Recurring revenue is crucial for any business, driving companies like Apple to focus on services alongside hardware. The rising cost of acquiring new customers underscores the significance of steady revenue streams. Subscription models face challenges in retaining customers, making churn reduction paramount.
Aiming to aid companies in addressing this issue, Danish startup Subsets offers an AI platform designed to bridge the retention gap in the subscription industry.
Understanding AI Decisions
Founded in Copenhagen in 2022, Subsets utilizes “explainable AI” to identify subscribers at risk of churning, providing experiments to engage them effectively.
Explainable AI ensures that AI can justify its decisions in understandable terms, contributing to the broader effort to enhance the transparency of AI systems. This approach is crucial in addressing concerns around the “black box” problem in AI.
Within six months of its launch, Subsets has attracted notable clients such as The Athletic and Danish newspaper Børsen. Building on this momentum, the company has secured a fresh $1.65 million in pre-seed funding led by Nordic early-stage VC Upfin and Y Combinator (YC), following its participation in YC’s Summer ’23 program.
Focus on Digital Media
While applicable to various subscription businesses, Subsets currently concentrates on digital media.
CEO Martin Johnsen mentioned, “Our current focus is digital media, including publishers, streaming platforms, subscription apps, and telecoms. We plan to expand to other digital consumer subscriptions like mobility, banking, and food delivery, as well as offline subscription categories, as digitalization advances.”
Subsets integrates with companies’ internal systems to collect subscriber data points for analysis, enabling insights on user behavior and preferences. The platform offers a web app where commercial teams can assess churn audiences through user-friendly visuals and natural language explanations.
Subsets: Audience discovery Image Credits: Subsets
Subsets empowers non-technical teams to conduct retention experiments on subsets of subscribers, experimenting with various methods to encourage customer retention. These experiments could involve notifications, discounts, or feature upgrades, tailored to each customer’s needs.
Subsets: Experiment Image Credits: Subsets
The successful retention strategies from these experiments are automated, assisting companies in streamlining their retention efforts based on proven results, reducing reliance on guesswork.
“Some experiments yield significant improvements in subscriber retention, leading to automation of successful strategies,” said Johnsen. “Subscribers exhibiting specific churn-inducing behavior receive these effective strategies as they join the identified audience.”
Subsets: Results Image Credits: Subsets
Technical Foundations
With expertise in mathematical modelling and computation, Johnsen noted that Subsets utilizes AI algorithms based on gradient-boosting models with temporal sequencing methods. These techniques enhance predictive accuracy by combining predictive models and considering time-related features.
Subsets leverages frameworks from Elon Musk’s xAI to make churn-driving behavior more comprehensible. While not currently using OpenAI’s GPT.x models, the company plans to incorporate downstream functionality from ChatGPT in its product.
Recent social media updates by YC’s CEO Garry Tan suggest a growing adoption of large language models (LLMs) among YC companies. Subsets benefits from YC’s support, given its strong network and connections, including successful partnerships with YC alumni like The Athletic.
Aside from Upfin and YC, Subsets’ pre-seed funding round attracted investments from various institutional and angel investors, further fueling the company’s growth.
