The “Gordian Knot” of the publishing industry is: How do we unlock the value of our growing digital audience?
Right now, new media is being played with old rules from old media. Traditionally, there was a lot of guesswork and waste involved in the placement of print media ads. Since it didn’t create a quantifiable effect, media companies priced units to advertisers using CPMs that had no functional basis in reality.
This hit-or-miss process has migrated to digital media, even though in digital there is much more data available about readers and their reading habits. Publishers have struggled to show advertisers the value of their audience because deals are still negotiated based on CPMs.
But why play by the old rules? Here is a way to get out of the business of CPMs and still be a high-value target for advertisers:
The audience in digital media such as the web is measurable and can be analyzed far beyond the means of traditional media. It is this aspect that should be emphasized when constructing revenue models for content businesses. Instead of the old commodity model in which one audience member is as valuable as the next, the Affinity Engagement Index (AEI) model proposes to specifically show how valuable an audience segment is to an advertiser, in a way that can form the basis for a negotiation (much in the way that CPMs work now).
The model is based on two critical factors:
Affinity, for the purposes of this model, is a measurement of how much the audience likes the publisher’s content. While affinity can be measured on the basis of an individual for a specific content item, for AEI purposes, we look at audience segment affinity for “clusters” of content, which are organized together by common metadata (such as author, category, topics, etc.).
Engagement is the level of an audience member’s interaction with, and attention to, a publisher. A high-engagement audience member is more valuable than a low engagement audience member, because they are paying more attention.
The model would create scores for each factor through various methods:
AEI = % Affinity x % Engagement x 100
For a publisher’s audience, we would look at the AEI numbers as well as audience size. With those two numbers an advertiser can make an apples-to-apples comparison among publishers.
Potential represents the value of an audience segment. Pricing for advertising should be compared against the Potential number (it is the equivalent of CPMs):
Potential = AEI x Segment Size
So for example if a segment had an AEI of 45 and contained 5,000 people, the Potential would be 225,000.
While you could describe a complete aggregate audience for a publisher in terms of their AEI score, it’s likely that a publisher will want to segment their audience, both by gross divisions in affinity clusters (this portion of the audience is into web development, this portion is into typography etc.), but also by engagement levels.
You could imagine a publisher having a rate card that showed their audience divided into affinity clusters and engagement levels, and audience sizes for each, and then showing the ad rates for each segment.
On the TWIT podcasting network, there’s a show called “This Week in Enterprise Tech,” which covers information technology concerns from the perspective of large enterprise. Recent shows have examined IPv6 (the Internet routing protocol), server virtualization, enterprise information security concerns, etc. It’s clear that the show would appeal to people who work inside enterprise tech, from IT staff through CIOs.
The show looks like a good target for companies selling enterprise IT services such as firewalls, network equipment, virtualization technology, etc. However, the content of the show varies, so it would be nice to segment the audience into more specific affinity pools. For example, one affinity cluster-based segment might focus on networking topics such as IPv6.
The TWIT network would create a rate card for this segment balancing the size of the segment with the highest affinity ratings. This segment would be sub-segmented into engagement levels (high, medium, and casual). Segment size and AEI numbers would be available for each segment. Let’s say that medium engagement segment here (70% engagement) represented 5,000 people and had an affinity level of 60%. The AEI score would be 42. (60% x 70% x 100).
To make this work we would need to create standardized measurement models for Affinity and Engagement. Measurement activities would need to be auditable for adherence to the standard model.
An engagement measurement model would essentially be a type of lead scoring in which engagement levels are inferred from user behavior. There would be a standard list of key user actions and associated scoring for each action.
Key user actions might include:
- Share content item on social network
- Bookmark content item
- Optimal amount of time spent with content item (too short indicates a bounce, too long means someone has walked away from the device).
- Print content item
- Return visit to content item
- Participation in online forum
Affinity measurement could trigger off of many of the same user actions as engagement measurement, but instead of scoring the user, it creates a level of association (the affinity value) between the user and metadata associated with the content item in question.
For example, a user who took an action indicating affinity with an article (such as sharing it on a social network) might establish an affinity relationship between themselves and the topic of the article (such as “networks”) and its tags (such as “cisco”, “IPv6,” and “Internet”).
A food magazine with 1 million subscribers creates an AEI model for bakers.
The baker segment has an average affinity for baking related topics of 75%. In the segment, 5,000 are highly engaged (80%+), 15,000 have medium engagement (50%+), and 30,000 are casual (25%).
So, for baking supply providers, for this segment the AEI chart is:
Interestingly, the smaller medium engagement segment is worth the same as the larger casual engagement segment. Pricing is based on the potential number.
Analyzing a segment of 100,000 wine lovers, it turns out that it makes sense to break the segment down into a couple of smaller affinity clusters:
Not particular to specific wines, just interested in what is trendy. Bifurcated between high and casual engagement levels: (15,000 high engagement, 8,000 medium engagement, and 37,000 casuals). Base affinity level for trend followers is 70%.
Here’s the card for trend followers:
(Interestingly, in this case reaching the 15,000 high-engagement layer in this segment is more valuable than the 37,000 casuals.)
We identify 20,000 wine enthusiasts that are just partial to California wines. The average affinity for this segment is higher (80%), but the engagement levels skew more casual.
Again interesting that the medium layer is half the size of the casual layer, but worth the same.