Readly is a digital subscription service that provides people who want to be informed, inspired and entertained, unlimited access to thousands of magazines, newspapers and podcasts directly in one app. The company was founded in Sweden and is today the market leader in Europe in digital magazines and newspapers, with a range of over 7,500 titles in 17 languages. It cooperates with 1,200 publishers worldwide and has users in more than 50 countries.
From aggressive growth to increased profitability – Readly optimized its marketing investments with predictive LTV
Together with Solita, Readly developed a new way to optimise its marketing budget and allocate the money to where it leads to most efficient subscriber growth. The result was a more profitable growth strategy and a smarter way of working that was instrumental in adapting to the tougher economic climate.
Higher Growth Efficiency and turnaround in several markets
Changed operational steering based on unit economics (LTV/CAC)
Significant LTV/CAC ratio uplift globally through market/channel adjustments after 6 months
+40-60 % improved channel profitability within key channels
New predictive LTV dashboards
New User Enagement framework & dashboards as decision making foundation
LTV implemented as a MLOps process to ensure automated monitoring, training and deployment of models
Readly has been a growth company for a few years, with very aggressive growth targets. We had always measured and used LTV as a metric, but the optimisation for our spend and marketing was largely based on acquisition cost and subscriber growth targets. Especially during the last two years, we were focused on double digital annual subscriber growth and operated as a growth lead company, so we invested heavily in our user growth across our markets.
Marie-Sophie von Bibra Chief Marketing Officer (CMO), Readly
Shifting from maximising growth to profitable growth
Many growth companies are adapting to the changed economic circumstances through trimming costs and increasing efficiency with the goal of building a strong competitive position for the future. This is also critical for Readly.
– We shifted our focus from pure growth to sustainable growth to be better positioned to reach our financial targets, including reaching a positive EBITDA latest in 2025. This meant that we also had to look at the framework our teams operated in, including changing LTV (Lifetime Value) from a lag to a lead metric that we could use for optimization and that would guide our marketing and strategy work. We wanted to move from acquisition cost-driven growth to a unit economics-driven way of marketing – an evolution and somehow also a complete 180 turnaround, says Marie-Sophie von Bibra.
Lifetime Value as a lead metric
Readly invests a lot of money in marketing and optimizes its marketing investments in correlation to the predicted LTV, Lifetime Value, i.e. the value that a customer generates during the time that the customer relationship lasts. LTV is used to calculate the value of a customer, especially in industries where you cannot calculate the value of the customer relationship by means of the customer’s first transaction, such as subscription services and apps for example.
The challenge is that customers have different LTVs per market and channels and there is a risk that the acquisition cost exceeds the value that the customer generates. This has a negative impact on unit economics. The problem is that it is difficult to know in advance which customers will be profitable and hence which markets and channels are expected to have higher unit economics.
This is a challenge that a lot of companies have to deal with. What we did together with Readly was to build a model for predictive LTV. Rather than waiting months to follow up on the profitability of marketing investments, we could predict the unit economics and expected profitability for the marketing investments just a few days into the customer journey. Being proactive rather than reactive is essential when both experimenting and optimizing marketing investments.
Niklas Liedholm Consultant at Solita and responsible for the project
The team built an advanced machine learning solution that is able to predict the value of a new subscriber based on their user engagement with the service. The model proved to be accurate in predicting LTV already a few days into the customer journey. Accuracy is however only one dimension in these types of projects. More important is to find a balance between model performance and the utility of the model and how that can be integrated into decision-making.
Increased profitable growth rate through enhancing decision making
The goal of establishing predictive LTV was to increase growth efficiency. This involved enhancing decision-making in three areas: allocating spend to markets, channels & campaigns with the best-expected return, evaluating new channels and partnerships more quickly, and improving the performance of individual channels. Predicting LTV was just one aspect of achieving these goals. Ultimately augmenting decision-making required different descriptive, predictive and prescriptive solutions that together built and showcased the forecasted payback times per market and channel in a way that could be efficiently incorporated into the everyday workflow of the employees.
– The expected return and unit economics for each market and channel were predicted by combining the total cohort LTV with the spend per market and channel. It was also important to predict the payback time. This was achieved by building an LTV model that forecasted the monthly expected gross profit for each customer. This model also improved monitoring and enabled forecasting of cash flow buildup, providing new means to invest in user acquisition, says Niklas Liedholm.
The result turned out to be very successful.
Predictive LTV has totally changed how we can predict, plan and evaluate the individual channel strategies and our marketing strategy as a whole. It has added a completely new and very important layer to our marketing strategy. We are still using all our metrics from before, and also all traditional marketing metrics, but now we have this layer of projected LTV and respective unit economics per market per channel per week which has been a game changer.
Marie-Sophie von Bibra Chief Marketing Officer (CMO), Readly
Business impact through data-informed ways of working across departments
An important part of the projects has been to implement a new way of working and make it part of the employees’ everyday life. At Readly, they have ensured that all teams, not only the marketing team but also the finance and product teams, have been integrated and use the same decision-making basis and metrics. This way they are able to make decisions fast and optimize towards the best-performing markets and channels every day. It has made it clearer to all employees that they are working towards the same goal and made it easy for them to see and understand how their daily decisions contribute to the business goals.
– This new level of projected LTV has enabled us to connect marketing and product metrics with the overall business metrics to a much greater extent, and therefore enable better business decisions in line with our overall business goal of sustainable and profitable growth. We have seen true business impact for example by prioritizing our markets differently based on the new insights, changing our channel mix across markets and evaluating channels and tactics more holistically. Now we have been able to show good progress in line with our financial goals and that is the path we are planning to continue on, says Marie-Sophie von Bibra.
- Combination of descriptive, predictive and prescriptive solutions
- AWS based MLOps solution, managing hundreds of models through automated training, deployment & model monitoring
- MLFlow, AWS ECS
Our work – See more cases
KONE Predictive maintenance gets more intelligent with advanced analyticsKONE moves more than a billion people every day. Together with Solita, KONE is enhancing the predictive maintenance capabilities for more than 150,000 elevators with sensor data, machine learning and analytics.
Ponsse Edge computing starts new era of intelligence in forest harvestingCollecting data from Ponsse forest machines for better services.
City Traffic Ltd. Real-time data analytics injects efficiency and punctuality into metro trafficCity Traffic Ltd. (Kaupunkiliikenne Oy) wanted to improve metro traffic control in the Helsinki Region with data analytics and provide the traffic controllers with a new tool for monitoring transit regularity and punctuality.