A placement audit cut retargeting CPA by 40% for a B2C marketplace

    A placement audit on the retargeting channel cut CPA (cost per acquisition) by 40% for a B2C marketplace, compared with the previous 3 months. To get there, we cross-checked clicks, spend and sales for each placement and excluded the ones that spent without converting.

    Profile

    B2C marketplace · Growth stage

    Problem

    For three months, the channel spent on placements that generated no conversions. The account average hid the waste.

    Solution

    Audit by placement, not by channel average: where the money goes in and where the click turns into a sale.

    What we did

    • We cross-checked clicks, spend and sales for every placement.
    • We excluded the placements that spent without converting.
    • We turned the process into a standard channel review, reused in later projects.

    Result

    -40% CPA compared with the previous 3 months.

    Key learnings

    • The account average hid the waste. The problem only shows up when you look placement by placement.
    • Click volume is not conversion. A placement can get plenty of clicks and zero sales.
    • The gain came from the method, not from a one-off clean-up. The audit became a routine.

    Do it yourself

    1. Export the placement, site or app report for the last 90 days and build a sheet with one row per placement and these columns: cost, clicks, conversions, CPA (cost ÷ conversions) and % of total spend. Sort by cost. If your campaign type doesn't show placements, that's the first problem to fix.

    2. Set the cut-off before you look at the data, for example: any placement that spent more than 2x the target CPA with zero conversions goes. Add the cut placements to an exclusion list and repeat the audit every month, because bad placements come back.

    3. Ask an LLM:

      I'll paste a placement report with cost, clicks and conversions. Calculate the CPA for each placement, flag the ones that spent more than 2x my target CPA of [TARGET CPA] without converting, tell me what share of total spend they represent and build the exclusion list ready to upload to the platform.