How can you boost your programmatic revenue? Choosing the right programmatic provider should be a data-driven decision

15. December 2019

Have you been using Google AdSense and not sure how well it’s working?

The name of the game when it comes to incorporating ads as a publisher is earning more from your content. It doesn’t do any good to enter the marketplace of programmatic ads without measuring the current impact it has on your revenue, and more importantly, the revenue growth you could unlock by monitoring the data trends and consistently optimizing your approach.

If you’re not happy with the limited possibilities offered you by AdSense, FatChilli can give you access to Ad Manager for your own use. But how might that help?

Where should I start evaluating my programmatic revenue?

Let’s start with a few key differences between AdSense and Ad Manager.

AdSense is essentially an ad network that connects advertisers directly with publishers, while Ad Manager is an ad server that acts as a middleman between publishers and ad networks or exchanges, and which determines how these ads are served on their website.

Put another way, the primary purpose of Ad Manager is to put publishers in the driver’s seat, giving you the tools to control how ads are served on your website (regardless of where they’re sourced from), as opposed to just providing the ads, as AdSense does.

The main benefit of using Ad Manager over AdSense is that while AdSense provides you the ads upfront but not the control over how they’re served, Ad Manager allows you more options for how to attract ads, serve them on your site, and gain higher programmatic ROI. 

Ad Manager offers you the opportunity to manage multiple demand channels, which ultimately helps you get the most value from each impression.

When it comes to analyzing the current state of your digital ad revenue, you want to make an informed decision rooted in resource efficiency. So let’s start at the beginning with the concrete evidence: by taking a look at your data.

How can you compare if the inventory managed by Fatchilli is performing better than AdSense?

The FatChilli Client Zone provides the same data that is available in AdExchange/AdSense, so it’s a breeze to compare AdSense data with what you’ll find in our reporting app, which comes via Ad Manager.

The first step is to create a custom performance metrics report that will help you gather the necessary insights towards optimizing your revenue.

Okay, great. But what metrics should I be interested in?

The most important metric towards evaluating your programmatic revenue is ad request eCPM, a custom metric available in Ad Manager

Note: Ad requests are counted whenever your site requests ads to be displayed. Google reports them each time a request is sent, even if no ads are returned. Ad requests represent the number of ad units that requested ads.

As you know, CPM means “cost per mille,” or in other words, the price for every 1,000th ad impression. CPM represents both a buying method for ad impressions as well as the price advertisers pay for those ad impressions.

On the other hand, ad request eCPM is the amount that you, as a publisher, receive for every 1,000th ad impression---regardless of what buying method is used (whether fixed price, CPM, CPC, CPA or CPO).

But, you ask, what about overall revenue?

That’s a good question. Before we continue, let’s dispel a common misconception. Of course you are interested in revenue; however, here’s why it’s not the ideal metric to use in evaluating programmatic performance:

  • Seasonality. Revenue can shift with the seasons, particularly in January and June, when ad agencies are in periods of low spending (for example, December, or the end of the fiscal year, is a period of high spending because of the holidays!).
  • Unstable pageviews. If your pageviews are unstable, this can lead to unstable revenue, meaning that one extremely good or extremely bad day can wrongly influence your findings and thus your assumptions about average earnings.

Why is eCPM a better choice?

First of all, ad request eCPM is designed as a benchmark metric in that it shows your revenue per every 1000 ad requests. 

It is useful precisely because it represents this universal standard of measurement regarding revenue for the impression sold. For that reason, it helps you compare site performance to averages and can be used as a critical indicator of campaign performance.

It helps you evaluate and optimize your current efforts by letting you compare ad revenue generated across multiple variables, such as ad network, region, location, etc.

For example, let’s say you want to understand which ad unit is performing better and making you more money. To do so, you’d calculate and compare the eCPM of both.

In a month, you see that rewarded video generated 400 impressions and $5.00 in earnings, while banner ads generated 700 impressions and $3.00 in earnings. It’s hard to compare the performance of the two ad units based on these numbers alone.

But after some quick math, we find that rewarded video has an eCPM of $12.50 and banner ads have an eCPM of $4.29. Now we see clearly that rewarded video is making you more money.

What if I don’t have eCPM data?

