All articles
Guide13 min read

MQL and SQL: definitions, criteria and how web agencies use them

MQL or SQL? Definitions, criteria and the sales handoff, adapted to a web agency doing outbound prospecting of local small businesses (TPE/PME).

By Nicolas Faglin, Co-founder and co-managing director
Isometric illustration of the lead qualification process

In short

  • MQL: a contact who matches your target (fit) and shows interest (intent), qualified by marketing, often through scoring.
  • SQL: a lead validated by a salesperson after a conversation, with a real need, a relevant contact and a credible timeline.
  • Lead or prospect? A lead has shown interest, a prospect is a lead that has been assessed and qualified.
  • In an agency: fit is checked before any contact, intent is replaced by a visible flaw (slow site, weak reviews), and the SQL is born at the first meeting.
  • What matters: written criteria and a handoff that triggers a specific action in the CRM.

An MQL (Marketing Qualified Lead) is a contact whose profile matches your target and who has shown some interest, but no buying signal. An SQL (Sales Qualified Lead) is an MQL that a salesperson has contacted and whose need was confirmed in a real conversation. For a web agency prospecting local businesses, the right reading is simpler than in the textbooks: you are not waiting for prospects to come to you, so your first filter is a need you have spotted yourself on their website, and your second is the conversation.

MQL and SQL: the two-minute definition

The vocabulary comes from B2B marketing, where a website, content and campaigns attract contacts that need sorting before they are handed to sales. Two stages structure that sorting.

According to Webmarketing School, an MQL is a prospect that closely matches the ICP, the ideal customer profile (company size, industry, seniority, location). It is assessed manually by marketing or automatically through a score. An SQL is an MQL that the sales team (SDRs or BDRs, the reps in charge of prospecting) has contacted, and whose interest and potential were validated in a first qualifying conversation. The opportunity then moves forward in the CRM pipeline.

The useful nuance: an MQL is qualified by data (profile and behavior), an SQL by a conversation. Rework sums up the difference this way: the MQL goes through scoring, while the SQL is validated by a salesperson after a human conversation that confirms a real project.

Lead, prospect, MQL, SQL: where each term fits

The four terms describe levels of maturity along the same journey.

StageWhat you knowWho qualifies
SuspectA visitor or a company you have spotted, with no signalNobody yet
LeadA sign of interest, with no proof of a concrete needMarketing
MQLFit with the ICP and enough interestMarketing, often through scoring
SQLNeed, decision-maker and timing confirmed in conversationSales
CustomerPurchase madeSales

The split between suspect, lead, prospect and customer is described by Axonaut. For La Poste, a lead shows a first sign of interest, and a prospect is a qualified lead, identified as matching the target with concrete commercial potential. In practice, "prospect" therefore covers the zone between MQL and SQL, and many teams use the word without distinguishing the two.

You do not need to settle this vocabulary debate. What matters is that everyone in your agency, including the partner freelancer or the business developer, means the same thing by "SQL".

The criteria that make an MQL, then an SQL

The software vendor Hyperops offers a clear grid. An MQL is judged on two dimensions.

  • Fit: the company matches the ICP, its size is right, it is active in your target markets, the contact's role is relevant and their potential need matches your offer.
  • Intent: downloading a resource, signing up for a webinar, viewing high-intent pages several times, replying to an email, asking for a meeting.

An SQL, on the other hand, is validated by sales when a real problem has been identified, your offer can solve it, there is a priority and a timeframe, a relevant contact is involved and a next step is on the table.

For complex enterprise sales, Rework cites the MEDDIC method (metrics, economic buyer, decision criteria and process, identified pain, internal champion). For selling a website to a plumber or a restaurant, that is overkill: a short version is enough, the one found in the grid above.

Two counterexamples that weed out false MQLs

Rerow illustrates the double-criterion requirement well. A student who downloads a white paper shows engagement but has no fit. The sales director of a target mid-sized company who opened an email six months ago has fit but no engagement. Neither is an MQL.

