One of the most difficult analytical challenges facing digital publishers is measuring the effect of quality on revenue and profits.
Quality is a characteristic, much like beauty, that is in the eye of the beholder. What counts as quality content and an excellent user experience to one reader may be irritating to another. Quality is hard to measure because there is no recognised “quality” attribute with definable dimensions.
Is it the frequency of publishing new articles? Is it the average length of posted videos? Despite the challenge of measurement and consensus, quality is a “thing,” and it is the attribute that differentiates digital publishers most profoundly.
High-quality products lead to greater user engagement. While not universally defined, engagement for a customer can be captured in the frequency of visits, time passed since the last visit, and time spent on the site.
Other dimensions of a visitor’s interaction with a site can be used to measure engagement too, but these three metrics are most often used by leading digital publishers.
User engagement is observed to lower churn among existing subscribers and increase conversion of non-subscribers.
Certainly, there is value in keeping more of your existing customers and getting more new ones, but what is that value? How much investment in quality content, and thus engagement, is justified? How many advertising positions should be removed, and revenue lost, to improve the customer experience?
These questions are at the heart of the publisher’s dilemma.
The publisher’s dilemma is the challenge of maximising total digital revenue, from advertising and audience customers, when the demands these revenue streams place on the product are contradictory.
In the most basic characterisation of this dynamic, advertising revenue depends on quantity of audience impressions while audience revenue depends on quality of content.
Digital advertising revenue is a function of the number of impressions delivered and the effective CPMs for those impressions. Digital audience revenue is a function of subscription sales attempts and conversion rates while acquiring new customers and retention and average rates while managing existing customers.
At their core, these two digital revenue streams are in opposition. And for editorial staff and newsrooms, balancing these two sources of revenue requires balancing the quality of the product with the quantity of the product. Executing integrated product design and product pricing strategies is crucial for a publisher’s success.
As digital advertising revenue begins to level off, and in some cases decline, on publisher sites, the need for greater audience revenue is paramount.
The trend in the publishing industry for years has been to erect paywalls to limit consumption of the content to non-paying visitors. But too often this desire for paying subscribers has not been supported by an increase in the quality of the digital product, which has led to poor conversion rates and low digital-only subscription sales.
Some publishers have tried multi-site strategies, notably The Boston Globe and Atlanta Journal Constitution. They have targeted one site at the large non-paying audience and another site at their paying digital subscribers. The site targeted toward the paying digital audience has a better user experience, higher-quality content, and less advertising. The site targeted toward the non-paying audience is the opposite.
Other publishers have focused on growing digital advertising revenue with minimal digital audience revenue. They have no paywall and do not offer digital-only subscriptions. To support the advertising-only revenue model, they capture as much data as possible about their audiences and use it for targeting ads to increase their average CPM.
Risks of this advertising-only strategy are that the additional tracking and targeting may cause site speed to slow down, and the data capture on the audience may become burdensome and annoying to the users. These sites typically have a lower-quality user experience in the form of more advertising and less quality content.
So, what performance metrics are important for balancing quantity and quality and thus advertising and audience revenue?
In addition to the user engagement metrics mentioned previously, important data to capture includes the total digital revenue by individual user. This can be difficult because it requires attaching advertising revenue from the impressions delivered to an individual user to the site traffic data for that individual.
The publisher should also capture data on what happened to individual users as they went through a paywall funnel so they can analyse the effectiveness of subscription offers. Using these data, a publisher can identify the combination of product attributes, including premium content access, quality user experience, and price point, where expected advertising revenue and audience revenue is maximised.
There are many analytical approaches that yield insights into the return on quality content and better user experience, including estimating a customer’s propensity to subscribe, segmentation of the online audience, hedonic pricing models that measure the effect of individual product characteristics on price elasticity, and retention models using survival analysis.
These tools are helpful when used within a framework for optimising total digital revenue where the effects of business rules on each revenue stream are quantified.
As with many economic questions, the point of maximum revenue for firms that have two distinct customer groups, such as advertisers and audiences, is where the next dollar of one type of revenue costs you one dollar of another type of revenue.
This “trade off on the margin” is a standard economic concept that supports many type of optimisation applications. In digital publishing, the optimal revenue model is unlikely to be 100% advertising revenue or 100% audience revenue because there are often “easy” dollars to be captured in both categories where you do not have to give up revenue from one stream to get incremental revenue from another.
Where the optimal balance between revenue streams falls for each publisher depends on their unique audience segments, advertising inventory, and products. Off-the-shelf algorithms are unlikely to fit a particular publisher’s business because no publisher’s site or products are “average.”
To take the example of optimising revenue streams on the margin one step further, consider the question of determining the optimal paid content threshold, whether it is in the form of a meter or a more general “premium content” strategy.
The optimal paid content strategy for a digital product would be where the paywall restricts access to content for an audience segment where the expected audience revenue from the next paywall “hit” is equal to the lost advertising revenue from limiting access to that content.
If the expected audience revenue was greater than the lost advertising revenue required to make that subscription sale, the access should be further limited so that more subscription sales are attempted.
The same logic holds in reverse. If the lost advertising revenue is greater than the expected audience revenue from the paywall, more content should be offered for free.
As you have likely surmised, this optimal strategy is dynamic across time and not consistent across the entire digital audience for a digital publisher. Segmenting the audience based on its propensity to subscribe and price elasticity, and understanding the demand for digital advertising inventory that different audiences provide, are key elements of optimising total digital revenue.
The publisher’s dilemma is a critical challenge not only for news media companies but also for many digital companies that rely on a mix of revenue streams, particularly those, like publishers, where the needs of their two customer groups have such different effects on the product.
Understanding how much to invest in quality content and user experience to support audience revenue, and knowing how much advertising revenue to put at risk to sell subscriptions, is how to solve the publisher’s dilemma.