MQL to SQL conversion rate: benchmarks and how to improve it

What is an MQL and what is an SQL?

A marketing qualified lead (MQL) has shown interest, usually by filling in a form, requesting a demo or engaging with content, and meets whatever threshold marketing has set to pass it on. A sales qualified lead (SQL) has been spoken to by a rep and graded against sales criteria, and the rep has accepted it as worth pursuing. This is the core of the inbound lead conversion engagement.

The gap between the two is where most B2B funnels leak, and it is almost always a definition problem rather than a lead quality problem. If sales has never written down what makes a lead qualified, marketing optimises for the number it can control, which is volume, and every handoff becomes an argument.

How do you calculate MQL to SQL conversion rate?

Divide the number of SQLs by the number of MQLs over the same period, then multiply by 100.

If 400 MQLs came in during a month and reps accepted 60 of them as sales qualified, the rate is 60 divided by 400, which is 15%. Moving that to 25% on the same 400 MQLs produces 100 SQLs rather than 60, which is 40 additional qualified conversations from no extra lead spend. That is the arithmetic behind every engagement described further down this page.

Two things quietly distort the number. Count both figures over the same cohort rather than the same calendar window, because a lead that arrives on the 30th has not had a fair chance to convert by the 31st. And decide in advance whether a recycled lead re-enters as a new MQL, or your rate will drift for reasons nobody can explain.

What is a good MQL to SQL conversion rate?

First Page Sage publishes a benchmark built from its client data between 2019 and 2025, with its definitions stated. It puts the average MQL to SQL rate for B2B SaaS at 13%, with industries ranging from 10% for legal services and real estate to 26% for business insurance and HVAC.

Read any benchmark against its definitions before comparing. First Page Sage counts a lead as an SQL only once a salesperson has vetted it as a good fit and a meeting has been booked. A team that calls a lead an SQL as soon as a rep replies will report a much higher rate on identical performance. Published figures elsewhere range from the low teens to over 40%, and few of them say how they define an SQL.

Use a benchmark to sanity-check, then baseline your own rate by source over the last 90 days. The goal is your trend. At Locus the rate moved from 10% to 25%, well above the published average, with no additional leads.

Why is your MQL to SQL conversion rate low?

A low MQL to SQL rate almost never means marketing is sending bad leads. It usually means nobody has written down what qualified means, so every rep applies a different standard and the handoff leaks.

At Locus, moving this metric from 10% to 25% took no additional lead volume. It took a written definition, a follow up standard, and reporting that made the gaps visible.

Why buying more leads makes it worse

Shobhit explains the mechanism in his short video on this. Marketing hands a hundred leads to three reps. Because nobody has written down what qualified means, each rep applies their own standard to decide who is worth calling. The leads that fail are not rejecting the product. Most are abandoned after a single attempt and lost to silence.

Now double the leads to make up the shortfall. The reps, already stretched, cherry-pick harder and abandon a larger share. The conversion rate falls and the cost of acquiring each customer roughly doubles. The fix is structural: a written definition of qualified, and a fixed number of follow-ups before any rep is allowed to drop a lead. That is steps one and two below.

How this was done at Locus

The method on this page was worked out one inbound call at a time.

The job: qualify the inbound nobody was calling

Shobhit Gupta was hired at Locus, a logistics SaaS platform, because inbound enquiries were arriving with nobody to call them. The brief was to qualify each one on BANT and hand the survivors to sales. For the first seven to eight months that was the whole role, call by call, including explaining the product and running demos where an enquiry warranted one. He had carried the SDR role before, at mFino and then Avancer Corporation, where he delivered over 100 SQLs in five months.

What BANT got wrong at first touch

Two things became obvious quickly from applying BANT several times a day. Budget is close to useless on a first inbound call, because the prospect usually does not know it yet, and pressing for a number produces either a guess or a defensive answer. And Need is where the entire judgement sits, while BANT gives no help with it. Whether “need” means a real operational problem or idle curiosity is the decision, and the acronym is silent on how to tell.

What made the difference was writing down what Need meant for Locus specifically, in language two reps would grade identically. In logistics software, value scales with throughput, which is why the logistics ICP is built around a volume floor rather than company size. Screening every lead personally is also where the habit of writing the disqualifying half down came from.

Tightening what was allowed to count

Most of the movement from 10% to 25% came from tightening what was allowed to count as qualified, not from working harder on the leads already in the system. The criteria were written as evidence rather than feeling. Not “prospect is interested”, but “prospect has named the problem in their own words and confirmed who signs”. Then they were enforced at stage exit, so a lead could not become an SQL without them.

Owning the CRM underneath it

A conversion rate is only as true as the CRM it is calculated from. At Locus the revenue operations work sat with the SDR leadership: Shobhit administered Salesforce, ran the data migration, and owned the tooling rather than filing tickets about it. Conversion was reported weekly by source and by rep. By source, inbound arrived steadily from South East Asia and India and barely at all from the United States, so the US and Europe were opened from outbound as a separate motion, with a distributed team of twelve built against that cadence.

