- Template item
A lead scoring audit is a structured review of your scoring model, thresholds, and routing against real closed-won and closed-lost data. The goal is simple: find out why sales keeps rejecting your MQLs and fix the four defects that cause it. Most B2B teams see their MQL-to-SQL conversion rate slide under 15% within 18 months of launching a scoring model, usually because the ICP shifted and nobody updated the fit criteria. A marketing operations specialist can run the audit, rebuild the model, repair routing, and sign acceptance criteria with sales, typically inside 4 to 6 weeks.
If you already have a scoring model and sales still treats the MQL queue like noise, you do not need another tool. You need an audit and the right ops hire to run it.
What a Lead Scoring Audit Is (and When You Need One)
A lead scoring audit reviews four things against 12 months of closed-won data: your fit model (who you sell to), your engagement model (what signals convert), your thresholds (what counts as MQL and SQL), and your routing rules (who gets what, and how fast). The output is a diagnosis and a rebuild plan.
You need one when any of these show up:
| Signal | What it usually means |
|---|---|
| MQL-to-SQL conversion under 15% | Thresholds are too low, or fit scoring is giving points to the wrong firmographics |
| Sales rejects over 40% of MQLs | Fit and engagement are mixed into one score instead of split |
| SDRs ignore the queue | Routing is slow, or lead quality has trained them to ignore it |
| Closed-won deals that never scored MQL | Scoring model missed signals that actually predict revenue |
The HubSpot State of Marketing data points to lead quality as the single biggest friction point between sales and marketing. Gartner's CMO research tracks the same complaint across B2B marketing orgs. The audit is how you resolve it with data, not with another dashboard.
A good trigger: your pipeline forecast misses by more than 20% two quarters in a row and sales blames marketing-sourced leads. That is a scoring and routing problem, not a lead-volume problem.
One nuance worth flagging. An audit is not the same as a model refresh. A refresh tweaks weights; an audit decides whether the model is right at all. If the last person who touched the model has left the company, run the audit, not a refresh. For the related question of what the model should produce in the first place, the MQL vs SQL distinction is the right primer.
The 4 Defects Every Audit Looks For
Every lead scoring audit hunts for the same four defects. Each has a diagnostic signature, and each has a fix. The audit's job is to confirm which ones you have, then prioritize the rebuild by revenue impact.
Defect 1: Fit Model Drift
Fit model drift means your scoring still rewards the ICP you had 18 months ago. Signatures: closed-won deals cluster in a segment the model under-weights, or MQLs cluster in a segment with a 2% close rate. Diagnostic: pull closed-won from the last 12 months, bucket by firmographic, compare to current fit scores.
Defect 2: Engagement Model Inflation
Engagement model inflation happens when every pageview and email open adds points, so junk leads cross the threshold on behavior alone. Signatures: 30% of MQLs come from existing customers or competitors, or SDRs call MQLs who have never requested contact. Diagnostic: audit which behaviors predict a scheduled meeting versus ones that just correlate with activity.
Defect 3: Threshold Miscalibration
Threshold miscalibration is when the MQL cutoff is set by instinct, not by back-testing. Signatures: MQL volume is twice what sales can work, or MQL volume is a trickle and reps run out of pipeline. Diagnostic: back-test the current threshold against closed-won, what score did winning deals actually have at the moment they got routed?
Defect 4: Routing Failure
Routing failure covers slow handoffs, wrong-owner assignment, and dropped leads. Signatures: median time from MQL to first SDR touch over 24 hours, or round-robin assignment that ignores territory and account ownership. Forrester's research on sales response time is consistent on this: speed to first touch is one of the strongest predictors of whether a lead converts.
Fix one defect in isolation and you will not move the number. Fix all four and MQL-to-SQL can double inside a quarter.
A pattern worth naming: fit and engagement get fused into one score in roughly three out of four scoring systems MarketerHire-matched ops specialists audit. The fix is small (split the score, require minimum fit AND minimum engagement for MQL) but the impact is large. If you only read one defect on this page, read defect 2 again.
The 7-Step Lead Scoring Audit Process
The audit runs in seven steps, in this order. Skip step 2 and you will rebuild the model on bad assumptions. Skip step 7 and sales will reject the new model inside a week.
- Pull 12 months of CRM and MAP data. Closed-won, closed-lost, open opportunities, MQL log, routing log. If your data lives across HubSpot, Salesforce, and a warehouse, unify it before scoring anything.
- Segment by closed-won firmographics and behavior. What size, industry, role, and signal path do winning deals share? What do losing deals share? This becomes your new fit and engagement baseline.
