40% Waste*, 100% Opportunity: Revolutionizing Digital Ad Spend with Supply Path Optimization

If you are new to this topic or need a refresher on the foundational aspects of SPO, start out by reading our previous article on Supply Path Optimization.

Note: The insights and figures presented in this blog post draw upon our expertise from Supply Path Optimization (SPO) projects, involving around ten clients across the US and EMEA regions.

This article was written by Aspram Grigoryan, Data Analyst at Jellyfish, and Thibault Martin, Data Insight Director at Jellyfish.

Hidden costs and inefficiencies in programmatic advertising

Every second, countless digital ads are bought and sold in a high-stakes bidding game played at the speed of light using programmatic pipes. Yet, beneath this rapid exchange, a confounding inefficiency exists—a staggering 40% of advertising budgets are lost to the depths of the programmatic supply chain.

In the complex landscape of digital advertising, advertisers share a straightforward objective: to deliver impactful impressions to users at a fair price. However, achieving this goal programmatically requires navigating an ecosystem of numerous actors and advanced technologies, all orchestrated to perform an auction to win impressions on websites and capture user attention. These auctions unfold behind the scenes, completing in mere milliseconds. Before an ad reaches the domain where it is displayed, as illustrated in the image below, it passes through several digital supply chain intermediaries, making its journey quite intricate.

As monetization strategies of publishers have evolved, significant shifts have occurred in the programmatic market: the shift from traditional waterfall strategies to widespread adoption of header bidding, and the transition to first-price auctions disrupting standard monetization rules. Although adtech providers on both the sell-side and buy-side have swiftly adapted their tools to these changes, they've provided marketers and publishers with minimal visibility into these processes.

Simply put, what this means is more intermediary reselling fees, less transparency and losing the bids you try to win. 

Our findings consistently confirm simple truths. Generally, the more intermediaries (SSPs, resellers, etc.) that there are in the supply chain, the higher the costs—per impression, per click, per visit, and per booking (see the chart). Notably, while the cost per impression increases to a manageable extent, the costs per click or booking rise steeply. This trend suggests that despite their high price, impressions obtained through complex paths might lack the effectiveness needed to generate leads.

See below the evolution of KPI performance across number of intermediaries

How does this affect programmatic investments? In an audit of more than 40,000 supply paths that a client bought through, we saw that around 75% of the supply paths are indirect, and about 50% of the investment still flows into these paths, raising concerns about the efficiency and transparency of spending in digital advertising.

So, how should advertisers respond to the potential inefficiencies of supply paths? With more paths, there are more puzzles, so simplifying can amplify results. As brand advertising dollars increasingly pour into the programmatic ecosystem, industry players are pushing back against complexity throughout the supply chains. In recent years, media buyers have closely examined their supply paths to make sure every advertising dollar is spent effectively.

However, while it may seem appealing to avoid investing in indirect supply paths altogether, this trivial solution would overlook the fact that many of the indirect paths are not only efficient but may also represent the best available options at certain times. No intermediaries Better paths! As evidenced below by some example paths purchased through a certain publisher, some indirect paths offer attractive Cost Per Thousand Impressions (eCPM) and competitive Cost Per Acquisition (eCPA).

This strategic insight calls for a more informed decision-making process, a necessity to optimize the balance between direct and indirect investments based on actual performance data rather than a blanket avoidance strategy.

Additionally, when advertisers purchase inventory through DV360, the investment tends to concentrate among a limited number of players. Should they consider diversifying into a broader range of marketplaces? 

Navigating Complexity: Our Approach to Supply Path Optimization 

At Jellyfish, we took the chance to tackle the inefficiencies media buyers are facing with SPO.  

The concept of SPO (Supply Path Optimization) has been around since 2016, but it has since developed in tandem with the broader programmatic industry. Nowadays, SPO strategies of varying maturity levels are being implemented by diverse buyer profiles around the globe. Yet the wide range of possibilities within SPO has led to some confusion and misconceptions regarding what SPO actually entails and who it affects.

