It turns out, the most crucial factor in user retention is not the “usefulness” of the service… but rather the “habit” of its use! Therefore, to reduce user churn, there’s no need to repeatedly convince them of the usefulness of your service 😉
It’s better to create a platform capable of automatically identifying and incentivizing subscriber habits. Then, such a platform can be sold to any subscription service. Paid blogs and online publications, in particular, will be keen on it, as they experience the highest churn rates 🙁
There are already 600 million blogs on the internet, and by 2028, their number is expected to increase by another 40%! So, it’s a vast and growing market where one can enter with their platform right now—taking into account such nuances:
The essence of the project
Subsets helps internet publications (magazines, paid blogs, and content newsletters) reduce the number of subscribers dropping off.
The problem is that subscription content services, as the startup claims, lose 20-40% of subscribers annually. However, the owners of subscription services or the people responsible for their business poorly understand what can be done to reduce this churn.
They say AI can help understand this 😉 But actually using AI technologies to analyze this issue is currently almost impossible without involving AI-savvy programmers familiar with publishing platforms.
Subsets intends to solve this problem by creating a platform to reduce user churn, which:
- Seamlessly integrates with the platforms already used in the publication,
- Can be used directly by business employees without involving programmers.
The first function of Subsets is to provide “understanding.” The AI engine of the platform can detect key factors in subscriber behavior that can predict whether they will continue to use the subscription or drop off soon. The AI engine can identify several such factors, each with a different level of influence.
Moreover, different factors can affect different subscribers differently. Therefore, the AI engine will segment the entire subscriber audience into different subsets—on which a similar combination of different factors has a similar effect.
Furthermore, the AI engine will also propose for each resulting segment its own set of actions that the resource owner can take to reduce subscriber churn from that segment.
Each of these strategies can first be tested manually—by performing the proposed actions once on a small percentage of subscribers from a specific segment and comparing the results with a control group of “untouched” subscribers from the same segment.
If the result favors the strategy, then the platform can be instructed to start executing the strategy automatically.
Meanwhile, the AI engine will continue to analyze user behavior and the effectiveness of the strategy. As a result, it will inform the resource owner of the emergence of new churn factors and the fact that the old retention strategy has stopped working—regularly suggesting new strategies adapted to the changed situation.
Subsets is based in Denmark; last summer, it joined Y Combinator, received the due $500,000 investment, and attracted the first clients from Danish internet publications to the platform.
The initial results were promising—subscriber lifetime increased by 6 months, the increase in Lifetime Value (LTV) over the subscriber lifetime was 20%, and the cost of retaining a subscriber turned out to be 6 times cheaper than acquiring a new subscriber.
The cost of retention and the disproportionate increase in subscriber lifetime contributing to the increase in LTV are explained by the fact that the platform offers, as almost a basic retention strategy, a discount offer for subscription renewal—like “get 20% off if you renew your subscription for another 3 months today.” This discount is the cost of retention, which also reduces the “standard” profit per subscriber.
Unfortunately, there have been no recent updates from the startup in the past few months—which seems quite disappointing because, in my opinion, they delved into a rather promising topic. Why do I consider it promising?
What’s interesting
In recent years, many have become internet writers. It all starts with innocently posting content on social media about topics that interest you, then you rejoice in the increasing number of followers… and then you realize that “better than fame can only be money” 😉 And you start thinking about how to monetize this.
Some limit themselves to posting paid advertisements on free blogs, while others create paid blogs. But both are important to ensure that as few subscribers as possible drop off.
The increase in such “worried” individuals grows as the overall number of blogs increases.
Just in the United States, from 2014 to 2020, the number of bloggers increased by 10 million people. As a result, there are already 600 million blogs on the internet, and 6 million new posts are published in them every day. But this is not the limit— it is expected that the number of bloggers will increase by another 40% by 2028!
And a significant part of them is concerned with the problem of subscriber churn, making them the target audience for platforms similar to today’s Subsets.
However, the problem lies in the fact that in the media and entertainment sector, the highest percentage of churn among paid subscribers is 20–30% annually. For comparison, cloud services (SaaS) lose only 5–7% of paid subscribers. These are the statistics from the payment system Stripe, which has every reason to claim this.
In other words, the problem of subscriber churn is most critical for the vast blogger market.
The catch, however, is that the most crucial factor ensuring user retention for digital products is not “necessity” but “habit.” This is hundreds and thousands of times more true for media services, which, on average, provide their subscribers with non-essential content 😉
Thus, fostering the habit of regularly consuming one’s content is the most effective method of reducing subscriber churn for any blogger and internet publication!
The problem is that different users may have very different habits. Let’s take, for example, our favorite fast founder.
Some read new reviews daily instead of ritually browsing social networks.
Some read all new reviews once a week after receiving regular newsletters announcing new articles on Wednesdays.
Some do it once a week too, but on their day off, for example, on Sunday.
And some don’t read for months but get hooked on reading accumulated reviews for half a day at some point. At the same time, they “do not doubt the value” and have been subscribers for over a year.
And these are just the patterns of habitual behavior of loyal subscribers shared with us in the chat. And surely there are other options.
This naturally leads to the idea that you cannot choose a single strategy for reducing subscriber churn by fostering the habit of regular reading of media content. For example, you can’t force those accustomed to reading once a week or in sporadic bursts to read new articles daily. This either won’t work or will become annoying.
What’s even more important is being able to recognize signs of some familiar behavioral pattern in each new subscriber to start “pushing” them towards it. And if we haven’t seen such a sign yet—start “pushing” different patterns one by one and see which one has a chance of catching on with a particular user.
Analyzing all this manually for each subscriber is not fea-si-ble! Therefore, the owner of a media resource cannot do without a special platform—that will automatically analyze subscriber behavior, identify habitual behavior patterns, and encourage subscribers to follow one of them. While clearly delineating different segments of subscribers with different patterns!
And this, judging by the description, is exactly what today’s Subsets has started to do, even if it hasn’t described it that way 😉 So I called this topic promising.
Where to run
Indeed, this is the direction of possible movement—creating platforms to reduce subscriber churn for media resources, utilizing AI to identify and nurture habits of reading them.
The key point is the ability to select the most suitable pattern of habitual behavior for each subscriber!
By the way, a similar approach can be seen in the startup OfferFit, which I wrote about in November. They use it not to cultivate habits but to tailor the most suitable offer for each client in terms of formulation, optimal delivery channel, and even the optimal time for sending it. OfferFit raised $39 million in investments, $25 million of which came in a single round in October last year.
So, the approach itself is quite effective. It just remains to learn how to apply it to the vast and growing market of content bloggers and other internet publications.
About the Company
Subsets
Website: subsets.com
Latest Round: $500K, 01.06.2023
Total Investments: $500K, rounds: 1