The problem
Recurring revenue is crucial for any business. That’s why Apple, Microsoft, and many other big names have started focusing on services in the last decade. For instance, 20% of Apple’s earnings now come from Apple TV+, Apple Music, and iCloud.
Simultaneously, the cost of acquiring new customers has increased by more than 200% over the past decade, underscoring the importance of predictable revenue streams. However, subscription-based business models present inherent challenges. Companies must not only focus on increasing the number of people willing to pay for a given service, but they also need to retain existing customers.
Subscription businesses lose 20-40% of their customers annually, and they lack automated initiatives to ensure continuous engagement for existing subscribers.
Modern CRMs do not have solutions available to determine what the subscriber needs right now to make the most out of their subscription.
So, how can churn be reduced effectively? This startup has an answer. They developed a service that helps companies lower their churn rate based on objective data. A few days ago, they received their first funding of $1.65M.
Image Credits: Subsets
The solution
Subsets allow non-technical teams to run retention “experiments” on subsets (hence the company name) of their subscriber base, to see what actions might lead a customer to stay on board. These experiments might be a series of push notifications or emails offering a subscription discount, or perhaps a free upgrade to unlock new features. The specifics of these “retention flows” can be tweaked by each customer.
Image Credits: Subsets
The steps discovered to decrease churn during the experimental phase are showcased as successful results. This method aims to eliminate some of the uncertainty from a company’s retention efforts, enabling the company to “automate what works.”
Experiments yielding positive results on subscriber retention can subsequently be automated. Given that an audience is characterized by specific subscriber behavior that prompts churn, these audiences are usually dynamic, with new subscribers continuously entering and leaving. All subscribers joining an audience receive the flows that have previously demonstrated positive outcomes.
Image Credits: Subsets
Technical details
Subsets’ AI algorithms are based on gradient-boosting models with temporal sequencing methods, allowing them to analyze subscriber behavior over time. Gradient boosting is an old and well-functioning classic Machine Learning algorithm used to make “yes/no” decisions in many data analysis tasks such as user recommendations or spam mail detection.
In the given case, Subsets takes many user-related signals for input and makes a balanced decision on whether the user is going to leave or not.
While currently, Subsets doesn’t use “real AI” frameworks, they plan to utilize OpenAI’s GPT.x models in the near future.
The types of signals Subsets uses to identify the potential of a user to churn are:
Usage Data
Information on how frequently and in what ways customers engage with the service, including login frequency, feature usage, content consumption patterns, and any changes in usage behavior over time.
Customer Interaction Data
Records of customer interactions with the service, including customer support inquiries, responses to marketing campaigns, feedback provided, and participation in community or social features.
Subscription Data
Details about the subscription itself, such as subscription type, duration, renewal dates, payment history, and any changes to the subscription plan.
Demographic and Behavioral Data
Information about the customer, such as demographic details, and inferred behavioral segments based on their activity and preferences.
Engagement Metrics
Metrics that indicate the level of engagement a customer has with the service, like session length, frequency of visits, and engagement with emails or notifications.
Experiments
Once identifying the risk of churn Subsets offers to perform various experiments to keep the customer stay subscribed.
These experiments involve running retention flows on subsets of the subscriber base, such as push notifications, email offers, or free upgrades tailored to each customer. The results of these experiments are presented to automate successful strategies.
A business value
The platform aims to make churn-driving behavior understandable and automate successful retention strategies.
Image Credits: Subsets
Subsets claims that businesses using its platform can see significant benefits, such as a 20% increase in lifetime value (LTV), an extension of subscriber lifetimes by up to 6 months, and cost-effectiveness, with the platform being 6 times cheaper than the cost of acquiring new customers (CAC).
The company
The company was founded in 2022 and is based in Copenhagen, Denmark. Its founding team includes Martin Johnsen, Oliver Brandt, and Nikolai Skelbo, who come from diverse backgrounds including management consulting, machine learning engineering, and revenue operations.
Subsets is the resident of YC S23 program.
The funding
Subsets received pre-seed funding of $1.7M from Upfin (https://www.upfin.io/) on February 27, 2024.