The KPIs You Must Look at When Analyzing Ad Creative Performance: An Insights-Driven Guide

By Alon Tvina

It’s time to advance the conversation around ad creative. Most of us have dealt with the fallout from iOS 14.5, and the announcements from across the industry about the reduction of reliance on user-level data make it clear that this was just the beginning. Users can no longer be the center of the online marketing universe, and creative assets are currently the best candidates to replace them.

Build your activity optimization around creative, and you’re immediately out of the privacy battle: aggregate metrics are all you need to be successful. Moreover, you’re tying your optimization flows to the one piece of the advertising puzzle you’re still in control of. Bidding? Automated. Audiences? Either automated or hiding from privacy watchdogs. But ad creative is still yours to shape and deploy, nor is it in danger of going away any time soon.

But what makes us think that ad creative, as important as it is, can be meaningfully optimized? Data. Lots and lots of data. Hundreds of thousands of assets. These images, videos, and playable ads, with at least $5.00 spend and at least one app download attached to them, were analyzed to bring you the good news: there’s still a lot of room for improvement.

In this report, we’ve assembled a few KPIs you should look at when optimizing your creative, and industry-wide insights associated with these KPIs.

IPM (Installs per 1,000 impressions)  is rapidly becoming the most important metric for creative analysis – simply because it reaches as deep as possible into the funnel without being affected by privacy protections and obfuscation. For these insights, IPM is our metric of choice unless stated otherwise.

KPI 1: Cross-Campaign Performance


You don’t have to overhaul your entire creative production in order to increase the success rate. Simply analyzing the activity from a creative-centric point of view can increase your performance. Bidalgo is one of the industry leaders in creative-centric tools, yet even on our platform, 45.2% of over-performing assets were discarded. They were positively impacting the performance, yet they were taken offline.

At the same time, 55.2% of underperforming assets were left online for more than 4 months. They were constantly dragging the activity down yet were allowed to run. While some period of activity for low-performing assets can be explained by testing, no one runs 130-day tests on new assets — and for a good reason.

While the IPM isn’t necessarily the most important metric at all times, it’s always important — simply because of what it says about the ability of the creative to attract users who will be persuaded to give the product a try. A good creative asset entices the right users to click, while discouraging those whose ad clicks will not result in conversions.

KPI 2: Creative Efficiency

To illustrate the problem of unnecessary creative production, we’ve looked at the creatives with a negligible contribution to the activity – expending less than 0.1% of the overall budget. Their overall contribution to the spend is measly, yet time and money were invested to create them. How many of these assets are there? Too many.

More than a quarter of assets have 0.1% of spend attached to them! Moreover, the bigger the budget, the lower the efficiency. And while the biggest marketers are both less sensitive to production costs and require more variety to succeed, 39% is an awfully high number.

When we look at the creative performance from the other direction, another insight emerges.

Think of the risk this kind of reliance poses. Fatigue is inevitable, and the spend concentrated on these assets makes each of them a potential threat to the activity. And the bigger the budget, the higher the risk.


KPI 3: Performance Distribution

Let’s dive deeper into the IPM metric to understand how data can help you measure the overall health of your creative asset library and its utilization.

In an ideal situation, all the assets are of similar quality, representing the best possible effort. In this case, the IPM for the vast majority of the assets should be somewhat similar. Yes, there will always be duds or breakaway successes, but not a lot of them.

What actually happens is very different. We’ve designated three very broad groups of IPM performance to see how many assets fall into each. Underperforming creative brings in less than 50% of the average IPM for the combination of product, channel and asset type. The average range is between 50 and 150% of the average. Breakout hits are the top creatives, bringing in more than 150% of the average IPM.

In the best-case scenario we’ve described, the “Average” group would’ve been by far the largest, with as few outliers as possible.

It’s easy to understand why we want few low-performing creatives, but why wouldn’t we want breakaway hits? Because they reflect badly on all other creatives. A meaningful percentage of creatives with much-higher-than-average IPM means that the average itself isn’t high enough.

The variance between images, videos, and HTML5 playable ads? Isn’t significant, it seems. The distribution of creative isn’t localized to a single type. There’s still a lot of work to be done across all media types before creative production is data-driven enough.

KPI 4: Creative Attributes

There’s no simple way to overhaul your creative-related processes and make them more efficient. There are often multiple teams involved, with different KPIs, priorities, and access to data.

Performance data can be combined with metadata to extract valuable, actionable insights. You should, for example, bundle creative assets by attribute to understand which common attributes work well on each of your target audiences. These can be sizes, formats, resolutions, or even designers who produce the assets – some of whom are bound to be more efficient than others at cracking the creativity code of what makes some assets better than others.

To point you in the right direction, here are some of the industry-wide metadata insights we see. You can treat them as best practices, or as a stepping stone for personalized optimization.

Top 10 most popular video lengths, and the percentage of breakout hits associated with them.

Top Resolutions

Top Aspect Ratios

Content Mix by Platform


These insights are just a drop in the ocean of data, which we’re trying to navigate day after day. The tools and services Bidalgo builds help make this navigation easier and get marketing teams to their goals faster. And while industry-wide insights can get you quite far on this journey of performance and optimization, to go the extra mile you need to work with your own data and optimize to your own goals. We’ll gladly show you the way. Let’s talk.



Published on October 7, 2021
Written by
Alon Tvina

Get Started With Bidalgo Today