- Back-test the current thresholds. For every closed-won deal, record the score at the moment it was routed. If 40% of wins came from leads that never crossed MQL, your threshold is wrong.
- Interview three to five SDRs and AEs. Ask what leads they actually chase versus which ones they skip, and why. The data will tell you what; sales will tell you why.
- Rebuild the model with explicit fit and engagement scores. Keep them separate. A lead needs both to qualify. Scoring fit and engagement as a single number is the most common defect in the field.
- Rebuild routing with SLAs. Define owner by territory, segment, and account. Define response-time SLA. Alert when a lead sits over threshold.
- Agree acceptance criteria with sales before go-live. This is a signed document, not a Slack message. See the acceptance-criteria section below for the exact fields.
A fractional ops specialist can run steps 1 through 7 in 4 to 6 weeks for a company under 500 employees. Full-time ops teams often take longer because they are also running day-to-day campaigns. The audit needs a dedicated owner.
The two steps most teams underestimate are 4 and 7. Interviewing reps feels soft; it is the single highest-signal input in the whole audit. And acceptance criteria feel bureaucratic until the first time a new model ships without them, at which point the scoring queue goes right back to being ignored. Build both into the project plan or the audit will not stick.
What a Marketing Operations Specialist Delivers
A marketing operations specialist brought in to run a lead scoring audit should deliver four artifacts and a 30/60/90 plan. Each artifact has an acceptance test, sales signs off before the project closes. Below is the deliverable set you should expect, with the acceptance criteria that make each one real.
| Deliverable | What it contains | Acceptance test |
|---|---|---|
| Current-state audit memo | Fit and engagement audit, threshold back-test, routing log analysis | Sales leader and marketing leader both sign |
| New scoring model doc | Separate fit and engagement rubrics, threshold logic, exception rules | Back-tested against last 12 months of closed-won |
| Routing + SLA diagram | Owner by segment, response-time SLA, escalation path | Pilot run on 50 live leads shows SLA hit on 90%+ |
| Dashboard + 30/60/90 plan | MQL-to-SQL rate, response time, rejection rate, pipeline sourced | Dashboard live in your BI tool, review cadence set |
Beyond these four, the specialist should also leave documentation your team can maintain. Scoring models drift. The audit is only the first fix; a quarterly review cadence keeps the model honest. If your ops hire does not hand off runbooks, you are buying a one-time project, not a system.
Acceptance Criteria to Agree With Sales Before Day 1
Acceptance criteria are the measurable targets you agree with sales leadership before the audit starts. Without them, you ship a new model, sales rejects it, and the audit is wasted work. Signed criteria flip the dynamic, the ops specialist builds to a spec, and sales commits to work the output.
Agree on these before day 1:
- Target MQL-to-SQL conversion rate, set a number (18%, 22%, 25%) based on segment benchmarks.
- SDR response-time SLA, time from MQL routed to first outreach. Common target: under 2 hours in business hours.
- MQL rejection ceiling, maximum percentage of MQLs sales can reject before the model is reviewed. Common target: under 20%.
- Pipeline-sourced quota contribution, what percentage of pipeline should marketing-sourced MQLs cover? 30 to 40% is standard for B2B SaaS.
- Review cadence, monthly scoring review for the first quarter, then quarterly.
- Veto rights, who can pause the model and request a re-audit, and under what conditions.
Write it as a one-page agreement. Both leaders sign. File it in the project doc.
One admission worth making: not every org can agree acceptance criteria in week 1. If sales is skeptical, run a two-week discovery sprint first, publish the current-state audit memo, show the back-test, and let the numbers earn the signature. Trying to force signoff before the data is on the table usually produces a thin agreement that nobody respects by week 6. See the demand gen team structure write-up for a parallel on how ownership gets negotiated.
Fractional vs Full-Time Ops: When Each Makes Sense
A fractional marketing operations specialist is the right hire for the audit itself and the model rebuild. A full-time hire makes sense once the system is stable and you need daily ownership of enforcement, reporting, and iteration. Below is the decision rubric.
| Factor | Fractional ops | Full-time ops |
|---|---|---|
| Scope | Audit, rebuild, one-time project | Daily enforcement, campaign ops, iteration |
| Stage | Series A to early B, under 150 employees | Series B+ or 150+ employees with 2+ MAP admins already |
| Timeline | 4-8 weeks to deliverable | 3-6 month ramp, ongoing |
| Budget | $8K-$15K/mo, month-to-month | $110K-$160K base plus benefits |
MarketerHire's match data across 30,000+ hires shows the same pattern across pre-Series B companies: a fractional ops specialist delivers the audit faster and cheaper than a search for a full-time hire, and the full-time role becomes clearer after the system is documented. Try fractional first. Convert to full-time once you know what the job actually is.
