Cheap Leads, Empty Pipeline? Build a Lead-Quality Scorecard

Your sales manager opens the CRM on Monday and finds plenty of new enquiries, but almost nothing worth adding to the forecast. Marketing reports a lower cost per lead. Sales reports wrong numbers, vague requirements and people who stop answering. Both teams may be describing the same campaign accurately—and still be measuring the wrong thing.

The useful answer is not to declare every unanswered enquiry poor quality. Build a lead-quality scorecard that separates contactability, customer fit, buying need and sales readiness. Then connect it to clear CRM dispositions and a denominator that cannot quietly shrink when results look uncomfortable.

For an Indian business selling software, equipment or professional services, this creates a practical shared question: Which enquiries became verified sales opportunities, which did not, and what evidence explains the difference? Start there before changing campaigns or buying another reporting tool.

Start with the disagreement, not another dashboard

Take a recent, manageable batch of enquiries and review the underlying records with marketing and sales. Read the original requirement, inspect the contact history and compare the salesperson’s conclusion with the recorded evidence. Do not begin with an average score.

Look for distinct failures. A submission may be automated spam, a genuine person outside your service scope, a suitable buyer contacted too late, or a qualified prospect whose purchasing window is distant. These are different operational problems, even if none has produced revenue yet.

Check whether routing worked. Did the enquiry reach the correct salesperson? Was the requested product visible? Did a reply land in an unmonitored inbox? An apparent quality problem can actually be a handling problem.

If your acquisition strategy also needs attention, the broader B2B lead-generation guide for India provides adjacent context. Keep this exercise narrower: define what happened after an enquiry arrived, before debating where the next enquiry should come from.

Write the qualification contract before assigning points

A qualified lead should mean the same thing to the campaign owner, salesperson and business head. Write a short definition based on observable buying conditions, not enthusiasm or a subjective impression of the prospect.

For example, your working definition might require a genuine business requirement, a serviceable use case, an identifiable buying role and agreement on a relevant next conversation. Adapt those conditions to your actual sales process.

Separate essential gates from helpful signals. An impossible delivery requirement may be a genuine disqualifier. A missing budget answer may simply be information still to collect. A prospect researching for a manager can have a legitimate buying role without final approval authority.

Define what qualification does not mean. It is not a promise to purchase, an accepted quotation or recognised revenue. Later stages need their own definitions.

Record who can change the definition and when a change takes effect. Otherwise, teams can improve reported quality merely by relaxing the criteria. Preserve the definition used for each reporting cohort so historical comparisons remain interpretable.

Give every enquiry a disposition that explains its state

Use a controlled primary disposition rather than allowing salespeople to invent labels. A workable sequence is: new, contact in progress, connected and assessing, qualified, nurture, disqualified, and closed outcome. Keep duplicate and invalid records as explicit administrative categories.

Each label needs an entry rule. “Connected and assessing” means a genuine exchange occurred but qualification evidence is incomplete. “Nurture” means the requirement is credible but not ready for an active opportunity. “Disqualified” requires a documented business reason.

Do not let “not interested” explain everything. Distinguish no relevant requirement, unsupported use case, unavailable service geography, and an explicit decision not to proceed. Record the prospect’s meaning rather than the salesperson’s frustration.

Make primary dispositions mutually exclusive, while storing supporting facts separately. A qualified lead can also have a demo booked; those should not compete as alternative statuses. Similarly, attempt count belongs in an activity field, not inside a status name.

Add a status date, owner and reason code. Keep a brief evidence note for consequential changes. “Asked for maintenance support; we only supply new equipment” is more actionable than “bad lead”. It also gives marketing something specific to check against the original promise.

Use a small scorecard with visible unknowns

The scorecard should make evidence comparable, not disguise judgement as mathematics. Keep its dimensions few enough that a salesperson can complete them during normal work.

The following is an illustrative scoring design, not a benchmark. Each dimension receives zero, one or two points when evidence exists. The possible maximum is four dimensions multiplied by two points, giving eight points. Mark missing evidence as “unknown”, not zero.

Dimension	Zero: evidenced weakness	One: partial evidence	Two: strong evidence
Contactability	Contact details verified unusable	Genuine exchange, identity unclear	Relevant contact identity confirmed
Customer fit	Confirmed outside service scope	Fit depends on clarification	Use case and delivery scope fit
Buying need	Confirms no relevant requirement	Exploring a relevant problem	Specific requirement described
Sales readiness	Declines any next step	Future review agreed	Relevant next action agreed

Do not set a universal passing score. In this illustrative design, qualification could require confirmed fit and need, plus an agreed next action. The total then helps prioritisation; it does not override essential gates.

