The gap between lead volume and pipeline
A familiar pattern in B2B: the campaign dashboard shows lead volume rising and cost per lead falling, while the sales team reports that very little is worth following up. Both observations are accurate. They are measuring different things.
Marketing is measuring form submissions. Sales is measuring opportunities. Without a shared definition of what qualifies as a real opportunity, and a system that records it, those two views never reconcile — and budget decisions get made on the easier number.
Lead qualification is the process that connects them. It is less a scoring algorithm than an agreement: what makes a lead worth a salesperson's time, who decides, and how that decision travels back to the campaigns that generated it.
If you cannot currently report cost per sales-qualified lead by campaign, the loop is not closed — and optimisation is running on form fills rather than pipeline.
MQL, SQL and SAL: agreeing the definitions
These terms are used loosely and that is the root of most marketing-sales friction. Write your own definitions down, with explicit criteria, and have both teams sign off.
| Stage | Definition | Who owns it |
|---|---|---|
| Lead | Any contact captured — form, call, chat, event list | Marketing |
| MQL | Meets the agreed fit criteria and has shown meaningful intent | Marketing |
| SAL | Accepted by sales as worth contacting, after review | Shared handover point |
| SQL | Contacted and confirmed to have a genuine need, authority and timeline | Sales |
| Opportunity | Active deal with a value and expected close date | Sales |
The SAL stage is frequently omitted and it is the most diagnostically useful of the five. A low MQL-to-SAL rate means marketing's criteria are too loose. A healthy SAL rate with a poor SQL rate means the criteria are right but the conversation or the follow-up is failing. Without the intermediate stage, both problems look identical.
Score on two axes: fit and intent
Single-score lead models collapse two different questions into one number and produce confusing results. Keep them separate.
Fit is whether this organisation matches who you can serve profitably: industry, company size, geography, budget range, technology or regulatory context. Fit is relatively stable and often available from the form or enrichment data.
Intent is whether they are actively looking now: pages viewed, pricing page visits, demo requests, email engagement, repeat visits, document downloads. Intent moves quickly and decays.
| High intent | Low intent | |
|---|---|---|
| Good fit | Contact immediately; this is your priority queue | Nurture with relevant content; revisit when behaviour changes |
| Poor fit | Review manually; may be a referral source or a smaller use case | Do not route to sales; keep in a low-frequency list or decline politely |
Keep the model simple at the start
A scoring model with forty rules is unmaintainable and nobody trusts it. Begin with five or six fit criteria and four or five intent signals, each with a weight. Review the model quarterly against actual closed-won data and adjust the weights that proved wrong.
Include negative scoring. Competitor domains, student or job-seeker enquiries, free email addresses where your business sells only to enterprises, and geographies you do not serve should all reduce the score rather than being filtered out silently — you want to see how much of your volume they represent.
What to ask on the form
Form length is a trade-off, not a principle. Shorter forms produce more leads; longer forms produce fewer and better ones. The right answer depends on whether your constraint is volume or sales capacity.
A practical approach: ask for the minimum that allows a qualification decision, and enrich the rest afterwards.
- Almost always worth asking: work email, company name, and one question that reveals the use case or problem.
- Worth asking when sales capacity is limited: company size, budget range or timeline — phrased as ranges rather than open fields.
- Rarely worth asking upfront: job title where it can be enriched, long free-text fields, and anything the salesperson will ask in the first two minutes anyway.
Add hidden fields to every form capturing UTM parameters, the click identifier, landing page and referrer. Without these, campaign-level qualification reporting is impossible later. Our guide to first-party tracking covers the capture mechanics.
One caution: a longer form reduces volume immediately and improves quality gradually. Judge the change on qualified leads after a full sales cycle, not on lead count after a fortnight.
Routing and response time
Qualification fails most often in the handover rather than the scoring. Three operational rules cover most of it.
Route automatically. Leads meeting the MQL threshold should create a task assigned to a named owner, with notification. A shared inbox that someone checks when they remember is where qualified leads die.
Set and measure a response-time target. Speed of first contact strongly influences whether a conversation happens at all, particularly for inbound demo and quote requests. Agree a target — same business hour for high-intent requests is a common standard — and report against it.
Define disposition codes. Every routed lead must be closed out with a reason: qualified, wrong fit, no budget, no timeline, unreachable, competitor, duplicate. These codes are the raw material for improving campaigns. A CRM full of leads in "new" status six weeks later makes qualification reporting impossible.
Close the loop back to the campaigns
This is the step that converts qualification from an internal process into a performance advantage.
