Online Video Advertising for Ecommerce Brands

Where OLV actually runs, and why it's different from streaming

OLV serves inside an active browsing session — Amazon's own shopping and browsing experience, or third-party apps and websites reached through Amazon DSP's publisher network. Unlike full-screen streaming video, it's frequently clickable, often skippable, and watched on a personal device rather than a shared living-room screen. That makes it structurally closer to display advertising than to streaming TV, even though both get called "video" — the viewer is doing something else, has an easy way to disengage, and the format has to earn attention in a genuinely different way than non-skippable CTV inventory does.

For an ecommerce brand specifically, that clickability is the main practical advantage over streaming TV. A viewer who's interested enough to click lands directly on a product or brand page, closing the gap between exposure and consideration in a way a non-skippable CTV ad, with no click at all, structurally cannot. That doesn't make OLV superior to streaming — it makes it a different tool for a different stage of the funnel.

The click-trap, with real numbers

A specific caution is worth stating plainly because it's easy to miss: a $0.41 CPC figure circulates widely in this category, and it's a real, accurate number for online-video specifically — not a blended DSP figure. Across 30 advertisers in our own book in July 2026, the account-wide blended CPC, including streaming and every other format, was $1.42. That gap matters because a cheap OLV click can look like an efficient one while converting nowhere near as well as a more expensive click elsewhere in the account — cost per click is not the same measurement as cost per outcome, and OLV is exactly the format where that distinction gets tested most often.

A worked example

At $8,000 a month and a representative $15 CPM, OLV buys roughly 533,000 impressions — enough scale for a genuinely useful mid-funnel discovery test. Set the campaign's read around both click-through and completion rate together, not click-through alone: a high CTR with a low completion rate suggests the video is earning a click without earning attention, which tends to convert worse downstream than a lower-CTR, higher-completion creative that holds the viewer through the message before any click happens at all.

Compare cost per click against cost per outcome at the end of the first month, not during it. A line item that looks expensive on CPC relative to another line item in the account can still be the better spend if its downstream conversion rate is high enough — the CPC comparison alone, made mid-flight, is exactly the kind of premature read that leads to cutting a genuinely working campaign.

What to do when OLV clicks aren't converting

If CTR looks strong but downstream conversion is weak, check the destination page match first — a click from a video ad that promised one thing landing on a generic product page is a common, fixable mismatch. If cost per click looks unusually low across the board, verify it's a genuine intent-driven click and not a low-friction accidental tap; our own DSP book shows this pattern clearly in an adjacent format, where third-party mobile in-app video generated 48.3% of all clicks recorded but only 3.5% of the sales, at a $26.94 cost per acquisition against a $6.83 portfolio-wide figure — cheap clicks are consistently the most expensive thing to optimize toward once you look past the CPC line.

The common mistake

The mistake is optimizing an OLV campaign purely on cost-per-click, chasing the cheapest available clicks without checking whether they convert at a rate that makes them worth having. A campaign can look highly efficient on CPC and be quietly burning budget on low-intent taps that a completion-rate or downstream-conversion check would have caught immediately. reMKTR reads every OLV line item on completion rate and downstream conversion alongside CPC, specifically because CPC alone has repeatedly proven to be the metric most likely to mislead on this exact format.