Sales Blames Marketing. Marketing Blames Sales. Here's the SLA That Ends the War.

Every revenue leader has sat in a pipeline review where the same conversation plays out like a script. Sales says the leads are junk. Marketing says Sales never follows up. Sales says they followed up on the ones worth calling. Marketing says Sales has no idea what qualified looks like. Heads nod. Someone takes an action item. Nothing changes. Two weeks later, the same conversation happens again.

This is not a people problem. It is an infrastructure problem. The two teams are operating from different definitions, different data, and different incentive structures — with no written agreement between them. A service level agreement, or SLA, is the document that ends the argument, not by forcing goodwill, but by making the rules of engagement explicit enough that both sides can be held to them. Here is exactly how to build one that works.

The data makes the cost of inaction unusually clear. Yet in practice, only a small fraction of B2B SaaS companies have formalized the relationship between marketing and sales in writing. The rest are running on trust, goodwill, and quarterly off-sites — none of which survive a missed number. What follows is the operational blueprint: five components of a real sales-marketing SLA, how to implement each one, and what the shared workflow looks like once it is running.

Section 1: Why the Blame Game Exists — and Why Good Intentions Don't Fix It

The misalignment between sales and marketing is almost always described as a cultural problem. That framing is comfortable because it implies the solution is a better relationship — more lunches, more joint planning sessions, more empathy. That framing is also wrong. Culture is downstream of structure. When two teams are measured on different outcomes with no shared definitions and no written handoff agreement, friction is not a failure of character. It is the predictable result of a broken system.

The Metric Mismatch at the Root of Every Pipeline Fight

Marketing is typically measured on MQL volume. Sales is typically measured on closed deals. When those two metrics diverge — which they will, because they measure different things at different points in a long funnel — both teams have data that proves the other one is failing. Marketing can show a spreadsheet full of leads generated. Sales can show a pipeline full of leads they never touched because the leads did not meet their informal standard of quality. Both datasets are accurate. Neither team is lying. The system is just designed to produce this outcome.

The result is what practitioners call the MQL graveyard: leads that pass marketing's threshold, get assigned to a sales rep, and then quietly die with no follow-up, no rejection reason, and no feedback to marketing about why. Industry practitioners have described this pattern bluntly: marketing pumps volume, sales gets buried in junk MQLs, and trust between the two teams collapses — the funnel looks like progress but the middle is a graveyard.

The Definition Problem Is More Serious Than Most Teams Admit

Ask the VP of Marketing what an MQL is. Then ask the VP of Sales what an SQL is. In most companies below $30M ARR, you will get two different answers that do not connect. Marketing has a lead score threshold. Sales has a gut feeling. Neither is written down. Neither has been agreed to by the other team. This is the single biggest reason MQL-to-SQL conversion rates vary so dramatically across companies — two companies in the same vertical can report 13% and 42% conversion rates, both accurately, because they are measuring different things.

Speed-to-Lead Is Not a Preference — It Is a Revenue Variable

The follow-up timing problem is equally structural. Research from MIT's 2007 Lead Response Management study, repeatedly replicated, found that companies responding to an inbound lead within five minutes are 100 times more likely to make contact than those waiting 30 minutes, and 21 times more likely to qualify that lead. Despite this data being two decades old, a 2024 study of 1,000 B2B SaaS companies by RevenueHero found that 63.5% never replied to a demo request at all — actually a worse outcome than the 23% non-response rate documented by Harvard Business Review in 2011. The problem has gotten worse, not better.

This is not because sales reps are lazy. It is because no one has written down an agreed response time, instrumented it in the CRM, or created accountability for it. Sales reps spend only about 30% of their workday on actual selling; the rest goes to admin, research, and CRM updates. Without a formal SLA, slow follow-up stays invisible until a deal is lost.

The Feedback Loop Is Usually Missing Entirely

The final structural gap is the feedback loop from sales back to marketing. When a sales rep rejects an MQL — deciding not to pursue it — the reason almost never makes it back to the marketing team in any usable form. There is no required field in the CRM capturing why the lead was rejected. Marketing therefore cannot adjust targeting, messaging, or scoring criteria because they have no signal about what sales is seeing. This is a solvable data problem dressed up as a people problem.

Section 2: The Five-Component SLA Framework

A sales-marketing SLA is a written document, ratified by both teams and their leadership, that specifies five things: a shared ICP definition, agreed MQL and SQL criteria, lead follow-up time commitments, a structured feedback loop cadence, and a shared dashboard both teams can see in real time. Each component is operational, not aspirational. Here is how to build each one.

Component 1 — Shared ICP Definition

The ICP definition is the foundation everything else is built on. It should be specific enough to be exclusionary. A strong ICP statement for a B2B SaaS company serving the VANDFORT target market might read: "B2B SaaS companies with $5M–$30M ARR, headquartered in North America, with a dedicated VP of Sales or VP of Revenue Operations, running HubSpot or Salesforce as their CRM, and showing indicators of recent GTM investment." Every word in that definition is a disqualifier for someone. That is the point.

