Stop using bad as a catch-all
Start with an agreed definition of a valid inquiry and a qualified opportunity. Valid means there is a real, relevant request you can investigate. Qualified means it meets your specific service, geography, timing or other legitimate buying criteria. Interested but not ready is a separate state, not automatically worthless.
Inspect what happened after submission
In a public LinkedIn post, Jon Loomer argues that follow-up, deliverability and process can shape what advertisers experience as lead quality. David Dinsmore separately describes the problem of sending people with different intent straight into the same sales motion. These are practitioner observations, not universal diagnoses. They suggest a better first question: what actually happened to this request? Jon Loomer, David Dinsmore.
Review a manageable sample of recent inquiries across campaigns and outcomes. Use actual records rather than a salesperson's general impression. Capture the submission time, routing result, first attempted response, successful contact, buyer expectation, stated need and next action. Respect recording permissions and avoid collecting information that is not needed for the review.
An automatic acknowledgement is not the same as a useful response. A task being created is not the same as someone completing it. Look for evidence of the next step, not merely evidence that the software ran.
Use a cause-based classification
Classify each inquiry into an actionable reason. Spam needs validation or filtering. Wrong service or location may require changes to targeting and copy. A real prospect who misunderstood the offer may point to a message mismatch. An unanswered request points to ownership or routing. A lost deal on price deserves a separate commercial review.
Illustrative example: an ad offers a free workflow assessment, but the follow-up immediately asks the prospect to commit to an ongoing retainer. The person may disengage because the promised next step changed, not because they were a bad lead. This is a hypothetical scenario, not a reported client result.
Ask sales and marketing to classify the same small sample independently, then compare disagreements. If their definitions differ, a new dashboard will simply automate the disagreement. Resolve the definitions before changing campaign optimization or adding more qualification fields.
Add useful friction, not an obstacle course
A public discussion by Prasanna Shrestha frames lead forms as a reflection of the intent attracted by the message. The useful takeaway is to make the invitation explicit before asking for more data. Read the discussion.
Explain the next step, who it is for and what the prospect should prepare. Ask only questions that affect routing, eligibility or the conversation. If budget is required for a particular service, explain why. Do not use a long questionnaire to compensate for vague positioning.
Test the effect on qualified conversations, not form completion alone. A shorter form might increase volume but overload the team. A longer one might discourage valuable buyers as well as unsuitable requests. The right amount of friction is a business-specific hypothesis to test, not a universal conversion trick.
Close the feedback loop
Record meaningful outcomes in the CRM and make them available to the people managing acquisition. Google documents ways to import offline conversion outcomes and use enhanced conversions for leads. That gives a technical route for better feedback when the underlying records and permissions are sound. Google Ads guidance.
Before automating that exchange, confirm the event definitions, identifiers, deduplication and treatment of cancellations. An AI system can summarize a call for review or flag missing data, but a model-generated label should not silently become a verified revenue event. Put ambiguous or consequential classifications in a human review queue.
The handoff audit checklist
Your next campaign review should produce an owner and a repair for each problem, not another debate over which department is responsible.
- Spam and duplicates are separated from genuine inquiries.
- Wrong-fit requests have specific service, location or timing reasons.
- The ad promise, landing page and first sales conversation describe the same next step.
- Routing, response attempts and successful contact are visible in the record.
- A person owns unanswered requests and automation exceptions.
- Accepted opportunities and closed outcomes feed back into acquisition reporting.
When the ads really do need changing
If the audited requests consistently come from people you cannot serve, change the audience, search terms, offer framing or placement strategy. If suitable buyers arrive but the process breaks afterward, repair that process first. Often both need attention. The goal is not to defend the ads or blame sales. It is to identify the step that prevents a relevant request from becoming a useful conversation.
