Sales and marketing alignment often fails at an awkward moment: marketing celebrates a pile of qualified leads, while sales says none is worth calling. Both sides have dashboards. Neither side agrees what happened to the buyer.
If that sounds familiar, the fix is not another debate about which department works harder. It is a short, evidence-led MQL audit that shows what qualifies a lead, when it moves, whether anybody responds and whether it becomes pipeline. Recent B2B research suggests many teams need that reset.
Key Takeaways
- A 2026 survey of 192 B2B practitioners found that only 7% of respondents reporting MQL volume trusted it to predict pipeline.
- An MQL audit should separate poor-fit leads from missed routing, inadequate follow-up and genuinely stalled opportunities.
- A 30-day repair plan can improve definitions and handoffs, while revenue outcomes may take longer to assess.
Table of contents
Jump to each section:
- Why is the MQL still causing arguments?
- What does sales need before it can trust a lead?
- How can you audit 100 recent leads fairly?
- Which metrics expose the real handoff problem?
- What should a shared lead SLA actually say?
- What should happen in the first week?
- What should change in weeks two and three?
- What should the team prove by day 30?
- Can AI qualification fix the gap?
- What tools are enough to start?
- What if the product or positioning is the problem?
- What should leaders stop rewarding?
Why is the MQL still causing arguments?
The MQL stays on executive dashboards because it is easy to count. But a count of contacts that scored enough points does not show whether sales can turn them into real opportunities.
The most striking evidence comes from The Trust Gap, a study fielded by Heinz Marketing for B2B vendor Docket in September 2026. It surveyed 192 marketing, demand generation and revenue operations practitioners. The sample is useful evidence of practitioners’ views, not a representative census of all B2B companies.
Nearly 90% report MQL volume to leadership, but among those who report it, only 7% are confident it predicts pipeline.
The denominator matters: the 7% finding applies to respondents who report MQL volume, not all buyers or all companies. The same research says 78% of respondents’ organizations have sales teams that reject more than half of the MQLs marketing sends. These are self-reported outcomes, so they do not prove that any single team’s leads are bad.
As the research partner’s analysis makes clear, the trouble is not confined to one handoff. Buyers abandon forms, marketing sends records without enough context, and sales follow-up often arrives slowly or generically. A lead report can look full while the buyer’s actual experience is poor.
The useful question is therefore not which department deserves blame. It is where otherwise plausible buying interest stopped moving and what evidence would have allowed the next person to act.
What does sales need before it can trust a lead?
Sales needs fit, a reason to talk, evidence of the buyer’s current problem and a credible next step. A form fill or content download can indicate interest without establishing those conditions.
A marketing qualified lead, or MQL, commonly means a contact that meets marketing’s fit and engagement threshold. A sales qualified lead, or SQL, is a contact that sales has vetted for a meaningful conversation. The Salesforce alignment guide argues that both teams must agree on the ideal customer profile and on the event that moves a record between these stages.
That sounds basic until someone asks to see the record. Does it name the right company and role? Did the buyer request a demonstration, or did someone merely download a broad industry report? Do we know the use case, buying timeline and who else is involved? Was the previous conversation recorded accurately?
| Evidence | Weak handoff | Useful handoff |
|---|---|---|
| Fit | Industry or email domain alone | Account type, role and relevant use case |
| Intent | A generic engagement score | A stated problem or specific buying action |
| Timing | A fresh timestamp | Known urgency and next-step window, when available |
| Context | A form field dump | Conversation summary and the buyer’s own questions |
| Ownership | Assigned queue | Named rep with a clear response obligation |
Not every buyer will disclose a budget or timeline before speaking with sales. That is fine. Treat missing information as missing, rather than automatically labeling the lead bad or inventing intent from a score. Your qualification model should distinguish not yet known from known not to fit.
How can you audit 100 recent leads fairly?
Take one recent, clearly defined MQL cohort, trace each record through its sales outcome and classify what actually happened. Review evidence before changing scores or targets.
A recent LinkedIn discussion by fractional CMO Sjeel K. proposes using 100 MQLs to settle this familiar dispute in 30 days. The suggestion is a useful starting sample, not a validated statistical threshold. For a small company, all leads from a quarter may be more informative; for a high-volume team, a stratified sample across sources prevents one campaign from dominating the answer.
- Freeze the cohort. Choose an exact creation window and inclusion rule. Export lead ID, source, campaign, fit score, MQL timestamp, owner, first human contact, disposition and opportunity outcome. Remove duplicates and record how many were excluded.
- Reconstruct the timeline. Compare the MQL timestamp with routing, first attempt, first response, meeting, acceptance and opportunity dates. Distinguish an attempt from a genuine conversation.
- Sample actual context. Read relevant notes or call transcripts where permission permits. Look for stated business needs, questions, objections and next steps. Do not assume that CRM drop-downs capture the full picture.
- Assign evidence-based outcomes. Mark each lead as poor fit, sales-unworked, worked but not progressing, converted or unresolved. Add a separate reason field for timing, product fit, contact quality and data gaps.
