How INSIGHT drives Stranger Things’ success
Last time we looked at how data is at the heart of Netflix’s success in helping them to predict successful movies and TV shows. If you missed that post, you can read it here.
As we move through the DATA > INSIGHT > ACTION loop, this blog post deep-dives further into the INSIGHT aspect of Netflix’s data.
We have already established that Netflix collects A LOT of data. Storage of that data so that it is readily accessible for analysis is a challenge. In order to do this, Netflix has to use a Data Warehouse combined with a Data Analytics tool to create a Single Customer View, measure KPIs and visualise the data to generate insights.
How does this insight maximise Netflix’s return on investment?
For Netflix, ROI means achieving the maximum customer happiness and loyalty per dollar spent. To maximise user happiness, Netflix has to continually provide really relevant content to its subscribers’ interests. However, with a price tag of $6 million per episode, totalling $45 million for the Stranger Things series, getting it wrong, could be a very expensive mistake to make. But Netflix has ensured that their spend has been successful through using historical data to predict subscriber satisfaction.
Their Machine Learning and Data Science approach to identify high potential content is so successful that, compared to the rest of the TV industry, where just 35% of shows are renewed past their first season, Netflix renews 93% of its original series.
To measure users’ happiness, Netflix uses various and complicated algorithms. This enables Netflix to quickly get quantitative answers (insights) to specific questions and to predict viewing rates for Stranger Things. When Netflix’s decision-makers were reviewing the Stranger Things project, I assume they had concerns such as:
Netflix will input those concerns along with some requirements into a smart algorithm. This algorithm will produce a list of indicators that helps Netflix’s decision-makers understand if the show would be a good investment. Going forward, using machine learning and data science, the system will also be able to share recommendations about how Netflix can improve the existing predictive score and maximise their investment.
For instance, for Stranger Things it could be:
Using powerful analytical tools such as Looker, Power BI and Tableau Netflix can investigate the recommendations that have been generated. Through manipulating the data and creating visualisations insights can be produced at a very granular level.
The main goal from this activity is to find a way to minimise the risk as well as define the action plan before and after making the decision to buy into a TV series or movie.
Predicting successful content is not the only area where machine learning and data science is being used to drive more outcomes as you can see below:
In our next Stranger Things post we complete the DATA >> INSIGHT >> ACTION loop and look at how Netflix puts the initial two stages of the process into action.