Ask a marketing leader and a sales leader in the same Indian B2B company, separately, what an MQL is, and you will very often get two different answers — sometimes wildly different. I have sat in the room where marketing celebrated hitting 400 MQLs for the month while sales quietly called maybe 30 of them worth a real conversation. That gap is not a sales execution problem or a marketing quality problem in isolation. It is a definition problem, and definition problems compound every month until nobody trusts the number at all.
An MQL definition that sales does not trust is worse than having no definition — it creates a permanent, unresolved argument about whether marketing's leads are "good," when the real issue is that nobody agreed in advance what "good" means.
Indian B2B lead volume tends to be heavily inflated by two patterns that are less pronounced in more mature digital markets: high volumes of student and job-seeker traffic mistakenly captured through broad-match search or content aimed at "software" or "marketing" generically, and a large tier of very small businesses that fill out a form out of curiosity with no near-term budget or authority to buy. A raw lead count in India, unfiltered, contains meaningfully more noise than the same raw count in a market like the US or UK, which makes a rigorous scoring model even more necessary, not less.
The mistake most lead scoring models make is combining everything into one number. A lead who is a perfect fit (right industry, right company size, right role) but has shown low intent (downloaded one blog post, nothing else) should be treated completely differently from a lead who is a mediocre fit but has shown extremely high intent (visited the pricing page three times and started a demo request). Collapsing both into a single score of, say, 62 out of 100 hides the difference and leads to identical treatment for two leads that need completely different sales approaches. I always build two separate scores — fit and intent — and only combine them into a single MQL threshold at the end.
A threshold-setting approach that works better than picking numbers out of the air: pull your last 6-12 months of closed-won deals, score them retroactively against your fit and intent criteria, and set your MQL threshold at whatever score the bottom 10th percentile of actual closed-won deals achieved. This grounds the threshold in what your business has actually proven converts, rather than an arbitrary round number, and it is the single fastest way to get sales to trust the model — because you can show them, concretely, "every deal we closed in the last year would have cleared this bar."
A scoring model without a service level agreement between marketing and sales is just a report. The agreement needs to specify, in writing and reviewed quarterly: what score constitutes an MQL, how fast sales must make first contact after a lead crosses that threshold (5-15 minutes for high-intent form types like demo requests is achievable and materially improves conversion versus even a same-day follow-up), what counts as sales accepting versus rejecting a lead, and — critically — a required reason code when sales rejects a lead as unqualified, because that rejection data is what lets marketing actually improve the model instead of arguing about it anecdotally.
I review the fit and intent model at minimum quarterly against two numbers: MQL-to-SQL conversion rate (are the leads clearing the marketing bar actually getting accepted by sales) and SQL-to-closed-won rate by lead source (are the leads that do get accepted actually closing at a healthy rate). If MQL-to-SQL conversion is below roughly 50-60 percent, the threshold is too loose or a specific source is polluting the pool and needs to be isolated and re-scored. If it is above 85-90 percent, the bar may be set so high that real, convertible leads are being filtered out before sales ever sees them — a quieter but equally costly failure mode that fewer teams check for.
Lead scoring often gets handed to whoever owns the marketing automation platform and treated as a configuration task. It is actually a strategic marketing decision, because the definition of what counts as qualified shapes what marketing optimizes campaigns for. A team optimizing purely for MQL volume against a loose definition will build campaigns that generate volume. A team held to an MQL definition anchored in actual closed-won data will build campaigns that generate the kind of leads the business actually converts — a completely different, and much more valuable, set of campaign decisions.
Most scoring models only add points for positive signals. I also build in explicit negative scoring for signals that reliably predict a lead will never convert, which is just as important in the Indian context given the volume of non-buyer traffic that legitimately lands on B2B forms. Common negative signals: a personal email domain combined with a job title suggesting student or job-seeker status, a company size field left blank or filled with an implausible value, an IP or form-fill pattern suggesting the same person submitting multiple different forms across a short window (common with lead-generation "students" filling forms for research assignments or content-download farming), and generic or placeholder text in open free-text fields like "test" or "asdf." Subtracting points for these signals, rather than only adding points for positive ones, keeps the model from over-crediting a lead that happens to check a few fit boxes on paper but carries clear noise signals underneath.
The fastest way to calibrate a scoring model in its first few months is a structured monthly conversation with the sales team reviewing a sample of both accepted and rejected leads together — not a one-off kickoff meeting, but a recurring habit. I ask sales to walk through five recently rejected leads and five recently accepted-but-poor-quality leads each month, and look for patterns in what the score missed: was there a firmographic signal the model was not capturing, a behavioral pattern that should have counted for more, or a source that was consistently producing leads that scored well but converted poorly regardless of score. This feedback loop, sustained for even three or four months, typically reshapes the scoring weights meaningfully and is what actually earns sales' trust in the number — not a polished scoring methodology document that sales never reads, but a visible process where their feedback demonstrably changes what marketing does next.
Back to all posts