Store unknown fields visibly and show assessment completeness beside the score. A partially assessed record should not appear directly comparable with a fully assessed record.

Also avoid double-counting correlated evidence. Downloading a brochure and opening its follow-up email do not independently prove buying need. Where possible, score the substance of the exchange rather than accumulated activity. This makes the score slower to complete but harder to inflate.

Lock the denominator before celebrating quality

A qualified-lead rate is meaningless without a clear denominator. A useful headline measure is qualified unique enquiries divided by eligible unique enquiries in the same intake cohort.

Define “eligible” narrowly. For this process, it means a genuine enquiry retained after documented duplicate and invalid-record exclusions. It does not mean “already suitable for our business”. Genuine enquiries outside your service scope remain in the denominator because they reveal a mismatch worth measuring.

Unreachable genuine enquiries also remain. Otherwise, difficult records disappear and the rate improves without the business receiving better demand. Track their unresolved status separately.

Publish a reconciliation: raw submissions, duplicates removed, invalid submissions removed, eligible unique enquiries, qualified enquiries and unresolved enquiries. State how you identify duplicates, including how separate requirements from the same organisation are handled.

Keep intake cohorts separate from event-period totals. Leads received this month may qualify later; leads qualifying this month may have arrived earlier. Both views are useful, but combining them produces a misleading rate.

Report at a declared assessment cutoff and allow later cohort updates. An immature cohort needs an unresolved share, not an early verdict. If you also show qualification among assessed records, label that as a diagnostic rate alongside—not instead of—the headline rate.

Make contact attempts observable and bounded

“No answer” describes an interaction attempt. It does not establish lack of need, poor fit or fraudulent intent. Require the CRM to capture the attempt date, method, result and next action.

Agree an appropriate contact sequence for your business, taking account of the enquiry context, prospect preference and normal operating hours. Do not turn a suggested attempt count into an industry standard. Different buying processes need different follow-up windows.

Use distinct outcomes for unanswered calls, undelivered messages, explicit refusals and successful conversations. A message sent is not necessarily a message delivered; a delivered message is not a conversation.

Set a stopping rule. Stop when the person declines further contact, when details are demonstrably unusable, or when the agreed sequence ends without engagement. Avoid repeated attempts through new routes simply to complete a target.

At that point, mark the record “unreachable after completed sequence” rather than “spam”. It can remain unqualified without being treated as invalid.

Review response handling alongside lead quality. Compare records with completed sequences against those abandoned early, and inspect whether assignment delays or missing notes explain differences. Consult appropriate privacy or compliance professionals when designing contact permissions, retention and access rules; the scorecard itself does not resolve those obligations.

Work through a hypothetical cohort without hiding losses

Hypothetical example: an industrial supplier receives 120 submissions and spends ₹60,000 on the activity being evaluated. These figures illustrate arithmetic only; they are not Adservex results or market benchmarks.

The team identifies 10 duplicate submissions and 10 invalid submissions supported by documented evidence. Eligible unique enquiries therefore equal 120 − 10 − 10 = 100.

At the agreed assessment cutoff, those 100 records comprise 25 qualified enquiries, 35 confirmed disqualifications, 20 credible nurture records and 20 unresolved enquiries. The reconciliation is 25 + 35 + 20 + 20 = 100.

The headline qualified-lead rate is 25 ÷ 100 = 25%. Qualification among resolved records is 25 ÷ 80 = 31.25%, because 100 − 20 unresolved records = 80. That second rate is useful for diagnosis but would overstate coverage if presented alone.

Cost per raw submission is ₹60,000 ÷ 120 = ₹500. Cost per qualified lead is ₹60,000 ÷ 25 = ₹2,400. The latter answers a more demanding question, but still says nothing about profitability without subsequent opportunity, win and margin evidence.

Suppose later follow-up qualifies 5 of the unresolved records. The updated cohort rate becomes 30 ÷ 100 = 30%, and cost per qualified lead becomes ₹60,000 ÷ 30 = ₹2,000. Preserve the earlier snapshot and show the updated assessment date. Progress should be visible, not mistaken for a reporting error.

Reject spam through evidence, not assumptions about people

Create a narrow invalid-record policy. Appropriate evidence might include meaningless automated payloads, demonstrably fabricated contact details, or submissions containing no enquiry and only unrelated promotional content. Document why the record meets the rule.

Do not use someone’s name, language, accent, gender, religion or perceived social background as a proxy for quality. Nor should a personal email address or an unfamiliar company name automatically trigger rejection.