Applied to your day-to-day: a tradesperson with a nice, fast, well-ranked website has geographic and industry fit, but no visible need. An owner whose site is dated and takes several seconds to load ticks both boxes.

Why the classic model is a poor fit for an agency doing outbound

The MQL-then-SQL model assumes inbound flow: a website that gets traffic, downloaded content, a score that climbs. But a web agency targeting local small businesses (TPE/PME) rarely works that way. The businesses you want to sign do not search for "website redesign" until the problem becomes glaring, and they do not subscribe to any agency newsletter.

Three adjustments are needed.

  1. Fit is checked before any contact. Industry, city, size, an identifiable owner: all of this is known before the first message.
  2. Intent is replaced by an observable need. With no behavior to score, you score flaws: an outdated or slow site, weak SEO, few or poor Google reviews, abandoned social media.
  3. The SQL is earned in conversation. Until the owner has confirmed a project, a rough budget and a timeframe, you are prospecting, not selling.

So you can keep both stages, mentally renaming the first one: a "lead qualified on fit and visible need", then an SQL after a meeting.

A qualification grid for agencies, ready to copy

Adapt it to your offer, for example with these thresholds.

  • Qualified lead (MQL equivalent): a local business in your target industries and cities, an identified owner, at least one concrete flaw found on the website or the Google Business Profile, usable contact details.
  • Engaged lead: the owner has replied, clicked or asked for more information.
  • SQL: a conversation has taken place, the owner acknowledges the problem, decides alone or with an identified partner, and a meeting or a quote is scheduled.
  • Back to nurturing: the need exists but the timing is wrong, so the lead goes back into spaced-out follow-ups rather than into the bin.

The handoff: written criteria and response time

According to Hyperops, the real difficulty is not finding the perfect definition but making sure everyone, CRM included, uses the same one, with criteria precise enough to know what to do when a lead changes status. Moving to SQL should trigger concrete actions: assigning a salesperson, a follow-up task, creating an opportunity, tracking response time.

The formal framework is called an SLA (Service Level Agreement): an agreement that sets the handoff criteria and the maximum response time, as described by Drift Digital. Rerow notes that this handoff is the most common point of friction between marketing and sales, and that it is solved with written criteria, scoring in the CRM and a commitment on response time.

In a small agency, marketing and sales are often one and the same person: the founder. The SLA then becomes a commitment to yourself: for example, every positive reply gets a call on the next business day. This is exactly the kind of rule that vanishes once the schedule fills up with client projects.

The loop back to nurturing

Not every rejected lead is lost. Rework recommends a feedback loop: sales send unqualified leads back to marketing nurturing so that long-term opportunities are not wasted. An owner who says "not now" today may become an SQL in six months, when their website starts causing problems. To keep that connection alive without chasing them relentlessly, your editorial presence matters: the approach to nurturing prospects on LinkedIn shows how to stay visible to these contacts between conversations.

The hidden cost: qualifying by hand

Qualifying a lead by hand is slow. Tensoria puts manual qualification at 20 to 45 minutes per lead, versus under 3 minutes with an AI agent, for inbound leads. For an agency, the logic is the same in outbound: open each business's website, test its speed, note its reviews, look for the owner, find their email, then write a message that is not generic. At 30 minutes per business, 40 leads add up to 20 hours, a freelancer's working week, before a single message has gone out.

Comparison of the time needed to qualify leads

This is where tooling makes the difference between steady prospecting and prospecting that stops as soon as a big project lands.

How AutoLeads applies this model

AutoLeads is Grezy's prospecting tool for French web agencies. Here is where it steps in at each level of the grid.

Grid stageWhat AutoLeads does
FitYou choose an industry and a city, for example "plumbers in Lyon" or "restaurants in Bordeaux". AutoLeads finds these businesses in real time on the web and Google Maps.
Visible needIt audits their online presence: outdated or slow site, weak SEO, Google reviews, social media.
ContactIt enriches each lead with the owner, a verified email and a phone number.
First messageIt writes an email and LinkedIn message sequence built on the actual flaws of each business's website.
SendingMessages go out in weekly waves from the agency's own mailboxes, or are pushed to Lemlist or HubSpot.
Follow-upReplies land in a built-in CRM as tasks.