Once the CRM records both the lead's source and its qualification outcome, two things become possible. First, you can report cost per SQL and cost per opportunity by campaign, keyword or audience — rather than cost per form fill. Second, you can import those outcomes back into the ad platforms as offline conversions, so automated bidding optimises towards leads that sales accepted rather than leads that merely submitted a form.
The effect is usually visible within a few weeks: volume often falls, cost per lead often rises, and qualified pipeline improves. Prepare stakeholders for that pattern in advance, because a rising cost per lead looks like a failure if nobody explained what changed. Our guide on cheap leads versus qualified leads is a useful briefing document for that conversation.
Working backwards to an allowable cost per lead
Qualification rates let you calculate what a lead is actually worth, rather than guessing at an acceptable cost per lead. Take a hypothetical B2B service business:
- Average deal value: ₹4,00,000
- Contribution margin: 40%, so ₹1,60,000 per deal
- Target acquisition cost: 25% of contribution margin, so ₹40,000 per win
- SQL to closed-won rate: 20%
- MQL to SQL rate: 25%
Working backwards:
| Stage | Conversion to next stage | Allowable cost |
|---|---|---|
| Closed-won deal | — | ₹40,000 |
| SQL | 20% close rate | ₹8,000 |
| MQL | 25% progress to SQL | ₹2,000 |
A ₹2,000 cost per MQL is therefore acceptable in this scenario, and ₹3,500 is not — regardless of how it compares to an industry benchmark. Equally, improving the MQL-to-SQL rate from 25% to 35% raises the allowable cost per MQL to ₹2,800, which is often easier than reducing media costs by the same proportion.
Run this calculation with your own numbers before the next budget discussion. It converts an argument about lead costs into an arithmetic question.
Common failure modes
- Scoring built without sales input. If sales did not help define the criteria, they will not trust the output and will revert to their own judgement.
- No disposition discipline. Leads left in an open status forever make every downstream report unusable.
- Thresholds never revisited. Criteria set at launch are rarely still correct a year later, after pricing, positioning or market changes.
- Nurture treated as a dumping ground. Good-fit, low-intent leads need genuinely relevant content on a sensible cadence, not a monthly newsletter.
- Optimising on MQLs alone. Without the SQL stage fed back, campaigns are tuned to produce whatever generates form fills most cheaply.
- Judging changes too early. In a business with a three-month sales cycle, qualification changes cannot be assessed after three weeks.
In summary
B2B lead qualification works when four things are in place: definitions both teams accept, a simple model that separates fit from intent, routing with a measured response time and mandatory disposition codes, and a feedback loop that returns qualification outcomes to the campaigns.
Start with the definitions and the disposition codes, because they cost nothing and make everything else measurable. Then calculate your allowable cost per MQL from real conversion rates. Once that number exists, campaign decisions stop being a debate about lead volume and start being a question of pipeline economics.
Questions B2B teams ask
What is the difference between an MQL and an SQL?
An MQL meets marketing's agreed criteria for fit and intent based on available data and behaviour. An SQL has been contacted by sales and confirmed to have a genuine need, appropriate authority and a realistic timeline. One is a prediction; the other is verified.
Should we make our forms longer to improve lead quality?
Only if sales capacity is the constraint. Longer forms reduce volume immediately and improve quality gradually, so judge the change on qualified leads over a full sales cycle. If volume is the constraint, keep the form short and qualify afterwards through enrichment and follow-up.
How many criteria should a lead scoring model have?
Start with roughly five or six fit criteria and four or five intent signals. Models with dozens of rules become impossible to maintain and lose the confidence of the sales team. Review the weights quarterly against closed-won data.
How quickly should we contact an inbound lead?
As quickly as the request's intent justifies. Demo and quote requests warrant same-business-hour contact. The specific target matters less than agreeing one, automating the routing and reporting performance against it.
How do we send qualification outcomes back to Google Ads or LinkedIn?
Capture the click identifier with the lead, store it on the CRM record, and use the platform's offline conversion import to upload the stage reached. Automated bidding then optimises towards qualified outcomes rather than form submissions.
What should happen to leads that do not qualify?
Record a disposition code rather than deleting them. Good-fit, low-intent contacts belong in a nurture track and can re-enter the pipeline when behaviour changes. Poor-fit contacts should be excluded from sales routing and, where relevant, added as audience exclusions.
How long before we can judge whether qualification changes worked?
At least one full sales cycle, and ideally two. Early indicators such as MQL-to-SAL acceptance rate and response time can be read sooner, but closed-won impact cannot be assessed until deals have had time to progress.