The metric you use to measure your ad revenue tells whether you are growing or not.
If you look at the wrong one, you might not get the whole picture and end up leaving money on the table.
Ad tech is volatile. Every year, we come across new advancements which lead to newer metrics such as CPM, CPC, CPL, and others. That’s why finding the right metric and sticking to it can be challenging for publishers.
Remember, you – being a publisher – work with site traffic, whereas these metrics deal in impressions/clicks – advertiser-side metric.
Hence, these advertiser-based metrics don’t apply to you.
So, how do you measure your growth exactly?
The answer is Page RPM
What is Page RPM
Page RPM is the revenue generated per thousand pageviews. It is an abbreviation of Revenue per Mille, or Revenue per Thousand pageviews.
RPM actually stands for “revenue per mille” — not mile, American’s way to measure length, but cost per mille with two Ls — which means “thousand” in Latin. The Roman numeral for 1,000 is also “M.”
Page RPM is a way to measure ad revenue for large numbers of pageviews. It’s like a shorthand, since expressing individual earnings per impression would become tedious and, in many cases, represent fractions of a penny.
How to Calculate page RPM
Formula to measure page RPM is simple and as follows:
Page RPM = (Estimated earnings / Number of pageviews) * 1000
For example, if you earned an estimated $15 from 2500 pageviews, then your page RPM would be equal to ($15 / 2500) * 1000, or $6.
Page RPM vs. RPM
Page RPM and RPM are generally used interchangeably. But sometimes these might mean different things.
For instance, Google AdSense’s definition for both RPM and page RPM is “Page revenue per thousand impressions (RPM) is calculated by dividing your estimated earnings by the number of page views you received, then multiplying by 1000.”
Notice that Google believes RPM is revenue per thousand impressions (not pageviews), however, the formula to calculate RPM includes pageviews not impressions.
Similarly, you can find more such acronyms in their Glossary – each more confusing than the previous one.
Impressions and pageviews are not the same. When a user appears on a webpage, it is counted as one pageview. Now this page can have 3 ad units, if a user is able to load all 3 ad units on the viewport, the server will count three impressions.
We will compare impressions and pageviews and their impact on revenue in detail, later in this post.
To reduce the confusion, it’s best to discuss with your ad networks what do they mean by RPM or page RPM.
Why’s Page RPM Important for Publishers?
Page RPM tells how much money a publisher makes for each pageview whereas the most of the other ad tech metrics (such as CPM) deal in impressions.
Consider this. When buyers agree to buy your inventory, they don’t mean to purchase your entire stock. For instance, you may have 1M impressions to sell per day, yet you only sell 750,000.
This means your fill rate is 75%. What happened to 250,000 other impressions? Do they go unsold?
If you just look at the CPM offered by advertisers, you would only know what they are paying you. You won’t know the actual price of your inventory and the mistakes you might be making.
This is where page RPM comes to the rescue. Page RPM tells you how much you are earning per thousand pageviews. You can further calculate page RPM for your individual pages, arrange them in order of high to low earning pages. This data can be very helpful in improving the performance of your low earning pages.
How to Increase Page RPM?
If you go on any search engine and look for ‘how to increase page RPM’, you will end up finding cliche results such as improving your content, engaging with the audience, or diversifying the revenue streams, among others.
However, we believe in actionable insight and here’s our take on page RPM optimization:
Ad refresh (also known as auto-refresh ads) is a technique that allows publishers to increase the number of impressions served per pageview by refreshing ads for active users based on predefined triggers. Obviously, more impressions directly translate to greater revenue generated.
Ad refresh works well on websites with higher-than-average session duration and time on page – such as online game sites.
This makes sense since the longer a user is on a webpage, the higher the chances they will pay attention to the new set of ads.
How is this related to page RPM?
Suppose you have a website that gets 100,000 pageviews a day and on an average, a user spends 5 minutes (or 300 seconds) on your pages. Right now, you have 100,000 impressions and pageviews.
Now if you deploy Active View Ad Refresh that refreshes ads after every 60 seconds, then for each user you can get 5 impressions. This means that your new impressions number is 500,000 – a clear increase in impressions.
Then why shouldn’t we calculate ad revenue by impressions?
Let’s elaborate on the above example to explain why calculating based on ad impressions might not be a good idea.
|Calculating for page RPM||Calculating for CPM|
|Ad revenue generated||$50||$50|
|Page RPM / CPM||Page RPM = $0.2||CPM = $0.1|
In the above example, we showed if we keep pageviews and impressions the same, CPM appears to be less than page RPM – even when the publisher has made the same ad revenue. This can confuse publishers about which metric to trust.
In the case of ad refresh, it is possible that the same advertiser is buying the impressions after the refresh. In this case, generally the bid price goes down.
