46 Amazon expects reported advertising impressions across its store to fall by an average of 63% when it changes how those impressions are counted on January 1, 2027. The adjustment could substantially alter the numbers brands see in their campaign dashboards, even where their advertising activity and commercial results remain broadly unchanged. Under the existing approach, an impression can be recorded when an advertisement begins downloading to a shopper’s device. The revised method will count it only after the advertisement appears on the screen. The change covers display, video, and audio advertising on Amazon’s retail properties, including its homepage, product pages, search results, and order pages. The estimated 63% reduction is an average, not a forecast for every campaign. Results will vary according to where advertisements appear. Placements lower down a page may be particularly affected because an advertisement can begin loading without the shopper ever scrolling far enough to see it. The financial effect on reporting could look dramatic. If a campaign’s spending stays constant while its counted impressions fall by 63%, its reported cost per thousand impressions rises by approximately 170%. Its click-through rate could also increase if the number of clicks remains unchanged. Neither movement, by itself, proves that advertising has become more expensive or more effective. Amazon says its separately reported, accredited viewable-impression metrics and attribution measurements will not change with the transition. Historical reports covering dates before January 1 will also remain intact. The revision brings a long-standing problem in digital advertising into sharper focus: a recorded impression has not always meant that an advertisement actually appeared before a consumer. The new method narrows that gap, although appearing on a screen still does not establish that somebody noticed an advertisement or acted because of it. For advertisers, the immediate difficulty will be comparing performance across the year-end boundary. Automated reports, campaign thresholds, and models that use impression volume may need recalibration. Contracts or internal targets built around historical counts could also require attention. The more useful comparison will be between advertising investment and outcomes such as incremental sales, customer acquisition, and profitability. Amazon’s new figures may be smaller, but they could offer a more defensible account of what advertisers have actually placed on shoppers’ screens. You Might Be Interested In Starbucks tests AI-built software to replace microsoft and IBM systems Meta set to overtake Google in digital ad revenue for the first time How 2026 US Open brand activations are becoming a test lab for experiential marketing Meta AI creative tools spark marketer concerns over control and consistency AI spending nears $760 billion as big techs test the economics of scale Meta Passes Digital Tax Costs to Advertisers With New Location-Based Ad Fees