Compare search-partner lead quality separately from Google search traffic

Separate surface from customer outcome

Search campaigns can serve on Google's search surfaces and participating partner sites. Those placements can create different customer contexts, so a blended campaign CPC may conceal a meaningful quality difference. Begin by asking whether each network produces requests your business can serve. Cheap clicks are not automatically poor, and expensive clicks are not automatically valuable. The business outcome provides the comparison; the surface identifies where to investigate.

Source checks: About the Google Search Network · About Search Partner Network full placement reporting

Export a network view with stable definitions

Use a consistent campaign, date range, cost column, and conversion-action set. Segment network traffic and check whether the business can associate delivered inquiries with that classification reliably. Preserve unknown or unattributed outcomes instead of forcing them into a network. Compare qualified inquiries and booked work where available. If only platform actions are measurable, label the result as a configured-conversion comparison rather than a verified lead-quality analysis.

Inspect the current placement report appropriately

Google now documents a Search Partner Network placement report that lists participating sites and impression data under specified reporting conditions. Use it to inspect distribution and raise relevant questions. Do not rely on outdated advice that partner sites can never be seen. Equally, a site-level impression table does not automatically supply site-level booked jobs or explain every lead. Keep the network outcome report and placement-distribution view distinct.

Compare quality and cost together

Calculate spend per qualified inquiry and completed job for each reliably measured network. Examine spam, wrong-service requests, and follow-up outcomes as separate reasons. A partner segment with more raw actions may still produce less usable work, while another may reach customers the main segment misses. Check whether inconsistent tracking or a conversion-definition change could create the apparent gap. Avoid treating one isolated cheap click as representative of a whole network.

Choose an inclusion test with a review date

If the evidence supports a change, document whether to maintain inclusion, test exclusion, or collect more data. Check the current campaign-level network settings and supported controls. Keep other major edits stable where practical and allow for booking delay. A before-and-after test remains affected by demand and auction changes, so report it cautiously. Do not promise that disabling partners will mechanically transfer the same spend into better customers.

Review useful work after the decision

Compare qualified requests, completed jobs, spend, and the unknown-outcome share after the planned review period. Check whether total useful work declined even if a headline rate improved. Preserve the previous inclusion setting and rationale. The owner should receive a clear explanation of what the network evidence showed, which controls were changed, and what remains unresolved rather than a blanket claim that all partner traffic is good or bad.

Worked example: illustrative figures

For example, Google search traffic spends $400 and produces 8 qualified inquiries, or $50 each. Partner traffic spends $200 and produces 2 qualified inquiries, or $100 each, despite its cheaper clicks. The blended qualified-inquiry cost is $600 ÷ 10 = $60. The owner investigates partner request types and tracking before testing a setting change. These figures assume reliable network attribution and identical qualification; they do not establish site-level results from a placement-impression report or predict the effect of disabling partners.

Questions about this task