For instance, an advertiser bid $5 for an ad unit, then the unit refreshed, the advertiser got the signal that he is about to show the ad to the same user. In this case, the advertiser might reduce the bid price. And as the cycle continues, the bid price might go down until another advertiser comes with a higher bid.
This means if you calculate your 500,000 impressions with your revenue, you might see a decrease in CPM, as for each impression you get a reduced price. However, your page RPM would show a boost – which is correct because you increased your number of impressions and ad revenue.
My bid price might go down, so ad refresh is good or not?
Again, for websites with high average time on site, ad refresh is the best solution. Bid price going down is not necessarily true for all publishers. But this is a case that we have noticed with a few, however, page RPM always goes up irrespective of CPM behaviour.
In short, applying ad refresh might show that your CPM is lower than the last month but your page RPM would certainly rise due to the increase in number of impressions per user.
Ad Layout Optimization
Layout optimization means testing and implementing ad units that gets you the best dollars in the market. It depends on:
- The layout of your website and individual webpages
- Places on your pages with high user interactions
- Popular ad units in your niche and among advertiser
- Device type – mobile, tablet, or PC
And you have to test for all these for individual pages. This is a tiresome process if done manually.
We, at AdPushup, use machine learning to run multiple tests for individual pages to get the best ad layout for your web pages. Once the test results are declared by the algorithm, we further check the layout manually for user experience before making it live.
How is this related to page RPM?
Let’s continue with the example we discussed in ad refresh.
Now we are testing inventories with optimized ad stack and non-optimized ad stack. As discussed, ad refresh is added to the mix which increases the number of impressions. However, for non-optimized stack, impressions are equal to pageviews. Also, let’s assume that CPM offered by advertisers is the same – $2 – for both cases.
Hence, the revenue earned by publishers will be (CPM * impressions) / 1000 or $1000 for optimized stack and $200 for non-optimized stack.
|Optimized ad stack||Non-optimized ad stack|
|CPM (by advertisers)||$2||$2|
You can clearly see in the table that page RPM shows clear growth in accordance with the ad revenue, however, CPM fails to tell the real story.
How’s ad layout optimized, exactly?
In the above examples, we assumed all pageviews resulted in impressions. However, in the real world, that doesn’t happen. Banner blindness is very real that quickly makes users ignore ad units on webpages.
In such a case, publishers need to constantly change the layout without being intrusive. Ad layout is optimized by testing various ad units and placements on the webpages. The algorithm A/B tests various ad layouts by complying with the Better Ads Standards and keeping user experience in mind.
We optimized the impressions (using ad refresh) and improved viewability (using layout optimization), next step is to increase bid price coming in from advertisers.
As a publisher, you would want advertisers to place higher bids on your inventory. For that, you need to increase bid pressure. This can be done by adding more tier 1 demand partners to your stack.
Basically, DSPs are equipped with systems where they can tell how many other partners are competing for this impression by looking at the bid request. If they see a lesser number of partners, they would assume that they have a good chance of winning and hence can be lousy with their bids. However, if they see bigger names (such as DFP, AppNexus, OpenX, etc.) competing for the impressions, they tend to increase bid prices.
Then there are publishers who are restricted with the type of ads they can show on their website. For instance, publishers wouldn’t want their competitors to show ads on their pages, this restricts certain advertisers to bid – reducing the size of demand pool. In such a case, having more options – in terms of demand partners who are willing to purchase impressions even with the restriction in place – is important.
How’s this related to page RPM?
Getting higher bids from advertisers definitely increases your ad revenue. Let’s go back to the above example again (to keep this simple, let assume there is no ad refresh enabled on ad stack)
|Without tier 1 demand||With tier 1 demand|
|CPM (by advertisers)||$2||$5|
In this example, you can see if we increase CPM, the page RPM increases directly.
We kept ad refresh and layout optimization out of the picture here. When these are added to the mix, page RPM should be better equipped to measure revenue uplift.
Putting it All Together
We haven’t added improving content quality, viewability scores, and user experience to our tips to increase page RPM. Reason being we believe you are already working on these things and understand the importance of offering value to users.
In case you aren’t, you can refer to our webinar on Improving Ad Revenue with User Experience.
Page RPM is a timeless metric that can be used to measure ad revenue on the publisher side. It directly depends on the pageviews, publishers can easily compare their performance from the previous month, quarter, or year – no matter whether traffic goes up or down. Additionally, it clearly shows where should publishers put their efforts – increase pageview or further optimize ad stack or both.
We, at AdPushup, use page RPM to compare publishers performance and generate reports. Since we work with multiple demand partners – who have different monetization models and NET payment cycles – page RPM makes it easy for us and our publishers to predict revenue that matches the actual earnings.