MQL Definition: How to Agree on Lead Quality With a Lead Scoring Model

A marketing qualified lead (MQL) is a lead that meets criteria the marketing team and the business agreed on in advance: enough fit and enough intent that it is worth the sales team's time. Not every form fill is an MQL. Someone outside your service area, a fake name, or a person "just looking for ideas" is a lead, but not a qualified one.

The written version of those criteria is often called an MQL definition or a lead quality agreement. It is one of the most useful documents an agency and a client can sign, because it replaces the argument "your leads are bad" with a simple check: did the lead meet the definition or not?

This guide explains the lead stages, how to build a lead scoring model with points and hard disqualifiers, two worked examples, and what to put in the agreement.

Quick summary

• Raw lead: anyone who submitted their details. Not checked yet.
• IQL (information qualified lead): a valid lead that answered your qualifying questions.
• MQL: a lead that scores above an agreed threshold and hits no hard disqualifier.
• SQL: a lead the sales team has spoken to and confirmed as a real opportunity.
• Score each lead out of 100 on a few weighted criteria, such as budget, authority, need and timeline, and set a threshold (for example 60).
• Add hard disqualifiers that reject a lead no matter how high it scores.
• Agree on a response time for sales, a rejection window for disputed leads, and a monthly review.

Why agencies and clients need an MQL definition

An agency running ads for a client can produce hundreds or thousands of leads a month. Without a shared definition, three problems keep coming back:

1. Nobody agrees on what a good lead is. The client judges by gut feeling, the agency by cost per lead, and both are frustrated.
2. The agency gets judged on sales it doesn't control. Whether a lead buys also depends on the price, the sales calls, the follow-up speed and the product. If everything is measured only by revenue, the agency stops focusing on what it can improve: the campaigns.
3. Cheap leads look like success. Open up the targeting and you can get far more leads at a lower cost. But if they don't qualify, more leads only waste the sales team's time.

With an agreed MQL definition, the agency's job is clear: deliver as many MQLs as possible at an acceptable cost per MQL. The client's job is just as clear: follow up fast and turn MQLs into sales. Both can be measured.

The lead stages, from raw lead to customer

Nobody should report on raw leads as a result. You can always get more raw leads. The useful numbers start at IQL and MQL.

The universal qualification dimensions

Almost every scoring model is built from the same few questions. The best-known checklist is BANT: budget, authority, need and timeline. Some teams prefer CHAMP, which starts with the challenge rather than the budget, so strong prospects aren't rejected just because they haven't set a budget yet.

Pick only the dimensions that matter for your business, and weight them by importance. Four to six criteria is usually enough.

How to build a lead scoring model

1. List your criteria. Use the dimensions above, written as specific, checkable conditions. "Budget over $15,000" is checkable; "good budget" is not.
2. Give each one points out of 100. The most important criteria get the most points (for example 25), the weaker signals fewer (for example 10). Partial points are fine: 20 for a timeline under three months, 10 for three to six months.
3. Set the MQL threshold. A lead at or above it is an MQL; below it, it stays an IQL. Start with a sensible guess, such as 60, and adjust it after a month or two based on which leads actually became customers.
4. Write the hard disqualifiers. Conditions that reject a lead whatever its score.
5. Add the questions to your forms, so every lead arrives with the data you need to score it.

Example 1: kitchen renovation company

A company that designs and fits complete kitchens in one metro area. Typical projects cost $15,000 to $40,000.

MQL threshold: 60 points.

Hard disqualifiers: address outside the service area; renting the home without the owner's approval; looking only for a single cabinet or worktop repair; fake or test details.

Scoring two leads:

• Lead A: budget $18,000 (25), homeowner (20), starting in 4 months (10), full kitchen (15), replied the same day (10), valid details (10). Score: 90 → MQL.
• Lead B: no budget given (0), homeowner (20), "maybe next year" (0), full kitchen (15), didn't answer (0), valid details (10). Score: 45 → IQL. Not a bad person, but not ready. Put them into a nurture sequence instead of the sales queue.

Example 2: IT support company for small businesses

A managed IT services company that looks after offices of 15 to 200 people.

MQL threshold: 55 points, since B2B leads often don't share every detail on the first form.

Hard disqualifiers: fewer than 5 employees; outside the regions served; job seekers, students or vendors selling to the company; an existing customer asking for support (send it to the support desk, don't count it as a lead).

The same structure works for any business: a language school, a car dealer, a solar installer, a wedding venue. Only the criteria and the weights change.

Hard disqualifiers: the lines a lead can't cross

A lead can score well and still be useless. Typical hard disqualifiers:

• Outside the target area agreed for the campaign. If ads target one city, a lead from another country does not count, however good it looks.
• Fake or test entries: "asdf", "test", 0000000000, a competitor checking your form.
• Duplicates of a lead already received in the last 30 days.
• Wrong intent: job applications, sales pitches, students doing research, people who say they are only comparing prices with no plan to buy.
• Something you don't offer: a service you don't provide, a product you don't sell.
• Legal or policy limits: for example under-18s where you can't sell to them.