What it produced

MQL to SQL conversion moved from 10% to 25% and lead to closure from 5% to 11%, with no additional lead volume. Over three years, Shobhit helped create more than $20 million in new sales opportunities there. Ruchi Mittal, GrowthStack’s co-founder, ran inside sales at Locus alongside him. The full record is on Shobhit’s profile.

Step one: write the qualification criteria down

This is the step that moved the number at Locus, and it has five parts.

Start from the deals you won and lost

Pull your last twenty closed-won and closed-lost deals. For each, note what was true at the first conversation: company size, the trigger that made them look, who was on the call, the problem they stated, and their timeline. The attributes that show up in the wins and not the losses are your criteria. The same method builds the ICP.

Write four or five criteria, each with a question and a pass rule

A criterion is only usable if a rep knows what to ask and what answer passes. WorkStore, which sells Dyson commercial hand dryers to facilities teams, qualified every inbound lead on four criteria. The questions and pass rules below are illustrative, to show the format:

Treat budget with care. WorkStore sells a physical product to facilities teams, where budget is usually known. For software sold inbound, Shobhit’s experience at Locus was that budget is close to useless on a first call, because the prospect does not know it yet. There it belongs at the opportunity stage, and the first-call criteria should lean on need and volume instead.

Unit count was the criterion that most often separated a real opportunity from an enquiry at WorkStore. Every business has one like it. For Easyship, the line was a single number: eCommerce businesses shipping a thousand parcels a month or more.

Write the disqualifiers

Most teams skip this, and it matters as much as the criteria. Existing customers, competitors, students and job seekers, regions you do not sell into, and anyone below your threshold number. A written disqualifier saves a rep from spending twenty minutes deciding something the business already knows.

Run the two-rep test

Take five recent calls and have two reps grade each one against the written criteria, without discussing it. Disagreement on any lead means the criteria are not finished, and the gap is usually a word that means something different to each rep: “engaged”, “interested”, “a good fit”. Rewrite until the disagreement disappears. The method is in the sales qualification framework, which also compares BANT and MEDDIC as starting points.

Enforce it when a lead changes stage

The criteria only change the number if a lead cannot become an SQL without them. At Locus, the written definition was enforced at stage exit. In a CRM, that means the criteria fields have to be filled before the stage can move. Keep those required fields to the criteria themselves. A CRM that demands more fields than a rep will complete mid-call stops being updated, and every report built on it becomes fiction.

Step two: fix follow-up before anything else

Most MQLs are lost to silence rather than rejection. A lead worked once and abandoned is not a bad lead, it is an unworked lead. The evidence on speed is stronger than on any other lever in sales development. Contacting a web lead within five minutes rather than thirty made a rep roughly 21 times more likely to qualify it in the Lead Response Management study (MIT Sloan and InsideSales.com, 2007). An audit of 2,241 US companies found the average first response took 42 hours, with 23% never replying at all (Harvard Business Review, 2011).

Write the follow-up standard down, the same way as the criteria:

The last row is where many teams lose SQLs they already won. At ConvertCart, fixing follow-up and the confirmation flow cut meeting no-shows from 44% to 15%. More on that in reducing no-shows and speed to lead.

Step three: agree the handoff in writing

Most MQL to SQL arguments are really about ownership. Put the handoff on one page that both teams sign:

• The MQL definition, owned by marketing: what a lead must do or be to be passed on.
• The SQL criteria, owned by sales: the written standard from step one.
• The response standard from step two.
• Rejection reasons from a fixed list, so every rejected MQL tells marketing something: wrong company size, no need, wrong region, no response.
• The handover note. At WorkStore, qualified leads went to the client’s sales team with the qualification reasoning attached, not just a name and a number.

If you lead marketing and sales will not accept your leads, the version of this written for you is for marketing leaders.

Step four: make the funnel visible

You cannot improve a handoff you cannot see. At Locus, conversion was reported weekly by source and by rep. This is the minimum weekly report:

Watch the fourth row. A rising MQL to SQL rate with a falling SQL to opportunity rate means the standard is being loosened, not met. The reporting build is covered in GTM analytics and reporting.

Step five: coach against the standard every week

A written standard drifts unless someone checks calls against it. At GoComet, weekly call reviews against the written qualification standard were part of how output moved from 2 to 6 SQLs per rep per month while the team grew from 10 to 23. Review one recorded qualification call per rep each week, and every SQL the account executive rejected. The cadence is in SDR coaching.

Then recalibrate quarterly against closed-won. If SQLs that meet every criterion still do not close, a criterion is wrong. Lead scoring built from closed-won uses the same check.

How do you track MQL to SQL in HubSpot or Salesforce?

The mechanics differ, the discipline does not. Both systems report this accurately only if the stage definitions are written down and the fields that drive them are filled in.

In HubSpot

• Use the lifecycle stage property. It already includes Marketing Qualified Lead and Sales Qualified Lead, and can be customised. Decide what event moves a contact into each, and set it by workflow or require the criteria properties before a rep can change it.
• Use lead status for rejection reasons, with the fixed list from step three, so rejected MQLs can be reported on.
• Build a funnel report. In Reporting, create a funnel report on contacts with the stages Lead, MQL, SQL and Customer. It shows next-step conversion (MQL to SQL) and cumulative conversion. Custom funnel reports need a Professional or Enterprise subscription, and HubSpot notes the date range filters by the date of the activity, not the date the contact was created.
• Break it down by original source, so the rate can be read per channel.

