Leading Digital Publisher Unlocks 110% Increase in Prebid Revenue Contribution with Mile

Rohan Sharma
July 1, 2024

Mile’s AI-Powered Dynamic Flooring optimizes ad auctions in real-time, boosting the Prebid revenue contribution within weeks. The publisher is witnessing its Prebid revenue double while AI outperforms GAM CPMs by 6%.

In the competitive digital landscape, the publisher stands out for its diverse content offerings. Their audience primarily hails from the US, with significant international following, particularly in Canada, the UK, and Australia contributing over 56% of their total traffic.

The Problem

The group was navigating a significant challenge within its programmatic advertising ecosystem. It relied predominantly on Google Ad Manager (GAM) for 92% of its revenue. In contrast, its Prebid Stack, a solution offering the potential for increased competition and higher earnings, was significantly underperforming.

Prebid Stack contributed only 8% to 10% of the publisher's overall revenue. This underwhelming performance was further compounded by a stark difference in cost-per-mile (CPM) rates. Prebid Stack's CPMs were a concerning 60% lower compared to those generated through GAM.

All eggs in one basket

In programmatic advertising, a publisher's success hinges on a diversified and optimized ad stack. While Google Ad Manager (GAM) is undoubtedly a powerful tool, relying solely on it creates problems:

  • Programmatic advertising is an ecosystem offering numerous demand sources beyond GAM. By solely relying on GAM, the publisher misses out on potential bids from these alternative sources. Prebid Stack functions as a neutral exchange, allowing participation from a wider pool of advertisers, including those not integrated with GAM. This expanded reach drives up competition, increasing the likelihood of receiving the best bids for each impression and potentially increasing CPM rates.
  • Also, overdependence on a single source, like GAM, creates a vulnerability. If, for any reason, performance on GAM dips, the publisher's entire ad revenue stream could be significantly impacted. Prebid Stack fosters diversification by facilitating participation from a broader range of advertisers. This reduces reliance on any single source and ensures a more stable flow of revenue, even if performance on one platform fluctuates.

The path forward required a well-optimized Prebid Stack that would lessen their dependence on GAM. This called for a strategic overhaul, leveraging technology to harness the true value of every impression and create a more competitive and diversified ad revenue landscape.

The Solution

The publisher, determined to bridge the gap between GAM and Prebid Stack, partnered with Mile to implement its AI-powered dynamic flooring module - a cutting-edge solution specifically designed for Prebid.

The dynamic flooring module stands out by using machine learning to analyze auction data comprehensively. It doesn't just look at past performance; it predicts future outcomes, setting optimal floor prices for each auction in real-time.

Precise predictions. Optimized floors.

The module employs machine learning algorithms to analyze a multitude of factors, including historical performance data, winning bids, ad unit performance, bidder behavior, device type (desktop, mobile, etc.), user location, ad unit size, and many more.

Based on the analysis, it predicts optimal floor prices for each individual auction request. Covering the extra mile, the ML module then enters a continuous feedback loop, where it constantly adjusts the floor in real-time based on the bid performance data. By factoring in these ever-changing variables, the AI sets intelligent floor prices at every fixed interval that are both competitive for advertisers and ensure a healthy return for the publisher.

As a strategic move, the module runs on a control vs. experimental group setup. In this case, the module applied dynamic flooring on 65% of the impression, while 35% remained as a control group for a comparative effectiveness study. The result speaks volumes.

The Impact

Since implementing the AI-powered dynamic flooring module in the fourth quarter of 2023, the publisher experienced a 110% increase in Prebid revenue contribution. Within a short timeframe (4-6 weeks), Prebid's share grew from 10% to 21% of total revenue.

This signifies a significant diversification of the publisher's revenue streams and a reduction in their reliance on a single platform (GAM).

CPM Enhancement

The gap in CPM performance between Prebid and GAM was not only bridged but reversed, with Prebid CPMs exceeding GAM by 6%, a stark contrast to being 60% lower in the preceding quarter.

Key Performance Metrics

Mile’s AI-powered dynamic flooring has been a game-changer for the company, helping it grow and diversify its ad revenue streams.

The solution's impact transcends mere numbers. By setting optimal floor prices, the module ensures that the Prebid Stack attracts higher-quality bids, ultimately attracting premium advertisers. This not only boosts revenue but also improves the overall quality of ads displayed on the publisher's website. It’s a total win-win.

Table of contents