Recurring revenue is almost the holy grail for any business. This is exactly why Apple has pivoted to become a services company as well as a hardware company, with 20% of its revenue now coming from Apple TV+, Apple Music, and iCloud. Among various other subscription-based and non-subscription-based services.
The cost of acquiring new customers has also reportedly increased by more than 200% over the past decade, highlighting the importance of predictable revenue streams. However, subscription-based business models have their own challenges. Companies should not only focus on increasing the number of people willing to pay for a particular service, but also be mindful of retaining the ones they already have. Reducing churn is key.
This is what Danish startup Subsets is trying to help businesses achieve with its AI-enabled platform designed to “close the retention gap in the subscription industry.”
explainable
Founded in Copenhagen in 2022, Subsets relies on so-called “explainable AI” to tell businesses which current subscribers are looking to switch, and to “experiment” to find the best way to retain them. ” is provided.
Explainable AI is a concept that describes the ability to “explain” or justify decisions and insights provided by AI in a way that humans can analyze, and is a fundamental part of broader societal efforts to make AI more trustworthy. It has emerged as one. Many of the major large-scale language model (LLM) providers are trying to appease their critics by addressing the much-maligned “black box” problem.
Six months into its launch, Subsets has already amassed some decent customers, including The Athletic, part of the New York Times, and Danish newspaper Borsen. But the company has now raised $1.65 million in new pre-seed funding led by Nordic early-stage VC firm Upfin and Y Combinator (YC) following Subsets' participation in YC's Summer '23 program. They are looking to double their recent momentum.
media business
The subset can be applied to any subscription business, but is currently focused on the digital media space.
“For now, we are limiting our use cases to digital media. This includes digital publishers and streaming.” [platforms], subscription apps, communications, etc.,” Subsets co-founder and CEO Martin Johnsen told TechCrunch. “Subsets can later be applied more broadly across a consumer's digital subscriptions, for example, we think the scope will expand to mobility, banking, food delivery, etc. As everything becomes more digital, offline subscriptions There are also more tion categories.”
Companies connect subsets to various internal systems such as CRM (customer relationship management), CMS (content management system), billing, and data warehouses to collect subscriber data points. This includes demographics, how each subscriber spends, etc. Specific parts of the content (time, category preferences, etc.) and other relevant items related to the product and how the customer interacts with the product.
On the front end, Subsets provides a web app that allows commercial teams to delve into their “abandonment audience” through easy-to-understand visuals and natural language. For example, you can focus on thousands of subscribers who were previously interested in a myriad of different titles under a particular media brand, but are now losing interest.
Subset: Audience Discovery image credits: subset
Subsets allow non-technical teams to run retention “experiments” on a subset of their subscriber base (hence the company's name) to see what actions are likely to lead to customer retention. You can check. These experiments could be a series of push notifications or emails offering a discount on a subscription, or a free upgrade to unlock new features. The details of these “retention flows” can be adjusted on a customer-by-customer basis.
Subset: Experiment image credits: subset
Any steps that were found to reduce churn during the experimentation phase will be presented in the form of results showing what worked. It is designed to take at least some of the guesswork out of a company's maintenance efforts, allowing companies to “automate what works.” ”
“Some of these experiments have very good results in terms of subscriber retention. The ones that have good results are automated,” Johnsen said. “Audiences are defined by specific subscriber behaviors that cause churn, so these audiences are typically dynamic, with new subscribers coming in and out of your audience. Receive flows that are proven to deliver positive results.”
Subset: Results image credits: subset
connection
Johnsen, who has an education in mathematical modeling and computation, told TechCrunch that the company developed its proprietary AI algorithm using “gradient boosting models with time-sequence techniques.” Gradient boosting essentially combines multiple “weak” predictive models to create a single, more powerful predictive model, whereas temporal sequences incorporate time-related features into the modeling process. Subset then uses Elon Musk's xAI framework to help it “understand the behaviors that drive churn,” as Johnsen puts it.
Subsets is not currently using OpenAI's GPT.x model, but is working on building some additional “downstream features” into the product using the same basic model that powers ChatGPT. Johnsen said.
in a social media post This week, YC President and CEO Garry Tan said that about half of all YC companies are “using some form of LLM,” and this is certainly the latest addition to a subset. Consistent with investment. Securing his YC as an early backer has also proven beneficial to Subsets' growth plans, given YC's historical ties. In fact, Subsets' client The Athletic is also a YC alum of hers, having graduated from the Summer '16 program six years ago before being acquired by the New York club. times.
“YC has a very strong network, which has given us some exciting opportunities,” Johnsen said.
In addition to lead backers Upfin and YC, Subsets' pre-seed round included investment from a number of institutional and angel investors, including Cuesta Labs. Sandhill Market; and Peakon founder Philip Chambers.