Your launch team commissions new product demonstrations from a production partner in China. The monthly email report then shows fewer clicks, and the next filming allocation comes under review. Before cancelling the shoot, ask one question: did the reporting settings change between the two periods?
That hypothetical situation illustrates a procurement problem, not an ONCE client result. A content supplier can be held accountable for performance only when the buyer and supplier understand which measurement they are comparing.
What the platform documentation actually establishes
Mailchimp says its bot filtering identifies most automated activity and hides it in reports without deleting the underlying account data. Settings apply across reporting, rather than to individual reports. Its documented historical coverage for filtered email clicks starts on August 10, 2025. That date describes data availability, not a product launch or a rule for every email provider. See About Bot Activity and Bot Filtering, checked October 31, 2026.
Filtered clicks should not be relabelled “verified human clicks.” Detection is not a guarantee of perfect classification. Nor should different automated mechanisms be conflated: Mailchimp explains that Apple Mail Privacy Protection can preload tracking pixels and distort opens. That is not evidence that MPP clicks every link. See its MPP FAQ.
Restate the baseline before judging the asset
ONCE proposes a measurement record attached to each budget review. Capture campaign scope, send date, observation window, delivered-message denominator, metric definition, filtering settings and export time. Note changes to tracking or delivery configuration. This is a proposed working method, not a Mailchimp feature.
If both periods can be re-exported under the same settings, do that first. Retain the original exports alongside the restated figures and explain the difference. Because settings can affect account-wide reports, coordinate with the reporting owner before changing them.
If the older period cannot be restated, mark the comparison as unavailable and establish a new baseline. Do not apply a universal percentage reduction to old results. A broken comparison is a reason to improve measurement, not evidence that the creative became worse.
A small example changes the spending question
Assume two campaigns have identical delivery counts and observation windows. An old reporting view shows 120 clicks for the first and 90 for the second. Under consistent filtering, they show 70 and 75. These invented numbers demonstrate the arithmetic; they are not a test, benchmark or typical bot share.
The apparent decrease of 30 becomes an increase of five. That does not establish that the second video caused an improvement. Audience, offer, timing and other differences still matter. Comparing filtered and unfiltered reports is not a randomized creative experiment.
Use the result to choose the next purchase: repair an incomparable baseline; run a bounded test where comparable evidence remains weak; consider more production when comparable results and meaningful downstream actions support the decision. There is no general click-rate threshold that automatically justifies a reshoot or cancellation.
A website counter is not automatically the truth
Mailchimp also warns that security scanning can generate link events, while its own automated link validation may be counted by external tracking tools. A larger website count therefore does not settle a disagreement with the email report. See Troubleshooting Click Tracking.
For teams turning clicks into sales follow-up tasks, review filtering and subsequent confirmed actions before treating an isolated event as intent. A direct reply or a valid brief adds useful context, but neither guarantees a qualified lead. Keep unknown identity matches unknown; do not manufacture a complete customer journey from unrelated records.
Scope measurement as part of the retainer
A video marketing engagement should distinguish asset production, campaign operations and reporting responsibility. If monthly analysis is included, specify the metric dictionary, configuration-change log, restatement scope and explanation of anomalies. Filming fees do not inherently include administration of the email platform.
Where engagement affects a performance bonus, agree in advance on the report, filtering setting, observation window and treatment of later definition changes. Historical data that cannot be restated should remain visibly incomparable. Commercial adjustments require agreement, rather than an improvised formula after results arrive.
For a brand managing a China-based production relationship, this makes the next brief more useful: bring the original report, the configuration history and the content versions. Then decide whether the work required is reporting, testing, editing or new filming. Over time, an explainable decision history can become part of the value of managed content services. That is ONCE analysis, not a revenue forecast.
Source scope: the three official help pages above were checked October 31, 2026; no publication date is inferred. This is evergreen procurement research. Platform documentation is not evidence of industry-wide adoption, and the numerical example is hypothetical.