In Salesforce

The process usually spans the Lead object and the Opportunity, so the equivalent measure is lead conversion with a qualification checkpoint before it. Use lead status with required fields at the qualifying status, and report on status transitions rather than record counts.

Shobhit Gupta has implemented and optimised both across the SDR and revenue operations functions at Locus, GoComet and Landmark Group. The failure mode was identical in each: too many required fields, a CRM that stops being updated, and conversion reports that stop being true. Which system you use matters far less than whether the team keeps it current, which is the argument in HubSpot vs Salesforce.

How do you track MQL to SQL back to Google Ads?

Everything above measures the funnel once the lead is in the CRM. If a meaningful share of those leads is bought, there is a second problem, and it is usually the more expensive one: the ad platform has no idea which leads were any good.

Left alone, Google Ads optimises toward the last thing it can see, which is the form submission. If your MQL to SQL rate is 15%, the platform is buying more of the 85%. That is why campaigns can look better in the ads dashboard every month while sales says lead quality is falling. Both are true. They are measuring different events.

Decide the event first

Send the SQL event rather than closed-won for most teams, because closed-won arrives too late and too rarely to train bidding on a normal budget. And do not send it until the two-rep test passes. Training a bidding algorithm on an inconsistent standard teaches it to optimise confidently toward noise.

If you use HubSpot

• Create a lifecycle stage conversion event. In Marketing, Ads, create an event, choose Google and your ad account, choose Lifecycle stage change as the trigger, and select Sales Qualified Lead. Add a value if you have one, and share email address for better matching. HubSpot’s guide has the full steps.
• Turn on enhanced conversions in Google Ads before syncing starts. HubSpot syncs a contact when it has a Google click ID or shared contact details.
• Know the limits. Only stage changes after the event is created count. Marketing Hub Starter allows 5 events, Professional 50 and Enterprise 100.

If you upload directly to Google Ads

• Capture the click ID with the lead. With auto-tagging on, Google appends a GCLID to the landing page URL. Read it there, carry it through the form, and store it on the CRM record. If it is not captured at submission it cannot be recovered. The GCLID is case sensitive.
• Create the conversion action, then wait. Google advises waiting 4 to 6 hours after creating it before the first upload.
• Upload SQLs as offline conversions against that action. Google’s offline conversion import guide sets out the timing windows, and Google now recommends starting with enhanced conversions for leads, which matches on hashed contact details rather than the click ID.

After this the bidding target changes. The platform optimises toward leads sales actually accepted, which usually means fewer leads at a higher cost per lead and a lower cost per opportunity. Tell whoever reads the ads dashboard before you switch it on.

Where the leak turns out to sit further down than the handoff, the same discipline across every stage is in how to audit a sales pipeline in 72 hours.

How long does it take to improve MQL to SQL conversion?

Inbound volume produces data quickly, so changes show up early. In GrowthStack’s client engagements, written criteria and a follow-up standard have moved qualification within the first one to two months. Give a change at least one full sales cycle before judging it on closed revenue.

Use the tool

Where has this actually worked?

The 10% to 25% figure comes from Locus, where Shobhit Gupta ran global SDR and revenue operations. It took no additional lead volume: a written definition of qualified, a fixed follow up standard, and weekly reporting by source and by rep. Lead to closure moved from 5% to 11% over the same period.

GrowthStack has since run the same sequence for clients. For Easyship, a multi-carrier shipping platform, that meant the full inbound path, from lead response to subscription closure and activation, with a team in India covering US and UK hours. For WorkStore, it meant a written four-criteria standard for paid media leads from Meta and Google Ads, with no change to ad spend. The figures for both are in the case studies.

In none of them did the fix involve generating more leads.

References

• Oldroyd, J.B. Lead Response Management Study. MIT Sloan School of Management and InsideSales.com, 2007.
• Oldroyd, J.B., McElheran, K. and Elkington, D. The Short Life of Online Sales Leads. Harvard Business Review, March 2011.
• First Page Sage. MQL to SQL Conversion Rate by Industry. Last updated December 2025.
• HubSpot. Create custom funnel reports. Knowledge Base.
• HubSpot. Create and customize lifecycle stages. Knowledge Base.
• HubSpot. Create and sync ad conversion events with your Google Ads account. Updated 9 September 2026.
• Google Ads Help. Set up offline conversions using Google Click ID (GCLID).
• Google Ads Help. Configure the Google tag for enhanced conversions for leads. All accessed 5 October 2026.

Related reading

• HubSpot and Salesforce setup, as an engagement
• If you own marketing and sales will not accept your leads
• ICP, persona and messaging frameworks
• GTM analytics and reporting
• Sales qualification frameworks: BANT vs MEDDIC vs yours
• How quickly to contact inbound leads
• Inbound lead conversion, as an engagement
• Sales pipeline audit
• Review your funnel on a call