How to get quality leads: define them, find them and check them before you send

This guide is for founders, SDRs and sales or marketing leads at B2B companies who are getting leads, just not enough of the right ones. It shows how to define a quality lead in terms your whole team can check, where quality leads come from, how to grade a list before you spend time on it, and how to tell from the numbers whether quality is improving. The worked example uses real New York City inspection data to show what a trigger-based lead source looks like.

What makes a lead a quality lead

A lead is a quality lead when it passes four tests. Fail any one and the lead costs time without producing a sale.

This mirrors how GTM-Bench, a June 2026 benchmark from Blackpearl Group built from real prospecting requests, grades a lead. A record counts as A-grade only if company fit, contact fit, offer fit, evidence, contactability, identity, claim support and consistency with the source data all score at least 4 out of 5. A B-grade record scores at least 3 on every field, and anything below B counts against the system that produced it. That last rule is the point: a bad lead isn't neutral. It takes a rep's time, it can hurt your sender reputation, and it can sour an account you'll want later.

Quality isn't the same as "qualified". Qualification happens after contact, when you confirm need, timing, authority and budget in conversation; the MQL vs SQL guide covers those stages. Quality is what you control before contact: which leads you choose to work.

How to get quality leads, step by step

1. Define quality from your best customers. Pull your last 10 to 20 won deals and look for what they share: industry, size, location, the role that bought, and what was happening at the company when they first talked to you. Write it in two sentences. It's usually narrower than the market you've been selling to.
2. Write down the disqualifiers. List what makes a lead a waste of time even when it looks right: too few employees, a parent company that buys centrally, a long contract with a competitor, a region you can't serve. Disqualifiers remove more bad leads than any scoring model.
3. Pick the trigger that signals need. Find the event that most often came before your won deals and make it a requirement. A dated, relevant trigger is the closest thing to a quality guarantee you can get before you speak to anyone. Buying signals lists the common types.
4. Source from the trigger, not from a bulk list. Start with the event (public records, job ads, news, an inbound request) and then find the company and the person, rather than starting with a big list and hoping some of it has a need.
5. Hand-check a sample before you send. Check 20 random records against the four tests. If more than three fail, fix the source before you use the rest.
6. Grade every lead and work them in order. Use the scorecard below. Contact A leads first, then B leads; park C leads until something changes.
7. Answer inbound leads fast. Someone who asked to hear from you cools quickly. Agree a response time with your team (the same working day is a sensible default) and track it.
8. Measure by source and feed it back. Every month, record cost per meeting and per customer for each source, and fold what you learn from won and lost deals back into step 1.

Where quality leads come from

Sources differ more in quality than in price. This table reflects common practice rather than measured data, so test it against your own results.

If you'd rather start from contacts that are already checked, Ready-to-Sell Leads come verified, with intent signals and a playbook for each one. Grade a sample of those too: the scorecard works on any source.

The quality lead scorecard

Fill this in for each lead, or for a 20-record sample of a new source. A 5 means you confirmed it at the source yourself; a 3 means it's likely but unconfirmed.

Lead: [name, role] at [company] ([website])
Source: [where the lead came from]    Date found: [date]

Score each field 1 to 5 (5 = confirmed at the source, 3 = likely, 1 = wrong or unknown)

COMPANY FIT: [ ] industry, size, location and model match our profile
PERSON FIT:  [ ] owns the problem or the budget, and is still in the role
REASON NOW:  [ ] dated trigger in the last 90 days: [trigger, date, source]
CONTACT:     [ ] email or phone verified; email domain matches the company
EVIDENCE:    [ ] every fact above traced to a source I opened

DISQUALIFIERS (any one means drop the lead): [your list]

GRADE
A: every field 4 or 5. Contact this week.
B: every field 3 or more. Contact after A leads, or firm up the weak field first.
C: any field below 3. Don't contact. Fix the field or drop the lead.

To rank a whole list rather than one lead at a time, the lead scoring calculator weights fit and intent and sorts the list for you.

Worked example: a pest control company compares two lead sources

Illustrative example. The company, its prices and its conversion rates are fictional. The inspection counts are real New York City data, pulled on 26 October 2026.

Brightline Pest Services (brightlinepest.example) sells monthly pest control contracts to restaurants in Queens. A contract is worth about $3,000 a year at a 50% margin, so $1,500 in first-year gross profit. Staff time costs Brightline about $35 an hour.

The trigger. New York City's health department inspects restaurants unannounced at least once a year and publishes every violation on the city's open data portal. Three codes matter to a pest control firm: 04K (evidence of rats), 04L (evidence of mice) and 04M (live roaches). From July to September 2026, 396 Queens restaurants were cited under at least one of them, out of 1,769 Queens restaurants inspected in that period. (The dataset covers restaurants that are still open and includes citations not yet adjudicated.) That's about 130 restaurants a month with a dated, documented problem and every reason to fix it before the next unannounced visit.

Two sources for one month:

What the numbers show. The two sources cost about the same in total, but the triggered list won four times as many contracts, so each one cost about a fifth as much. Both clear Brightline's ceiling of $1,500 per contract, so on paper both "work". The difference is that the bought list caps growth at one contract per 17 hours of work, and it carries a risk the table doesn't show: 7 failures in 20 suggests about a third of the list is bad, and every email to a dead address or the wrong person adds to bounces and spam reports that drag down the emails that follow.

The note Brightline sends names the inspection date, offers a site survey that week and mentions one similar restaurant it looks after nearby. It doesn't repeat the violation back in detail: referring to a public record is fine, lecturing an owner about it isn't.

Why more volume makes quality worse

When results drop, the reflex is to send more. Two pieces of evidence argue against it.

Deliverability. Google's sender guidelines ask anyone emailing Gmail accounts to keep the spam rate shown in its Postmaster Tools below 0.1% and never to reach 0.3%. At 1,000 emails, 0.3% is three people clicking "report spam". A loose list of people with no reason to hear from you gets there quickly, and then your good leads stop seeing your email too.

Overproduction. In GTM-Bench, four of the AI agents tested finished with a negative net score, meaning they returned more below-standard rows than good ones, and the paper names overproduction as a repeated cause: one agent returned 6,342 rows for a single task. Even the best-scoring system met the A-grade bar on about 41% of the rows it returned. However a list was built, it needs a sample check before you use it.

How to tell whether lead quality is improving

Track each figure by source and by grade. If B leads close nearly as well as A leads, your bar may be too strict. If the C leads you contacted anyway never close, the scorecard is doing its job.

Self-check

• You can describe a quality lead in two sentences a new hire could check.
• You have a written list of disqualifiers.
• Every lead you contact has a dated reason and a source for it.
• You hand-check 20 records from every new source before sending.
• You know your cost per customer for each source, including staff time.
• Won and lost deals update your definition at least once a quarter.

Pitfalls

• Treating every form fill as a quality lead. Inbound shows interest, not fit. Apply the same four tests.
• Scoring without disqualifiers. A large company locked into a three-year contract can score well on fit and still waste a month.
• Counting cheap records as cheap. Price per record ignores the hours spent on the ones that go nowhere. Put time in the sum.
• Stale triggers. Last year's expansion is history. Use triggers from the last 90 days where you can.
• Definitions that drift. If sales and marketing judge quality differently, write the definition down and review it together every quarter.

Next step: turn your definition into a working list with how to build a prospect list.