Why Are My Ad Leads the Wrong Fit?

“The leads are bad” is too broad to guide a useful change. It can mean the wrong service, a missed callback, a price objection or a duplicate record. Those problems need different fixes. Create a short codebook that describes what your team actually learned.

Write codes that point to different actions

Start with a small set your employees can use consistently. Give every code a definition, the evidence needed and an owner for reviewing it. Add an “unknown” option so an agent or employee never has to invent a reason.

• Code: Service mismatch; Use it when: The prospect explicitly requests work you do not offer; Review next: Ad wording and service description

• Code: Coverage mismatch; Use it when: The requested job address is outside supported coverage; Review next: Campaign geography and coverage explanation

• Code: Timing mismatch; Use it when: The requested deadline cannot be met; Review next: Availability promises and intake timing

• Code: Unreached; Use it when: The agreed contact process produced no conversation; Review next: Contact details and response process

• Code: Duplicate inquiry; Use it when: The same known person repeats an existing request; Review next: Counting and record matching

• Code: Unknown; Use it when: There is not enough evidence; Review next: A specific follow-up task

Do not confuse an objection with a fact

“Too expensive” needs context. Did the customer see your minimum charge, receive a full quote, or simply assume a price? Record the customer’s actual statement and the point at which it happened. A campaign attracting people who cannot meet a stated minimum differs from a salesperson explaining a complex quote poorly.

Keep budget unknown separate from budget unsuitable. Avoid labels about attitude, intelligence or presumed personal circumstances. They are difficult to verify and do not identify a marketing action. For sensitive categories, keep the process focused on information appropriate to the service and have your team review the rules.

Give the AI agent a narrow recording job

An agent can suggest a code from the conversation and attach the sentence that supports it. Have a person confirm uncertain cases. The useful instruction is: “Choose service mismatch only when the requested service clearly falls outside the approved list. Otherwise choose unknown and explain the missing information.”

Require one primary reason and allow a secondary note. If every lead receives six labels, your report becomes hard to interpret. Preserve the original inquiry so a reviewer can tell whether the code accurately represents the conversation.

Use a weekly sample to improve the definitions

Review a small selection of accepted, rejected and unresolved leads together. Ask whether two reviewers would choose the same code. When they disagree, fix the definition before changing a campaign. If an entire category is growing, inspect examples rather than treating the count as an explanation.

An illustrative review might find that ten “bad” leads include four repeat requests, three missed callbacks and three genuine service mismatches. Only the last group directly supports changing which service the advertisement describes. The other groups call for measurement or follow-up repairs.

Google distinguishes qualified leads from later converted leads. Your internal codebook should make those stages understandable before you share conversion feedback. For related work, see using phone inquiries as marketing feedback and assigning ownership to incoming leads.

For help connecting useful qualification rules to your campaigns, Build Your AI Agents. Escencion handles setup and ongoing management of the agents you choose.