Why you should downplay CPM ad rates and think 'earnings per visitor' instead

Maximizing the amount of advertising revenue a news organizations gains from each visitor to its site relates to a lot more than traditional CPM (cost of an ad per thousand impressions) rates, according to Gavin Bechtold, a business development specialist at advertising optimization firm Ezoic, which is a sponsor of this year's LION Summit, the country's largest gathering of local independent online news publishers.

Bechtold will be speaking at the 2017 LION Summit at Columbia College in Chicago Oct. 26-28 about how user experience affects ad revenue. We asked him some questions in advance of that talk.

1. "ARPU" – average revenue per user, or unit – has become a buzzword in publishing. In what ways are news organizations leaving money on the table and how can this measure help inform that?

ARPU is what we refer to as EPMV (earnings per thousand visitors). Organizations that are not using these metrics but, instead focusing on RPM (the AdSense metric for revenue per thousand page views) could be leaving a lot of money on the table. RPMs and CPMs have been somewhat useful in the past, but they're really a metric for advertisers, not publishers.

EPMV or ARPU can give publishers a true north. It can help them account for things like session length (by looking at page views per session inside the calculation). This will allow publishers to account for the value of good user experiences on ad revenue  We see sites all the time where the publisher adds more and more ads to their pages in the (mistaken) belief that more ads = higher RPM = more money. This is misguided. Even though adding one or two ads might earn you more per page, those extra ads might be driving your users away. Driving users away pushes your bounce rate up, your pageviews down and time on site down.  This causes fewer pages to be viewed overall and actually lowers your ad earnings!!!

2. How significant is the role of overall news site design in the performance of advertising?

Not as important as you might think. Creating a site that is 'pretty' to the eye but poor in usability is a more costly mistake. Creating a pretty site that lacks engagement and drives away users equals a poor user experience and will ultimately hurt your advertising performance. There are really two ways to measure user experience; objectively or subjectively.  If done objectively (the preferred way); the most common and accepted form of objective data for user experiences right now are bounce rate, time on site (session duration), and pageviews per visit. These are the best high-level ways to see how users feel about a digital piece of content. They are essentially voting with their behavior. However, we think that when you look at these metrics with even more granularity that you can go a step further and better understand how user engagement is really affected by changes on a page.

If done subjectively, this would look something like a survey. It is commonly performed by asking for specific feedback regarding how a visitor feels about certain elements. While this kind of data can be helpful in guiding some decision making – if it's not balanced with objective data – you put yourself at risk for creating a poor user experience.

We've seen a lot of publishers — I mean a lot — spend valuable time on website redesigns, relaunches, and the alike only to lose money when it launched. Its super-important to ask WHY you're looking to redesign and reformulate something. Testing is your best friend here.

3. Is there a traffic threshold that local news sites should reach before programmatic makes sense as a supplemental revenue source to local ads?

Sites of all shapes and sizes should be looking at programmatic as a revenue solution. Major publishers like The Guardian are saying that they will be 100% programmatic by mid-2018 now. The question has to be asked, is the current form of ad sales and ad operations sustainable, efficient, and profitable for your business? We know that programmatic can be for everyone. It is something that is growing and must be explored. The longer publishers wait to explore these transitions and improve their programmatic earnings, the more risk they put their business in long-term.

What is a publisher actually earning from these direct deals once they take into account the effort it took to secure that deal — salaries, health insurance, lunches, etc.?

4. What are the biggest mistakes a publisher can make in working with programmatic advertising networks?

Using CPMs (cost per thousand impressions) to measure performance. While ad networks traditionally use this, your goal should be to maximize the amount of revenue that you earn per visitor. CPM does not measure that in any way.  It was created to try and measure the value of showing a single ad in a single location. However, CPM does not take into account the effect an ad has on user experience metrics (like page views / visit). If an ad provider doesn’t display per-user revenue – their interests aren’t aligned with yours. The most important metric for publishers is session revenue or EPMV (earnings per thousand visitors).

Signing contracts with ad networks. These guys will always try to lock you into long-term deals. They have a number of tricks they like to use to make it seem like the greatest thing in the world up front, but once under contract, things usually get a lot less great.

Not testing or making these guys all compete. Listen, competition is a good thing. You want as competitive environment as possible for your ad space. This is what drives up the cost. This is why tricky CPM measurements and contracts are ultimately bad for publishers. Ask networks to compete or to test against each other.