If you want context on where ops sits in the broader team, see where marketing ops sits in your org and marketing team structure. One honest caveat: if sales leadership refuses to sign acceptance criteria, no fractional hire will save the project. In that case, the first job is organizational, not operational, and a short engagement with a marketing analyst to produce the diagnostic case can be the lever that gets executive buy-in before you bring in ops.
The Scope Brief: Hand This to a Specialist
The scope brief below is what you hand to a fractional marketing ops specialist on day 1. It makes the engagement concrete, auditable, and finishable. Copy it, fill the brackets, send it. A good ops hire will push back on anything underspecified before quoting a timeline.
Project: Lead scoring audit and model rebuild
Context: We run [HubSpot / Marketo / Pardot / Salesforce] with [MAP] feeding [CRM]. Current MQL-to-SQL conversion rate is [X%]. Sales rejection rate on MQLs is [X%]. Current scoring model was last updated [date].
Deliverables:
- Current-state audit memo covering fit model, engagement model, thresholds, and routing.
- Rebuilt scoring model documentation with separate fit and engagement scores, back-tested against the last 12 months of closed-won.
- Routing rules and SLA diagram covering owner, response time, and escalation.
- Dashboard in [BI tool] tracking MQL-to-SQL rate, SDR response time, rejection rate, and pipeline sourced.
- 30/60/90 handoff plan with runbooks.
Timeline: 6 weeks from kickoff.
Acceptance criteria: All four deliverables signed by [sales leader name] and [marketing leader name]. Target MQL-to-SQL rate [X%] and SDR response-time SLA under [X hours]. See attached acceptance-criteria doc.
Data access needed: CRM admin read access, MAP admin read/write, BI tool read, 30-minute interviews with 3 SDRs and 2 AEs.
Decision-maker: [Name], [Title], final sign-off.
Reporting cadence: Weekly 30-minute check-in; final review in week 6.
Hand that document to any vetted ops specialist and you will get a credible quote back inside 48 hours. If a hire needs more than that brief to scope the work, you have the wrong hire.
FAQ
Audit lead scoring every six months at minimum, or any quarter where MQL-to-SQL conversion drops more than 20% off its trailing average. ICP, product, and buying behavior shift faster than most scoring models. A twice-yearly cadence keeps the model honest without burning ops cycles on constant retuning.
A healthy MQL-to-SQL conversion rate for B2B SaaS lands between 18% and 25%. Mid-market and enterprise segments usually sit higher because fit scoring filters volume earlier. If you are under 15%, the model is broken or thresholds are too low. HubSpot's benchmarks are a useful reference point.
Marketing operations owns the scoring model, but sales owns the acceptance criteria. The ops team builds and maintains it. Sales leadership signs off on thresholds, SLAs, and rejection ceilings before go-live. Shared ownership without a signed agreement is the top reason scoring audits fail inside the first quarter after launch.
A full lead scoring audit, model rebuild, and routing fix takes 4 to 6 weeks with a dedicated fractional ops specialist. Larger CRMs with messy data history add 2 to 4 weeks. The data pull and firmographic segmentation (steps 1 and 2) are the longest phases; the model rebuild is relatively fast once the back-test is clean.
The specialist needs 12 months of closed-won and closed-lost CRM data, the MQL and SQL log, routing history, current scoring rules from your MAP, and read access to pipeline dashboards. Plan on 30-minute interviews with 3 to 5 SDRs and AEs. Without the historical data, back-testing is impossible and the rebuild is a guess.
Yes, a lead scoring audit works on any CRM that logs activity and stage changes. The platform matters less than the data quality. Salesforce with any activity-tracking layer is enough. The audit's output (the new scoring logic and SLAs) can be implemented in HubSpot, Marketo, Pardot, or a custom warehouse rules engine.
What to Do Next
If your MQL-to-SQL conversion rate is under 15%, if sales rejects more than 40% of what marketing routes, or if your scoring model has not been touched in a year, you have a lead scoring audit to run. Hand the scope brief above to a vetted fractional marketing operations specialist and you will have a diagnosis in weeks, not quarters. MarketerHire matches you with a senior ops specialist in 48 hours, with a two-week trial and no long-term contract.
By Jenny Martin, Growth Marketing Editor, MarketerHire. Jenny has led growth across DTC and B2B SaaS, scaled revenue to $50M, and cut CAC by 40%. She writes about marketing ops, growth hiring, and what works at the $2-50M stage.
