Intelligent Paywalls and Dynamic Paywalls: Similarities and Crucial Differences
Prefer to listen instead?
To answer that, we first need to understand other types of paywalls that came before them.
The Historic Paywalls
For the first decade or so of mainstream digital news (and other media services), paywalls were typically built in one of three ways:
- Hard paywall – where all content is behind the wall.
- Freemium – where premium content is behind the wall and lower value content is public.
- Metered paywall – where limited content consumption is permitted before a paywall is dropped.
While these three approaches can work for some businesses, it’s worth assessing them in more detail to understand their weaknesses, and the need for more advanced solutions.
- A “hard paywall” sends the unambiguous message that the product is premium and high value. However, for a consumer who has never experienced the brand’s content, it’s hard to foster engagement before subscription. The propensity to buy is entirely driven by brand and reputation. This is the polar opposite of the purely ad-funded approach, where the perception of content value diminishes greatly.
- “Freemium” tries to solve the engagement problem above by giving access to selected pieces of content so that the consumer can get a taste for the publication. However, in most cases, the freemium model either isn’t appropriate for media businesses or has been poorly implemented. This is due to the fact that typically, the highest value content is behind the paywall. That means the taste a non-subscriber gets of content is of the lowest value, and doesn’t create the strongest first impressions.
- And finally, “Metered paywalls”, which are the predominant model for digital news & periodicals today, are another attempt at trading off some value-perception for more top-of-funnel engagement. In this case, the meter is used as a proxy for engagement. If a subscriber has used up their meter, the publisher considers them engaged and shows a paywall; if a subscriber has consumed less content than the meter limit, they are considered unengaged and given more content. This theory is good, but it can still be improved…
The Rise of Intelligent Paywalls
Early generations of the meter can often be considered far too blunt. A publication might settle on ten free articles per month. But why ten? Why monthly?
As split testing and data-driven product management practices became more widely adopted among publishers, there was a push across the industry to discover the “best” meter configuration. But pretty quickly, it became clear that this exercise was doomed to failure, because, in fact, there is no “best” meter configuration.
The optimal meter differs from one consumer to the next. This is where intelligent paywalls come in…
Today, in the age of commoditised machine learning, many publishers that talk about “intelligent” paywalls are referring to using artificial intelligence to choose the best meter limit for each visitor. Machine learning is used to score a visitor’s propensity to buy, segmenting the visitior into high-propensity / low-propensity buckets and sending them along a journey to match. Low-propensity visitors get 10 free articles; while high-propensity visitors only get 2.
This is a far more effective meter for converting users based on their individual preferences, but the question we should now ask is, “why optimise a meter at all”? The meter brought a solution to yesterday’s technical limitations. It was always a proxy to engagement, used because the industry didn’t have any better tools of accurately measuring engagement.
Consider the journey a meter drives:
Free article 1-> Free article 2-> Free article X-> Meter Paywall 1-> Paywall 2
Let’s imagine, for a minute, that for a single visitor, at a single moment in time, there exists an optimum journey. Is it possible that journey might start with a paywall instead?
Paywall 1-> Free article 1-> Paywall 2-> Free article 2-> Paywall 3
It certainly seems like that could be a feasible subscriber journey. It establishes the perception of value and it nurtures the visitor with free content. But, it could not ever be described through the paradigm of a meter.
So even with “intelligent” paywalls that route through to a meter, publishers are still placing limitations on the journeys they can serve. And, hence, there is a ceiling on how close to optimal they can get.
Stepping Up into Dynamic Paywalls
There are a few media companies who have been able to break out of the meter-trap. These businesses have built what can be considered as truly “dynamic” paywalls. To be dynamic, these systems must make real-time decisions about every interaction. The decision for one pageview is not driven by the segment the visitor was put in last night, nor the meter they consumed 5 days ago. Live consumption information, user state, context, content metadata, and machine learning results are assessed each and every time. Using machine learning in these kinds of decisions means a paywall can be both dynamic and intelligent, and this is likely the future of effective subscriber journeys.
Ultimately, all these advancements boil down to one key takeaway: Build a system that is optimised for customers and their value-perception – not optimised for a meter or model. Taking an audience-first approach is what will continue to drive innovation in the industry.