The cost-per-mille (CPM) has been the foundational currency of advertising for over a century. Pay for a thousand impressions. Hope some of those impressions translate to awareness. Measure success through surveys, brand lift studies, and correlation with sales data. The model worked for print, adapted to television, and carried over to digital display and streaming video.

But in a live sports streaming environment where viewers can tap a prediction, click a shoppable overlay, build a bet slip, vote in a poll, and share a reaction — all without leaving the video stream — measuring success by "was the ad viewable for two seconds" is like judging a restaurant by whether the door was open.

Interactive ad formats are forcing a fundamental shift in how streaming advertising is bought, measured, and valued. The metric that captures this shift is Cost Per Engagement (CPE) — and the platforms, advertisers, and ad tech companies that adopt it first will have a structural advantage over those still trading on impressions alone.

Why CPM Worked (and Where It Fails)

CPM was an effective metric for an era when advertising was a broadcast activity. A brand bought a million impressions on a network, and the assumption was that some percentage of those impressions would translate to awareness, consideration, and eventually purchase. The model's strength was simplicity — everyone agreed on what an impression was (more or less), and the pricing mechanism was transparent.

CPM still works for brand awareness campaigns where the goal is broad reach. But in streaming environments with interactive ad capabilities, CPM has three critical blind spots:

It doesn't distinguish passive from active. A viewer who sees a 15-second pre-roll ad while checking their phone is counted the same as a viewer who actively clicks a shoppable overlay, adds a product to their cart, and returns to watching. Both are "impressions" under CPM. The gulf in value between them is enormous.

It can't capture interaction depth. Interactive ad formats generate a spectrum of engagement — from a glance at an overlay to a multi-step interaction (view odds, build a bet slip, tap to transfer to sportsbook). CPM flattens this entire spectrum into a single binary: seen or not seen.

It incentivizes the wrong behavior. When revenue is measured by impressions, platforms are incentivized to maximize the number of ads served, which leads to higher ad loads, lower completion rates, viewer frustration, and eventually subscriber churn. This is the classic "race to the bottom" that has plagued ad-supported digital media.

What CPE Measures

Cost Per Engagement (CPE) shifts the currency from "was the ad delivered" to "did the viewer do something with the ad." The specific definition of "engagement" varies by format and campaign objective, but common engagement actions include:

Tapping or clicking an interactive element — a product link, a CTA button, an expandable unit, a poll response, or a prediction prompt within an ad experience.

Spending time with an ad unit beyond the minimum — actively hovering over an expandable overlay, scrolling through a product carousel, or watching an extended brand experience that the viewer opted into.

Completing a conversion step — adding a product to a cart, signing up for a newsletter, downloading an app, or transferring to a partner platform (like a sportsbook) from an in-stream prompt.

Sharing or social interaction — sharing a sponsored prediction result with a watch party, posting a reaction to a branded moment, or inviting friends to a sponsored challenge.

The key distinction is that engagement requires volitional action from the viewer. It can't be faked by autoplay or inflated by viewability tricks. The viewer actively chose to interact with the ad, which is a qualitatively different signal than passive exposure.

The Economics of Engagement vs. Impressions

The shift from CPM to CPE restructures the economics of streaming advertising in several important ways:

For Advertisers: Pay for Signal, Not Noise

Under CPM, advertisers pay for every impression regardless of whether the viewer noticed, cared about, or acted on the ad. Under CPE, advertisers pay for demonstrated interest. A viewer who taps a shoppable overlay during a live game is sending an explicit signal of intent — that engagement is worth dramatically more than a passive video impression.

Performance data from interactive CTV campaigns consistently shows that engagement-based pricing delivers better ROAS (return on ad spend) than impression-based pricing, because the denominator is filtered for actual interest rather than raw exposure.

For Publishers: Premium Pricing for Premium Inventory

Interactive ad formats that generate measurable engagement can command CPE rates that translate to effective CPMs significantly higher than standard video. When a sponsored prediction prompt generates a 15-25% engagement rate (compared to the 0.5-2% click-through rate of standard display), the per-engagement price can be set at a level that delivers far more revenue per available impression.