- Review disputed cases together. Have one marketing owner, one sales owner and a neutral RevOps participant examine the ambiguous records. Save disagreements rather than forcing false consensus.
- Repeat by source. Compare inbound demo requests, webinars, downloads, referrals and paid campaigns separately. The blended average can hide a channel whose handoff works well.
Protect private buyer data during the audit. Only grant transcript or CRM access to people who need it, and avoid copying personal details into an external AI tool without organizational approval. The goal is to improve the funnel, not create a second uncontrolled customer database.
Do not decide the verdict first. A bad-fit lead is not proof that all marketing leads are poor, just as an unworked lead does not prove sales deliberately ignored it. Routing failures, capacity constraints and missing contact details may be separate causes.
Which metrics expose the real handoff problem?
Track accepted leads, timely contact, meaningful conversations, opportunity creation and eventual revenue by cohort. MQL volume on its own cannot distinguish poor fit from poor follow-up.
The 2026 Unbounce survey of more than 500 SMB go-to-market professionals found that 87% expected better alignment to improve performance, but only 56% described their teams as highly aligned. In that sample, highly aligned organizations were more likely to report strong cross-functional data sharing. That is an association, not evidence that one meeting or tool causes growth.
59% versus 16% reported strong cross-functional data sharing among highly aligned teams compared with less aligned teams, a 3.5-times difference.
| Metric | What it reveals | How to calculate |
|---|---|---|
| Sales acceptance rate | Whether the lead passes the shared threshold | Accepted MQLs / MQLs routed to sales |
| Time to first human contact | Routing and capacity delays | First verified contact time minus handoff time |
| Worked-within-SLA rate | Whether owners acted in the agreed window | MQLs worked on time / eligible routed MQLs |
| Meeting conversion | Whether outreach creates real conversations | Meetings held / relevant routed MQL cohort |
| MQL-to-opportunity rate | Whether qualification predicts pipeline | Opportunities created / MQL cohort |
| Revenue per source | Long-run business contribution | Closed-won value attributable to a defined source and cohort |
Keep definitions stable. A meeting booked is not a meeting held. A lead marked accepted is not necessarily sales qualified. And influenced revenue is not the same thing as revenue caused solely by marketing. Every dashboard should state its denominator, date window and attribution method.
For a new month-long experiment, lead response and meeting metrics can move first. Opportunity and closed-won metrics may need longer because B2B sales cycles rarely fit into 30 days. Present early process improvements as such, rather than claiming they already proved revenue impact.
What should a shared lead SLA actually say?
A useful service-level agreement defines an eligible handoff, the assigned owner, a response window and the feedback that must return to marketing. It also explains exceptions.
The Highspot alignment guide recommends shared lead-scoring rules and explicit handoff agreements so that campaigns do not generate leads that sit untouched. But signing an SLA changes little if the CRM cannot show who acted or the criteria are impossible to apply.
Start with a short operational contract: marketing sends only leads meeting agreed fit and action criteria; RevOps checks routing; sales accepts or returns each eligible record with a reason; both teams review unmatched records weekly. Assign a single owner to maintain the definition and change log.
Set different response targets for different intent levels. A demo request may merit a rapid human response during staffed hours, whereas an educational download might enter a nurture sequence. Choose actual targets from your staffing and buyer expectations, not a universal claim that every lead must be called within five minutes.
Spell out holidays, time zones, weekends, invalid contact data and existing customers. Without those exclusions, teams can game the timer and produce a misleading compliance number. A fair SLA is one people can check from a shared system, not a promise written only for a leadership slide.
What should happen in the first week?
Week one should establish a single data set and produce a map of where leads fail. Do not rewrite the scoring model before learning whether the original records were routed or contacted correctly.
Ask RevOps to export the sample using one query and preserve the field definitions. Compare it with the sales team’s report. If totals disagree, diagnose duplicate contacts, lifecycle dates, multi-account ownership and manual overrides before debating the campaign’s quality.
Next, inspect the timestamp chain. A lead that became an MQL on Friday evening may have been assigned on Monday and contacted on Tuesday. Whether that was a failure depends on the buyer’s request, coverage hours and stated SLA. Document the rule that applied at the time.
Finish the first week with a practical diagnostic: how many leads lacked fit, how many were never routed, how many were routed but never worked, how many received a relevant response, and how many converted? Mark each outcome as known, disputed or missing. This creates an audit trail for the next meeting.
What should change in weeks two and three?
Weeks two and three are for testing a revised handoff and making the buyer experience better. Change the smallest number of rules that account for most observed failures.
If bad-fit records dominate, refine the ICP and form qualification fields. If good-fit leads go unworked, fix ownership, alerts and capacity. If replies are generic, include the buyer’s topic, recent interactions and requested next step in the handoff summary.
If marketing and sales disagree on what qualifies, run a calibration meeting using anonymized real examples. Have each participant decide whether the lead should be contacted now, nurtured or disqualified, then explain why. Turn agreements into specific CRM fields and routing rules.
