By Kris Tait, Managing Director at Croud
With audiences spending on average $47 a month on streaming services, the likes of Amazon, Netflix and Disney+ are having to work hard to maintain their pride of place with viewers. New kids on the block like HBO Max and discovery+ are swooping in with appealing offers but audiences are already maxed out on multiple services, turning this into a marketing battle to stay in the subscription mix and win the streaming wars.
Attracting subscribers through high-quality and original content certainly helped the big players build a strong lead, but production delays due to the ongoing pandemic have forced platforms to reassess their strategies. To stay competitive, streaming services need to think more long-term, approaching customer acquisition in a data-led way, not just relying on short-term boosts around content.
With millions of subscribers, and viewing time increasing up to 60% during the pandemic, tailoring efforts to each and every prospective customer is a hard task. But streaming services will want to ensure they are focusing their efforts on the customers that are the most likely to convert into loyal, long-term subscribers in order to drive profitable revenue growth.
So, how can streaming services begin to create a lifetime value (LTV) strategy?
Unifying customer data
The first step is to create different customer segments based on their value. To figure this out, you can use sources that are already available, be it looking at current subscriber data to profile users, search terms or the various ways users interact with the content and shows. Effectively combining data from those already subscribed and engaged with the service and matching this up with a history of site and marketing data will help streaming platforms gain a single customer view and inform an LTV calculation.
Data preparation and modeling
Having the data is great, but what you do with it is the key; data is crucial for decision-making but can be hard to sift through, so you can get processing services like BigQuery to do the hard work.
Unifying and sifting through data can be a daunting task for services like Netflix, which has more than 200 million subscribers globally. Such established complex streaming giants will likely have invaluable data stored in multiple places across different systems. Utilizing an agency partner to help prepare and model this data will allow for it to be used most effectively. Agencies can help by utilizing data processing services that use machine learning to prepare data and build models.
Once this data is modeled accurately, it’s possible to join up the dots of a customer’s engagement journey and create models to understand the drivers of customer LTV. We may want to find out, for example, if a customer is promotion-driven, or if we see a certain location having a higher predicting factor of LTV.
Understanding the drivers of LTV
Once you have the data in a good place, it’s possible to create a model that accurately predicts the LTV of a new customer, but on its own, this is not all that helpful. We need to understand how these predictions are made to understand the audience and tailor strategies from there. Using analytics to look at features of audience members such as their user type, location, device type and loyalty, it’s possible to see what high-value audience members have in common and adapt strategies accordingly.
Understanding customer clusters and audiences
Audiences are made up of diverse individuals and it’s important to keep this in mind while assigning LTV. The limitation of most customer profiling is that they include only a few dimensions and are therefore rather static. Modern audiences are more diverse than ever, and so a better approach is to use a multi-dimensional clustering approach, from the data. Here we can look at the shows customers are watching, location, recency and the marketing channel that delivered the sign-up. An example cluster could be ‘prefers to watch Horror’ or ‘shows little response to promotions’. This kind of insight means we can adapt emails or ads to this genre or create landing pages that are in line with this behavior.
Activation in audiences and bidding
The final step in the process is then using machine learning to automate the process of predicting LTV in real-time and finding new audience segments. By using an online service like Google Cloud Platform’s AI Platform Prediction, we can host the model for accurately predicting LTV. When a new customer converts, their data will be processed and given a predicted LTV, helping build audience lists from our customer clusters. This real-time data allows marketing strategies to be adjusted in-flight, and the speed of this process will help aid lookalike targeting to gain subscribers that are high-value.
Short-term subscriber gain is easy, but in order to deliver the right customers, who stay for the long-term, streaming services should look within. To win the streaming wars, great content will always be the number-one pull, but focusing on data quality, LTV modeling and smart activation will mean platforms maintain strong business growth as the battle continues to rage.