Skip to main content
    European Search Awards 2026 · Best Small PPC Agency

    Primary research · Withdrawn pending review

    The 2026 Ecommerce POAS Benchmark isunder review and not available

    We have withdrawn the v1 conclusions of this study. They should not be used for planning, budgeting or investment decisions, and they should not be cited.

    Withdrawal notice

    The published v1 figures, and the cohort descriptions that framed them, could not be reconciled. The conclusions, the downloadable version and the structured data describing this as a published dataset have been withdrawn pending source validation, along with every figure from the study that appeared elsewhere on this site.

    Nothing from this study should be used as a benchmark, or for planning, budgeting or investment decisions. Two broader claims built on the same cohort, a universal healthy POAS level and Performance Max causing branded-search cannibalisation, have been removed as unsupportable.

    The intended method is described in the questions below, and named client results are on the case studies.

    Version history

    • 22 May 2026: v1 published, with a median contribution lift, a cohort description and a downloadable version.
    • 8 September 2026: v1 conclusions withdrawn pending source validation. Download and dataset markup removed, and downstream citations taken out of site copy.

    What you can look at instead

    Named client results, each with its metric, period and definition attached. They are individual outcomes rather than an average, which is a smaller claim, and a clearer one.

    Questions about this withdrawal

    What is the 2026 POAS Benchmark?

    It was intended as JudeLuxe primary research measuring how Profit on Ad Spend changes between agency handover and twelve months of POAS-first Google Ads management, across a small cohort of UK DTC ecommerce brands.

    Why are the figures no longer published?

    During an internal review the published headline findings did not reconcile with the underlying per-brand records. Rather than restate the study around a revised number, we have withdrawn the v1 conclusions in full while the source data is reconciled.

    Can I cite or rely on the v1 figures?

    No. The v1 figures should not be cited, quoted or used for planning or investment decisions. Anything previously downloaded or referenced from this page should be treated as withdrawn.

    Will the benchmark be republished?

    Only if the reconciled dataset supports a defensible result. We will not publish a benchmark that a reader cannot check, and we are not committing to a republication date.

    What evidence can I look at in the meantime?

    Named client results on our case studies, each with its metric, period and definition stated. Those are individual outcomes rather than a portfolio or cohort average.