Location can be relevant when a business genuinely cannot deliver there. Record that as a serviceability disqualification for a genuine enquiry, not as spam. The distinction protects both measurement accuracy and respectful treatment.

Likewise, limited English or a short message may require clarification rather than exclusion. Evaluate the requirement using the same business criteria wherever possible.

Send uncertain invalid classifications for review. Periodically inspect excluded records, especially when an automated rule suddenly rejects more submissions. Preserve enough evidence to audit the decision while avoiding unnecessary personal-data collection. Automation can accelerate a defensible rule; it cannot make a weak rule fair or accurate.

Turn sales feedback into a calibration routine

Run a recurring review around disagreements, not just totals. Include qualified records, borderline cases, genuine disqualifications and invalid exclusions. Ask marketing and sales to classify the same examples independently before discussing them.

When judgements differ, locate the missing rule. Does “budget unknown” mean incomplete assessment or disqualification? Does a requested callback count as readiness, or must the buyer agree to discuss a defined requirement? Update the rubric rather than criticising individual judgement.

Capture feedback in structured reason codes with a short explanatory note. Keep the code list stable enough for comparison, but retire vague categories when they stop producing useful action.

Assign an owner to each pattern. Unsupported requirements may call for clearer messaging. Incomplete assessments may call for better discovery prompts. Slow assignment may require an operational fix, not a campaign change.

For founder reporting, place these measures inside a weekly marketing dashboard without reducing them to a single traffic-light score. Show the denominator, unresolved share and principal rejection reasons beside cost per qualified lead. That preserves the connection between a number and the work needed to improve it.

Copy this worksheet before buying automation

Use this worksheet to create your initial operating agreement. Complete it with sales and marketing together; unresolved decisions are more valuable than prematurely polished reporting.

• Business scope: What requirements can we genuinely serve, and which restrictions are essential?
• Qualified definition: What verified evidence must exist before an enquiry is qualified?
• Intake cohort: Which arrival period and assessment cutoff will this report use?
• Eligible denominator: Which records remain, and what exact evidence permits exclusion?
• Duplicate rule: How do we distinguish repeat submissions from separate buying requirements?
• Disposition dictionary: What allows entry into each status, and who owns updates?
• Scorecard: Which dimensions are scored, which are gates, and how are unknowns displayed?
• Contact sequence: What methods, timing, stopping rules and outcome fields apply?
• Evidence note: What short explanation must accompany qualification, rejection or exclusion?
• Review process: Who checks disputed decisions and audits invalid records?
• Reporting reconciliation: Can raw submissions be traced through exclusions to every eligible status?
• Version control: When did definitions change, and which cohorts use each version?

Only then assess tooling. Your existing CRM may already support the required fields, validation and history. A spreadsheet can test the definitions, although access control and concurrent editing become practical limits.

Google Ads provides official guidance on qualified leads and converted leads. Use it when aligning applicable advertising conversion categories with downstream outcomes. That platform guidance does not prescribe this article’s scorecard, denominator or contact process. Resolve those internally before configuring measurement or imports.

If you need help reviewing the measurement brief, contact Adservex with your current definitions and anonymised examples rather than an unnecessary export of identifiable prospect data.

Practical questions before you roll it out

Should an unreachable lead receive a zero score?

Not for fit, need or readiness. Those dimensions remain unknown unless evidence supports a judgement. Record the failed contact sequence separately. Treating silence as a zero across every dimension confuses missing information with confirmed weakness and can conceal a sales-handling problem.

Should sales be allowed to override the score?

Yes, with a recorded reason and reviewable evidence. A score simplifies reality and may miss a legitimate exception. Preserve the original component scores, show the override separately and inspect recurring overrides. Frequent exceptions suggest the rubric needs repair, not that every salesperson needs tighter restrictions.

What if the qualified-lead rate falls after launch?

First check whether the denominator became more complete, exclusions became stricter or previously hidden unresolved records appeared. A lower reported rate can reflect more honest measurement. Compare cohorts using consistent definitions and similar assessment maturity before concluding that actual enquiry quality has deteriorated.

Can cost per qualified lead decide the budget alone?

No. It is a useful intermediate measure, not a complete commercial outcome. Consider qualification consistency, opportunity progression, sales effort and eventual contribution. A less expensive qualified lead can still be costly to serve or unlikely to close. Keep downstream outcomes linked without rewriting earlier qualification decisions to match hindsight.

Source notes: Google Ads Help: qualified leads and converted leads is the official reference for the relevant Google Ads guidance. The operational definitions, worksheet, illustrative scoring design and hypothetical arithmetic above are editorial recommendations, not requirements or performance claims from Google.