The benefit for your pipeline is direct: you go into the first conversation with a concrete diagnosis rather than a generic message, which makes validating the need, and therefore moving to SQL, more natural. And since replies create tasks, your response time becomes a reflex rather than a good intention.

Measuring what matters

The key metrics remain the MQL-to-SQL and SQL-to-customer conversion rates, as Rerow points out. For an agency, they translate into four numbers to track every month.

  • The number of qualified leads contacted.
  • The reply rate, positive and negative replies combined.
  • The number of meetings held, in other words SQLs.
  • The number of signed quotes.

If many leads reply but few become SQLs, your fit criterion is too broad: narrow down the industry or the size. If few reply, the message or the flaw you chose to highlight is the problem. If SQLs do not sign, look at response time and the quality of the quote. Each number points to a specific stage, which is the whole point of separating the two statuses.

A worked example over one month

A worked example over one month

A three-person WordPress agency targets restaurants in Bordeaux. It sets four fit criteria: open for business, identified owner, existing website, at least one visible flaw. Out of 200 businesses found, 120 pass this filter and become qualified leads. These figures are an illustrative calculation, not an observed result. If 10% reply, that makes 12 conversations. If half confirm a need and a timeframe, 6 SQLs enter the pipeline. With a one-in-three close rate, two websites are sold.

The point of the exercise is not the precision of the percentages but knowing where to act: if the result is zero, you can see immediately whether you are short on leads, replies or meetings.

Our take

We built AutoLeads from one observation: a web agency that relies on word of mouth, freelance platforms or public tenders is at the mercy of its client flow instead of steering it. The MQL and SQL vocabulary is useful, but it was designed for teams that receive leads. An agency has to go out and find them.

That is why our tool detects businesses with a visible need in real time, on the web and Google Maps, without relying on a static database or outdated open data. Every lead comes with a concrete audit (SEO, advertising, customer reviews) and with a message that starts from that diagnosis. The prospect is not qualified by a hunch, but by a flaw you can show them.

The SQL, though, remains a human matter: it is your conversation with the owner that creates it. Our job is to get you there with the right contacts, and to turn every reply into a task so nothing falls through the cracks.

Tom

Filling your calendar with AutoLeads

If you want to apply this grid without giving up your evenings, AutoLeads offers three monthly plans: Starter at €249 per month, Growth at €499 per month and Scale at €899 per month. The subscription runs on credits, consumed according to lead volume and the enrichments you turn on. A "We call for you" option is available for outsourced appointment setting.

Grezy

You pick a trade and a city, AutoLeads delivers audited businesses with the owner, a verified email and a phone number, and you stay in control of what matters: the conversation that turns a lead into an SQL. To get started, head to the Grezy plans page.

Frequently asked questions

SQL stands for Sales Qualified Lead, meaning a lead qualified by sales. It is a contact that a salesperson has reviewed and accepted as a genuine prospect, with a confirmed need, decision-making authority and a realistic buying timeline, according to Rework. In the context of database languages, SQL means something else entirely, unrelated to prospecting.

An MQL is qualified by marketing based on its profile and behavior, often with a score. An SQL is validated by a salesperson after a conversation that confirms a real project. So an MQL is a signal, an SQL is a conversation.

A lead is a contact who has shown interest without in-depth qualification. A prospect is a lead that has been assessed and qualified, matches the target and shows concrete commercial potential, as explained by La Poste and Effy. The difference lies in the level of maturity and how clear the need is.

Yes, as soon as you prospect regularly. Two written statuses are enough: a lead qualified on fit and visible need, then an SQL after a conversation. This stops you from counting contacts who never replied as opportunities, and gives you conversion rates you can actually read.

A real problem identified, an offer that solves it, a priority and a timeframe, a relevant contact involved and a next step in view, according to Hyperops. Record these five points on the lead's record at the time of the meeting.
Book a demo