Playbook
How to Fix the MQL-to-SQL Handoff: The Canonical Playbook
Why does the MQL-to-SQL handoff leak?
Short answer: because two teams with opposite incentives score against different definitions, and nobody owns the leads caught in the seam. Marketing is measured on the volume of qualified leads, so its threshold drifts down until a whitepaper download counts as intent. Sales is measured on closed revenue, so it discards anything that looks thin without recording why. Both teams behave rationally, and the pipeline absorbs the cost.
The most expensive failure is not the disagreement, it is the silent rejection: a lead routed to a rep who never accepts it, never rejects it, and never logs a reason, so it ages out of the queue invisibly. It does not show up as churn or as a loss. It shows up as nothing. That is why the handoff is a pipeline-integrity problem before it is a lead-quality problem: you cannot trust a funnel whose largest leak is unmeasured. Firmographic-only qualification compounds it by scoring who an account is instead of whether it is ready. For where the leak shows up in the numbers, see the MQL-to-SQL handoff data brief.
How do you fix the MQL-to-SQL handoff?
Treat the handoff as a transaction with a receipt, not a hope. Five moves turn it from a weekly argument into a governed, measurable process.
Step 1: Write one shared, signed definition
Put both teams on a single written definition of what qualifies as a handoff-ready lead and what a rep must confirm to accept it. The document names the fit bar, the intent bar, the fields that must be populated at handoff, and the valid rejection reason codes. Keep the reason-code list to five to seven items: enough to separate an actionable rejection (bad timing, wrong contact) from a non-actionable one (not ICP fit), few enough that reps actually select carefully. Both leaders sign it, and it is reviewed quarterly, not filed and forgotten.
Step 2: Score for pain, not just firmographic fit
A shared definition only holds if the score behind it predicts a deal. Replace firmographic-only qualification with a two-model score: firmographic fit capped near 20% of the weight as a disqualifier, and a pain and intent model weighted 60-80% that watches trigger events like recent funding, a new exec in a target role, or a pricing-page return. This is the Pain-First ICP Scoring framework, and it is the reason the handoff bar becomes defensible: sales stops rejecting leads that "look thin" because the score now reflects readiness, not just resemblance. The deeper comparison lives in firmographic vs pain-based ICP.
Step 3: Set a tiered handoff SLA with a timer
The most common SLA failure is a single blanket rule like "sales must follow up within 24 hours," which ignores that not all MQLs are equal. Tier it instead:
- Tier 1, high intent (demo request, pricing page plus contact form, trial signup): respond within minutes to an hour.
- Tier 2, moderate intent (content download plus a return visit, webinar engagement): respond within one business day.
- Tier 3, fit-only MQLs (scoring threshold met, mostly fit): respond within two to three business days.
Then define what "follow-up" counts as (a logged outbound activity in the CRM, not an unrecorded "I reached out") and wire the timer into the CRM so the clock is enforced. The urgency is not arbitrary: Harvard Business Review's 2011 study put the average first response to a web lead at about 42 hours, and the same lead-response research shows qualification odds fall sharply as first response is delayed.
Step 4: Route to a named owner with an acceptance step
Route each lead to a specific rep or queue using deliberate territory or segment logic, not an unowned shared inbox, and require an explicit accept-or-reject action within the SLA window. Acceptance is a discrete, logged event ("I have reviewed this lead and I am accepting it for outreach"), and every route needs a fallback owner so a lead is never stranded on a rep who is out of office. Pass the score breakdown, not just the total: a 72 that is mostly fit means something different from a 72 that is mostly behavioral signal, and the rep needs the composition to open the conversation well.
Step 5: Govern the silent-rejection loop
Track the share of MQLs that reach sales and receive no disposition at all. That silent bucket is your handoff leakage, and it is more honest than any conversion rate, because conversion rates only count the leads someone bothered to process. Govern it toward zero: every MQL gets a yes with a logged next step or a no with a structured reason code, and the rejected leads flow back to marketing weekly so the scoring model calibrates instead of shipping the same miss next month. A rejected lead is not a failure, it is the fastest signal you have for improving the next batch.
The MQL-to-SQL handoff is where most B2B pipeline leaks, and the leak is usually structural, not effort. Firmographic-only qualification caps MQL-to-SQL conversion around 50-60% because it scores whether an account looks right, not whether it is ready. The fix is a written, signed definition both teams share; a two-model score that caps firmographic fit near 20% and weights pain and intent 60-80%; a tiered response SLA enforced by a CRM timer (HBR's 2011 study put the average first response at about 42 hours); an explicit accept-or-reject step; and a handoff-leakage metric governed toward zero, where leakage is the share of MQLs that reach sales and get no disposition at all.
What tools operationalize the handoff?
A handoff is only as good as the CRM that enforces it and the engine that acts on an accepted lead inside the SLA window. Attio is where the two-model score, the required-fields gate, the acceptance step, and the routing rules live, so the SLA is wired into fields and workflows rather than a slide nobody opens. Affiliate link.
Amplemarket sequences the accepted lead the moment it is routed, watching the same trigger events your score weights, so a Tier 1 handoff becomes a personalized first touch inside the response window instead of a task that waits for a rep to check a queue. Official Artemis GTM partner. Affiliate link.
The Lead Scoring agent ($349) builds the part of this playbook that most teams get wrong: it caps firmographic fit as a disqualifier, weights the pain and trigger signals that predict a near-term deal, validates the thresholds against your closed-won data, wires the BDR queue rules and the acceptance step, and installs monthly re-tuning so the handoff bar does not drift. See how the queue itself is built in BDR queue routing from lead scores, or read the full scoring model in the lead scoring guide.
Frequently asked questions
What is the MQL-to-SQL handoff?
Why does the MQL-to-SQL handoff leak pipeline?
What should a marketing-sales SLA include?
How does a lead scoring agent fix the handoff?
What is a good MQL-to-SQL conversion rate?
Sources & References
- Oldroyd, McElheran & Elkington, "The Short Life of Online Sales Leads," Harvard Business Review (2011): the lead-response research on how sharply qualification odds fall as first response is delayed. The 21x lift for acting within five minutes comes from the same lead author's Lead Response Management study (Oldroyd, 2007).
- HubSpot, "Service Level Agreements: 7 Steps to Writing an Effective One": the reciprocal marketing-sales SLA structure: shared definitions, response commitments, and closed-loop reporting on accepted and rejected leads.
- Gartner Sales Survey, 2025 (reported by Demand Gen Report): finds 73% of B2B buyers actively avoid suppliers who send irrelevant outreach, the case for scoring pain before the handoff, not raw fit.
Run this play in your own stack
Read the guide, then install the engine.
The Artemis AI GTM Engineer prices this leak in dollars before it recommends anything, then builds the fix with you inside your own Claude. See how an agent installs and buys, or start with the free audit that prices all seven leaks.