This is particularly important for live sports, where inventory is finite (there are only so many games per season) and demand is high. Selling engagement rather than impressions captures more of the true value of live sports attention.

For Platforms: Alignment Between User Experience and Revenue

CPE pricing incentivizes platforms to create better ad experiences rather than more ad experiences. If revenue is tied to engagement, the platform is motivated to deliver ads that viewers actually want to interact with — contextually relevant, well-timed, non-disruptive, and genuinely useful.

This aligns the platform's revenue incentive with the viewer's experience for the first time. Under CPM, the platform makes more money by serving more ads, even if the viewer hates them. Under CPE, the platform makes more money by serving ads that the viewer chooses to engage with.

Interactive Formats That Drive Engagement

Not all ad formats generate meaningful engagement. The formats that perform best under CPE pricing share common characteristics: they're contextually relevant, non-disruptive, and offer genuine value to the viewer.

Sponsored Prediction Prompts

"Tonight's BMW Player Performance Pick: Will Steph Curry score 30+ points?" These prompts tap into the same psychology that drives prediction market engagement — fans want to express opinions about outcomes. When a brand sponsors the prompt, the brand becomes associated with the interactive moment rather than interrupting it.

Shoppable Overlays Triggered by Game Events

When a player scores wearing a specific brand's gear, an overlay surfaces the product with a one-tap purchase path. The engagement is contextually perfect — the viewer just saw the product in action and can act on impulse without leaving the stream.

Interactive Stat Experiences

"Presented by Gatorade: Track tonight's player hydration timeouts and performance correlation." Branded interactive stat experiences give fans deeper insight into the game while associating the brand with valuable content rather than interruptive advertising.

Expandable Brand Experiences

A small overlay invites the viewer to "Learn more about the new Ford F-150." If the viewer taps, the experience expands to a larger interactive unit with video, specs, and a configurator — while the game continues in a PiP window. The viewer opted in, making the extended exposure genuinely valuable.

Measurement Infrastructure for CPE

Shifting from CPM to CPE requires measurement infrastructure that can track, verify, and report on interactive ad engagement at scale.

Engagement event taxonomy. Advertisers and publishers need a shared vocabulary for engagement actions — what counts as an engagement, what's the minimum threshold, and how are different engagement types weighted. IAB's interactive ad measurement guidelines are evolving to address this, but the market hasn't yet converged on a single standard.

Real-time reporting. Unlike CPM-based campaigns where post-campaign reporting is sufficient, CPE campaigns benefit from real-time engagement data that enables mid-flight optimization. If a sponsored prediction prompt is generating low engagement during hockey games but high engagement during basketball, the campaign can be adjusted in real time.

Fraud prevention. Engagement-based metrics are inherently more fraud-resistant than impression-based metrics (it's harder to fake a multi-step interaction than a viewable impression), but bot traffic and accidental clicks still need to be filtered. Platforms need engagement quality scoring that distinguishes intentional interaction from accidental taps.

Cross-platform consistency. Engagement manifests differently across devices — a tap on mobile, a remote click on CTV, a mouse click on web. The measurement framework needs to normalize these different interaction modalities into consistent engagement metrics.

The Transition Period

The industry won't abandon CPM overnight. Brand awareness campaigns will continue to trade on impressions for the foreseeable future. But the transition is already underway for interactive formats — performance-focused advertisers are increasingly asking to buy engagement rather than impressions, and platforms with interactive capabilities are able to offer that option.

The hybrid model that's emerging — CPM for standard video ads, CPE for interactive formats — actually makes sense. Not every ad is interactive, and not every campaign goal requires engagement. The key is that platforms which offer interactive formats and CPE pricing give advertisers a choice, and advertisers with performance objectives will increasingly choose engagement over impressions.

For streaming platforms evaluating their ad technology roadmap, the ability to deliver and measure interactive ad engagement isn't just a feature — it's the foundation for a pricing model that captures the true value of live sports attention.