya describes the technical and data requirements in discovery; the seller's preliminary check finds no known blocker, while IT validation remains pending.
• Sources checked: regional business directory listing, Northstar team and services pages, and operations job post, checked October 22, 2026. The job post suggests a timely question but does not prove a problem.
• Person and route: Maya Chen, VP of Sales, role checked against the company page and professional profile on October 22; partner agreed to introduce the seller. No cold-email address was inferred.
• Buyer evidence: Maya's quoted reply, received October 23; she requested a handoff mapping discussion and accepted a call.
• Decisions, owner and next action: A. Seller accepts for discovery on October 23 and holds the October 27 call. Maya's answers confirm the active issue, evaluator and COO approval path, described integration requirements, reserved budget range and October 30 map review. The seller's preliminary scope check finds no known blocker. A. Seller marks SQL on October 27; next action is the map review on October 30. IT validation, approval and purchase remain unconfirmed.

Now compare a second fictional account: Harbor Retail downloaded a general checklist, but sells directly to consumers and has no sales-to-delivery handoff. It fails the offer's fit rule despite the download. Mark Stop — outside ICP, and do not count the download as a qualified lead. This contrast protects the meaning of both stages.

The example is a decision record, not a claim that these companies exist or that this sequence produces a predictable reply rate. In your own process, change the criteria and handoff threshold to match what sales can actually verify and accept.

Measure which sources produce usable conversations

Report each step with its own denominator: candidate accounts found, fit-approved accounts, accounts with one verified buyer and route, contacted people, meaningful buyer replies, people accepted for discovery and SQLs after discovery. Keep accounts and people in separate counts. Record source and dates so duplicate people or accounts do not inflate the apparent yield.

Separately from the Northstar story, suppose a different small fictional pilot starts with two sources and assigns each account to its first recorded source. The team researches one buyer per account and contacts no more than that one person during this pilot:

• Partner introductions and customer referrals: 20 candidate accounts → 15 fit-approved → 12 with a verified buyer and route → 8 people contacted → 3 substantive replies → 2 accepted for discovery.
• A public business directory: 20 candidate accounts → 10 fit-approved → 6 with a verified buyer and route → 4 people contacted → 1 substantive reply → 0 accepted for discovery.
• Combined: 40 candidates → 25 fit-approved → 18 with a verified buyer and route → 12 people contacted → 4 substantive replies → 2 accepted for discovery. No SQLs are counted until discovery meets the separate SQL rule.

These are illustrative counts, not benchmarks or a conversion forecast. In this small sample, referrals produce 15 fit-approved accounts from 20 candidates, while the directory produces 10 from 20. Of people contacted, 2 of 8 in the referral cohort and 0 of 4 in the directory cohort were accepted for discovery.

These tiny groups cannot establish that one source will outperform the other. They point to the records worth inspecting. Why did directory candidates fail fit? Could its filter be improved?

Do not divide two discovery acceptances by 40 and claim the outreach copy failed. Locate the loss first: source fit, role verification, contact route, response or actual buyer need.

Review rejected and held records weekly. Let the observed reasons change your ICP criteria, source mix or research method. Do not loosen the sales-accepted definition merely to show more leads. If the accepted leads progress to opportunities, assess whether that later outcome supports the original fit rule; preserve the lag and the small sample.

Questions sellers ask about qualified leads

Is an ideal-customer-profile match already a qualified sales lead?

It can be a fit-approved prospect under your rule. It is not automatically sales-qualified. You still need the buyer evidence and handoff conditions your team defined. A company database and a person profile are useful for finding where to investigate, not for proving a current need.

What if the account fits but nobody responds?

Keep fit and response as separate facts. Follow your outreach policy and any requested opt-out. If the person has not responded after the appropriate sequence, stop that route or place the account on a dated hold; do not declare a buyer need or mark it SQL. A new attributable event may justify fresh research, but silence alone proves neither a bad fit nor readiness.

How many qualified leads should we aim to find each week?

There is no safe universal target. Start with the number your team can research, contact appropriately and follow up. Track fit approval, meaningful replies and sales acceptance by source. Set a capacity-based goal after observing your own funnel, not from an external percentage presented without your context.

Take ten current names and record, for each, the fit rule and exclusion check, source and check date, buyer role and verified route, observed buyer evidence, owner and next action. Assign Contact, Research, Hold or Stop using the four rules above. The missing field will show whether the next hour belongs to sourcing, verification or a buyer conversation.