Avoid changing the score threshold, the sales script, the target audience and the email sequence all at once. That would make any later gain difficult to interpret. Pick one or two changes, log when they went live, and compare the new cohort with the original using the same definitions.
The goal is not to find a winner in the argument. It is to create a repeatable test that a new hire could understand. Write down what counts as accepted, what counts as worked, and what evidence qualifies a lead as an opportunity.
What should the team prove by day 30?
By day 30, the team should be able to explain where leads are lost, who owns each fix and whether short-term handoff metrics improved. It should not promise that one month proves a higher close rate.
Show leaders a before-and-after cohort table. Include sample size, source mix, accepted lead share, worked-within-SLA share, meeting rate, opportunity rate where mature, and a breakdown of rejection reasons. Also show how many cases remain unresolved.
One shared number is helpful only if both sides can inspect the inputs. Give marketing and sales access to the same pipeline view, while allowing each team to retain tactical leading indicators. Marketing still needs campaign conversion data; sales still needs activity and forecast data. The executive view should connect the two.
If the experiment changed definitions, label the resulting time series accordingly. A smaller MQL count after tightening fit criteria is not automatically a decline. Better leads may mean fewer leads, while a more permissive SQL definition can inflate apparent improvement. Compare outcomes at equivalent stages and disclose methodology changes.
Can AI qualification fix the gap?
AI can help summarize buyer context, route records and surface follow-up tasks. It cannot prove lead quality unless the resulting records are measured against actual sales outcomes.
In Docket’s survey, about seven in ten respondents said they were piloting, deploying or scaling an agent for buyer engagement, while only about one in ten organizations running an agent said they read results from a dashboard or CRM report. Those bases are different, and the figures do not establish whether AI-qualified leads convert better.
Follow-up research on MQL-to-SQL rejection makes the distinction clear: interest captured by an interaction is not the same as a buyer who has supplied relevant qualification context. A conversation agent may capture a problem, timeline and agreed next step that a basic form misses, but those details still require a consistent acceptance standard.
For an early pilot, ask whether AI-routed leads received faster appropriate responses, whether sales found the summaries accurate and whether the meetings progressed. Keep the human review process for uncertain fit, sensitive data and unusual purchasing journeys.
An automated system that produces more confident-looking summaries without improving opportunities is simply a more polished version of the old MQL problem.
What tools are enough to start?
You can run a useful audit with a CRM export, a shared sheet and a small weekly review. Dedicated revenue-intelligence or workflow tools become more useful after the definitions are stable.
| Tool or system | Use in the audit |
|---|---|
| CRM reporting, such as Salesforce or HubSpot | Reconstruct lifecycle timestamps, owners, dispositions and opportunity outcomes |
| Spreadsheet or BI dashboard | Compare one common cohort, denominator and campaign source |
| Call recording or sales notes with approved access | Check buyer questions, relevance and contact quality |
| Workflow automation | Alert an owner when eligible records are not worked within the agreed window |
| Revenue attribution platform | Link campaign sources to opportunities once stages and CRM data are trustworthy |
Tool selection should follow the failure you found. ContentGrip previously covered how Integrate and CaliberMind intend to connect demand capture with revenue data; that type of platform helps only when the underlying handoff definitions are usable. A sophisticated dashboard built on ambiguous lifecycle stages will make disagreement faster, not smaller.
What if the product or positioning is the problem?
Sometimes marketing and sales can execute the handoff correctly and still fail to create pipeline. That points toward product fit, positioning, pricing, competitive pressure or the wrong target segment.
Check whether the rejected accounts are genuinely in your ideal customer profile and whether qualified prospects repeatedly cite the same missing feature or commercial objection. If buyers request a capability your product cannot deliver, the remedy is unlikely to be a faster callback.
Also distinguish demand captured from demand created. Many B2B buyers research vendors before making contact; the absence of a fresh MQL does not imply the company lacked influence earlier in the journey. Account-level buying signals and customer interviews can help interpret the funnel, but they should not be manufactured into deterministic lead scores.
Send repeat objections to product and positioning owners as a separate action stream. A good audit can reveal uncomfortable strategic questions, not merely operational fixes. It should make those questions visible rather than forcing them into a sales-versus-marketing scoreboard.
What should leaders stop rewarding?
Stop rewarding a volume target that nobody trusts without also showing what those leads became. Reward cleaner handoffs, qualified pipeline and reliable feedback instead.
Sales and marketing alignment is an operating agreement. It is made of definitions, timestamps, ownership, buyer context and a shared view of outcomes. The executive meeting should review those ingredients before it celebrates the headline MQL count.
The next step is straightforward: export a recent cohort, inspect 100 records if that is an appropriate sample, and reconcile each lead’s fit, follow-up and outcome with both teams in the room. After that, write one clear lead standard and test it for 30 days. You may discover marketing needs better targeting, sales needs a better follow-up system, or both need a product story that buyers believe. The data should decide.