The ICP definition should be jointly drafted — not handed to marketing by sales, and not handed to sales by marketing. Both teams bring data to the conversation: marketing brings lead source conversion data by segment, and sales brings closed-won and closed-lost patterns from the CRM. The output is a one-page document that lives in the CRM as the reference definition, reviewed and updated quarterly as the market evolves. This is the foundation of your GTM operations infrastructure.

Component 2 — Agreed MQL and SQL Criteria

MQL and SQL definitions must be written in terms both teams agree on, not in terms of a lead score number that only the marketing automation system understands. A workable MQL definition requires both firmographic fit — the account matches the ICP — and behavioral signals, such as a pricing page visit, a demo request, or repeated high-intent content engagement. A lead that downloads a whitepaper from a company that does not match the ICP is not an MQL. A lead that downloads it from a matching company and then visits the pricing page twice is.

The SQL threshold should be the point at which a sales rep has had a qualifying conversation and confirmed: a real pain point exists, the prospect has authority or clear access to authority, there is a realistic timeline for decision, and budget is present or accessible. BANT-style criteria still work here, with the caveat that in modern SaaS buying, budget is often the last thing confirmed, not the first. Build flexibility into the SQL definition around budget while keeping the other three criteria firm.

Component 3 — Lead Follow-Up Time SLAs

Follow-up time commitments should be tiered by lead intent signal, not applied uniformly. A prospect who submits a demo request is not the same as a prospect who downloaded a content asset three weeks ago. The SLA should reflect that difference explicitly. Companies with a formal response-time SLA respond within 15 minutes at nearly twice the rate of companies without one — 54.9% versus 29.5% — which means the act of writing down the commitment meaningfully changes behavior even before any tooling is added.

Component 4 — Feedback Loop Cadence

The SLA should specify a standing meeting cadence between marketing and sales leadership — not as an optional alignment ritual, but as a contractual obligation of the agreement. A bi-weekly 45-minute session is sufficient for most companies in the $8M–$30M ARR range. The agenda is fixed: review MQL volume versus target, review MQL-to-SQL conversion rate versus baseline, surface the top three rejection reasons from the prior two weeks, and agree on any definition adjustments. This cadence is what makes the SLA a living system rather than a document that gets signed once and ignored.

Component 5 — Shared Dashboard

Both teams need to see the same numbers in real time, not in separate reports built by their respective ops people that produce different figures depending on how the data was pulled. The shared dashboard should show: MQLs created versus target (by source), MQL-to-SQL conversion rate (by source, so attribution is visible), average follow-up time against SLA (by rep), rejection reasons by category, and pipeline sourced from marketing as a percentage of total pipeline. This is precisely the kind of revenue intelligence infrastructure that turns alignment from an intention into a measurable operating condition.

Section 3: Implementation — Building the SLA in Six Steps

Section 4: The Operational Workflow — What the SLA Looks Like Running

An SLA is only as good as the process it governs. Here is the operational workflow that makes the five components above function as a system rather than a set of aspirational commitments.

Section 5: How to Tell This Story to Your Board

A sales-marketing SLA is not just an operational tool. It is a board-level narrative about how you are building a repeatable, scalable revenue engine. Here are the three frames through which investors and board members typically evaluate this work.

Section 6: The Gap the SLA Cannot Close on Its Own

A sales-marketing SLA solves the handoff problem. It does not solve the broader GTM architecture problem. Companies that implement a strong SLA often discover, in the process, three other issues that the SLA surfaces but cannot fix on its own.

The first is lead scoring that was never validated. Most companies are running lead scoring models built years ago by someone who no longer works there, never updated against actual closed-won data, and producing scores that have no statistical relationship with likelihood to buy. The SLA requires a valid MQL threshold — which requires a valid scoring model — which requires going back to the data and rebuilding the model against outcomes. This is GTM operations work, not a conversation.

The second is territory and routing logic that does not reflect the ICP. If you have sales reps covering territories defined by geography, but your ICP is defined by company size and tech stack, your routing rules are systematically creating coverage gaps. The best leads may be going to the wrong reps, or to no rep at all. This surfaces immediately when you start tracking MQL-to-SQL conversion by rep against lead source — and it requires rethinking the routing logic, not just the SLA.

The third is that the ICP definition exercise almost always reveals fundamental questions about which segments you are actually winning in, which you are losing in, and why. Those questions deserve rigorous answers — not gut feel — before you publish a shared ICP definition that both teams are going to be held to. A proper GTM Audit is the structured process for answering them: it examines your CRM data, your closed-won and closed-lost patterns, your current scoring logic, and your handoff process, and produces a prioritized diagnosis rather than a list of assumptions.

The SLA is the right place to start. The GTM Audit is what ensures the SLA is built on a foundation that will hold.