Keep this list short and specific. If it grows long, the targeting or the form probably needs fixing.

Design your forms to collect the scoring data

You can't score what you don't ask. For each criterion, add one simple question, ideally multiple choice so answers are easy to score:

• Use ranges for budget, not an open box.
• Ask the timeline as fixed options: within 1 month, 1–3 months, 3–6 months, just researching.
• Mark the essential questions as required and the rest as optional.
• On Meta lead forms, the higher intent form type adds a review step before submission, which cuts accidental submissions.

Every extra question lowers the number of leads but raises their quality. That trade-off is the whole point: agree with the client whether they want more leads or better leads, and test it.

The response time agreement (lead SLA)

An MQL that nobody calls for a week is no longer an MQL. Speed matters a lot: a widely cited 2011 Harvard Business Review study found that companies that tried to contact a web lead within an hour were nearly seven times as likely to qualify it as those that waited even an hour longer.

Write the client's side into the agreement too:

• First contact within a set time during business hours, for example 15 minutes for phone leads.
• A minimum number of attempts, for example three calls and one message over two days.
• Outcome recorded in the CRM or the shared sheet within 24 hours: reached, not reached, qualified, not qualified, with a reason.
• If the response time was missed, the lead can't be marked as bad quality later.

Preventing disputes

Most "the leads are bad" arguments happen because there was no definition to check against. With one, disputes become a routine check:

• Rejection window: the client can reject a lead within, say, 5 business days, with the reason from the disqualifier list or the scoring. After that, it counts.
• Monthly review: go through a sample of rejected leads and MQLs that didn't buy, together. Look for patterns, not blame.
• Fix the root cause together. Common complaints and where to look:

Neither side should simply impose its view. The definition is a shared tool, and it should be updated when the data shows it is wrong.

What to put in a lead quality agreement

A simple agreement, one or two pages, covers:

1. Parties and scope: business, agency, industry, campaigns and channels covered.
2. Lead stage definitions: raw lead, IQL, MQL, SQL, as above.
3. Scoring model: criteria, points, the MQL threshold.
4. Hard disqualifiers.
5. Required form questions for each campaign.
6. Lead response SLA for the sales team.
7. Dispute process: rejection window, evidence needed, monthly review.
8. Reporting: what is reported, how often and where.
9. Review date: when the definition is checked and updated, for example every quarter.
10. Sign-off from both sides.

Agencies that charge per MQL, which is common in lead generation, need this document even more: it decides what they get paid for.

What to report every month

• Raw leads, IQLs and MQLs, by campaign.
• MQL rate: MQLs ÷ raw leads.
• Cost per MQL: ad spend ÷ MQLs. This matters more than cost per lead.
• MQL to SQL rate and SQL to sale rate, from the client's side.
• Rejected leads, with reasons.

Cost per MQL is the same calculation as cost per lead, just with MQLs as the count.

If MQLs are high but sales are low, the problem is after the handover. If raw leads are high but MQLs are low, look at the targeting, the ad message and the form. For a full view of each step, see types of marketing funnels.

Do you need all this for a small campaign?

Not always. If you spend a small daily budget and get a handful of leads a week, a full scoring model is more work than it's worth. Write down three or four must-haves, keep a simple spreadsheet, and call every lead.

A formal MQL definition pays off when:

• many leads arrive every day and several people handle them;
• an agency and a client need to agree on results;
• the agency is paid per qualified lead;
• you want to feed qualified leads back into Google Ads or Meta as conversions, so the algorithm finds more of them instead of more cheap form fills.

Most CRMs, including HubSpot, Zoho CRM and Salesforce, can score leads automatically with rules like the ones above, so you don't have to add up points by hand once volume grows.

To see what an MQL is worth to you, start from your sale value and close rate, as in how to calculate target CPA and CPL.

Frequently asked questions

What is an MQL in simple words?

A lead that meets criteria the business and the marketing team agreed on in advance, showing enough fit and intent to be passed to sales.

What is the difference between an MQL and an SQL?

Marketing decides an MQL from the lead's data and score. An SQL is a lead the sales team has spoken to and confirmed as a real opportunity with need, budget and intent.

What is an IQL?

An information qualified lead: a valid lead that answered your qualifying questions but hasn't reached the MQL score.

What is a good MQL score threshold?

There's no universal number. Many teams start around 55 to 65 out of 100 and adjust it after checking which scored leads actually became customers.

What is a lead quality agreement?

A short document an agency and a client sign that defines lead stages, the scoring model, hard disqualifiers, the follow-up time for sales and how disputed leads are handled.

How fast should sales respond to a lead?

As fast as possible, ideally within minutes during business hours. Research has long shown that leads contacted within the first hour are far more likely to qualify than leads contacted later.

Should an agency be judged on sales or on MQLs?

Mainly on MQLs and cost per MQL, since sales also depend on the client's pricing, sales team and follow-up. Both sides should still review sales together to keep improving the definition.