Do you have more network partners? It is possible they are not providing a metric that supplements ad request eCPM. If that's the case, you should look at your ad server data. You can use the "Total code served count" metric from Ad Manager and calculate your own alternative to ad request eCPM with this formula: 

(overall revenue / "code served count")*1000 (visualize)

See more about the available metrics here

Remember: If data analysis or interpretation is not your cup of tea, Fatchilli is here to help!

How can I improve my revenue?

Now for the (literal!) money question: Naturally, every publisher is looking for the most profitable solution. We recommend A/B testing and weekly evaluation of your metrics.

A/B testing

A/B or Split-Testing is the process of creating multiple variations of an existing web page to boost your eCPM (alongside click-through rate) and thus increasing your revenue. This is the most accurate way to test and optimize your programmatic performance.

In the simple version of A/B testing, you choose one (and only one!) variant of your ad to test. The steps are 1) Theory/Hypothesis, 2) Test, 3) Result, 4) Repeat. 

Your goal is to gather data by testing the ad on a 50/50 split of new viewers until you can tell which version performs better---but not too long, otherwise factors like seasonality or instability  may start to factor in, making your data unrealistic.

To begin with, there are more ways to gain more ad revenue that just replacing the source of the ads if it doesn’t seem to be performing well. Instead of segmenting your inventory, run A/B tests on all pages in a 50:50 ratio with your audience (or a proportional ratio if you are testing more partners). For example, you can use different website layout, ad units, or ad formats. Once you’ve done some tests, you may also start on questions like, “What ad combinations generate better revenue?”

Here are some examples to start:

    • What ad format gets better CTR? Consider what type of ad is best for you; you can choose between banner ads, interstitials, video ads (you don’t need pre-existing content with FatChilli), and so on. You can even test a display version of an ad with a text one. Each have their own advantages depending on the content you’re monetizing.
      • Best practices: Experiment on a page with little traffic to start so that you don’t disrupt popular content.
    • What ad placement gets the best CTR? The best ad format also depends on where an ad is placed in the content flow. Best practices will tell you to serve the ad when readers are most engaged, but not when or where it will disrupt their experience. You can also experiment with ads on the left versus the right (we naturally scan from left to right, but this also means that people get used to ads being on the right and tune them out), and above versus below the fold (but if above, make sure the whole ad can be seen without scrolling).
  • Best practices: Your site’s UX is key. Make sure that ads do not overlap with content or otherwise distract from it; readers dislike this most, and disruptive ads--whether in placement, color, or sound--do not perform well.
  • What ad dimension gets the most impressions? You’ll want to consider your website layout to see what ad size has the most potential. Sizes like the leaderboard (728×90) and small rectangle (300×250) have proven results.

What did you find?

  • Did your hypothesis about which ad would perform better, and why, hold up?
  • Do you have statistical confidence in your results? (95%)
  • Is your data inconclusive? Test again! And then test another variant.

Now, check your ad request eCPMs! If it’s higher than before, you’ve run a successful test. connect back to highlighted math section above?

Week-based metrics evaluation

Here are some tips for choosing the right period for comparing metrics. It doesn’t make sense, for example, to compare December with January, since advertisers spend much more money in December because of Christmas. 

Instead of a month-based comparison, use a week-based method and pick two weeks in the middle of the same month to evaluate your metrics. We recommend the 2nd and 3rd weeks of the month, since while you can use the 1st or 4th weeks, these may not give you the most accurate results (marketing budgets are often tricky at the beginning and end of each month).

Best practices: Be sure to avoid weeks with public holidays, events, special occasions (like Christmas, the Super Bowl, Black Friday, and so on) because these will not be an accurate or relevant representation of usual ad performance. Plus, people will be focused on other things---and certainly other, more public ads!

While evaluating, check your ad request eCPMs. If they’ve gone up, it’s a signal to continue in the direction you’re going!

I want to get started. What do you recommend?

It’s said that programmatic ads can reach 98% of the internet, so if you want to maximize your revenue, don’t give up on them too quickly.

Even if A/B testing is more time-consuming at the beginning, it pays off. You’ll benefit from the informed, data-driven decision for a long time afterwards.

With access to the ad network inventory through FatChilli, we can give you at least 7 great reasons why you’ll see better results in the form of a revenue boost. Not only does our artificial intelligence make miracles real, but we’ll put the power back in your hands to do things you could never dream of in AdSense---like running direct campaigns and closing Preferred Deals, incorporating high-value video ads even without video content, and get stronger reporting tools that allow you to analyze, act and evaluate for long-term improvement.