Think of SPO as our pipeline for navigating the complex network of ad placement routes—a complete workflow that integrates various stages of the project, from data ingestion and preprocessing to understanding the supply mix, model training, evaluation through A/B testing, deployment, and Continuous Innovation and Continuous Development, all while ensuring a seamless process. 

Using this framework, we start by thoroughly analyzing the programmatic data. Leveraging Google Ads Data Hub (ADH) — Google's privacy-conscious data clean room environment — allows us to access detailed supply path data otherwise unavailable through native reporting tools. With this information, we address critical questions: Do our customers truly understand their supply mix as well as their own pillars for success? Where is their media budget currently allocated across sellers, and through which exchange marketplaces? How many sellers are they engaging with? Are the majority of impressions funneled through a relatively small number of partners?

We then audit the paths that ads take to identify bottlenecks. Each element of the supply chain—from exchanges and sellers to domains—is then evaluated using a machine learning approach. It allows us to identify critical areas that could impact an ad's effectiveness. Our findings are then transformed into a real-time bidding algorithm that encourages the best paths (see J+ Bidding - our proprietary technology for real-time bidding), reduces costs per impression and boosts key performance indicators. This sophisticated approach allows us to optimize for not only direct paths but also to discover efficient routes among the indirect ones. 

However, optimizing a supply path cannot be a ‘set it and forget it’ effort. Which is why we use an iterative process that requires careful evaluation and continual assessment. We always reserve a portion of the traffic without optimization to maintain a control group, serving as a reliable benchmark to assess market conditions at any time. Through continuous evaluation and refinement we determine whether the ideal path one day is still optimal the next. 

Transformative Results: Lower CPM and High Performance through SPO

In a recent collaboration with a client, we tested our scoring methodology for a display pre-roll video format in an A/B experiment setting across a whitelist of digital-ad-trust domains. We achieved a 40% reduction in CPM (Cost per Thousand Impressions) and enhanced the ROI (Return on Investment). We also maintained key performance indicators like Ad Viewability and Video Completion Rate intact in the optimized group. Essentially, we delivered the same volume of impressions at 40% lower cost, while maintaining or improving the performance.

Key learnings? Our customized bidding strategy has revealed substantial variations in sellers effectiveness. For instance, not all sellers associated with Google Ad Exchange deliver consistently superior performance. As a result, the SPO strategy has broadened our reach, allowing us to diversify investments across various exchanges. The optimization algorithm reduced the investment in Google Ad Exchange, which freed up funds for allocation to other platforms. Alternative exchanges like Index Exchange, Magnite DV and Pubmatic demonstrated lower CPMs and better performance in certain paths resulting in an increase in investment in those exchanges***. Despite these shifts, Google Ad Exchange still dominates the investment portfolio, as most of the paths through which we buy via Google Ad Exchange continue to showcase high performance and offer competitive CPMs.

These adjustments are dynamic, continually refined based on performance data to ensure that our investments align with the most effective supply paths. This approach maximizes ROI and boosts the impact of our digital advertising efforts, setting a new standard for efficiency in programmatic advertising and ensuring that each dollar spent is a strategic investment toward measurable success.

On the supply side, SPO can be a powerful tool for sellers, facilitating more transparent buying processes and enabling them to maximize revenue. However, implementing SPO requires a nuanced approach, as increased cost transparency from removing intermediaries does not automatically correlate with maximized performance.

Step Up Your Advertising Game with Jellyfish

Excited about refining your programmatic spending and boosting your metrics? Partner with us at Jellyfish. We'll help optimize your investments and transform ad spend uncertainties into predictable, profitable outcomes with our innovative solutions.

*40% waste refers to the cost savings achieved by using Supply Path Optimization (SPO), indicating the excess expenditure that could have been avoided in the absence of SPO.

**All the top-spending indirect paths shown in the chart and purchased through a specific publisher, involved exactly two intermediaries.

***It should be noted that a brute force strategy of simply increasing investment in certain exchanges may not yield the desired outcomes. A thorough evaluation of all elements involved in the path is essential.

Credentials Information: 
Contributors : Thomas Allemand, Rafael Lasnier, Marie Racary , Karl Adaimy, Astrid Lesort, Theo Blou