It’s been a tough summer for Netflix.
Last month its shareholders got a nasty surprise when the streaming giant reported that it had lost U.S. subscribers for the first time in nearly a decade, and global customer growth was roughly half of what it had projected.
In the following week, Netflix shed $24 billion in value. Ouch.
Netflix has a famously tight lipped culture and somehow they managed to keep their bad news to themselves until the last minute.
But there was a small group of people, mostly congregated in New York and London, who were not surprised by the drop in users at all. Those of us who deal in alternative data – the business of predicting company movements before they happen based on outside indicators – saw this one coming.
In the months preceding Netflix’s announcement, a number of subtle signals had started to appear all over the web. Combined, they painted a picture of a stock that was about to tumble.
Here are four data types that were ideal leading indicators ahead of Netflix’s reporting.
1. Product review data
Firstly, product reviews of Netflix were steadily becoming more negative.
Based on sentiment analysis by Meltwater (which works in 16 languages) user satisfaction was on a downward trajectory. For example, early this year Netflix had 7.4% more negative reviews than it had in the previous quarter.
2. Social media sentiment
Signs on social media were also foreboding if you knew what to look for.
Meltwater picked up about 100 million social posts mentioning Netflix globally over the last year, across plenty of sources including Facebook, Twitter, Instagram, YouTube, review sites, forums, and article comments. But the posts that were really interesting to look at were on the more intense ends of the emotional spectrum.
The number of people raving about Netflix as it continued its transition from cool new technology to a mainstream product declined sharply towards the end of last year and continued to stagnate in Q4 and the early part of this year. The graph below shows how social media posts with strong positive language like “love” in reference to Netflix have declined.
As you will see in the graph below (measuring emotional reactions to Netflix from the last few weeks) people still feel overwhelmingly positive about the company, but sadly for markets that isn’t enough. Unless Netflix can find a way of re-energising its user base, it will be hard for the company to continue to post the kinds of growth figures that its investors have become accustomed to.
3. News media
Another place offering clues was news media.
There were approximately 3 million global articles mentioning Netflix in the last year, which certainly seems like a healthy number from a PR standpoint. However in order to use news data predictively, it is important to look at what the news is saying and whether it matters.
3 million articles is a lot of data to wade through, but this is where AI comes into play. Using its Signals tool Meltwater was able to use machine learning to determine the news stories that could actually be important for Netflix’s user numbers.
Take for example this unfortunate situation:
A month before Netflix’s earnings annoucement, pro-lifers decided to drive a “mass exodus” from Netflix as they didn’t appreciate the company saying it would leave Georgia in response to the state’s draconian new abortion ban.
An angry band of pro-lifers vowing to drive Netflix out of business certainly didn’t help matters at what was already a sensitive time for the company.
This news story didn’t make the global headlines (in fact it was barely mentioned in U.S. national news), and so it could easily have been missed without the right technology to flag it as a potential problem. Pro-life group Live Action is making an unverified but interesting claim that the boycott was responsible for the 126,000 lost US users that turned Neflix’s quarterly statement into such a disaster. Whether or not this is true, uncovering this type of movement at an early stage serves both as a PR catch for Netflix and as a possible indicator of a growing trend for outside observers.
4. App downloads
Of course, Meltwater were not the only ones who saw that Netlfix wasn’t doing so well. The alternative data industry was all over this story.
SimilarWeb wrote a great blog in advance of Netflix’s announcement where they pointed out that the number of people using the company’s Android app was hardly growing at all.
Ed Lavery, SimilarWeb’s Investor Solutions business manager, explains: “We saw a strong correlation between Netflix international subscribers and the number of Daily Active Users of their app. Even though Netflix users are likely to use television to access its programming, we believe that many users will download the app and log into it occasionally as well, which makes this data more predictive.”
So the fact that Android app users were flatlining indicated that Netflix’s most important metric – new subscribers – was likely to be less than projected as well.
Backing gut feel with data
What many found to be most surprising about the Netflix story is in fact quite how many people were surprised by it!
SimilarWeb published its concerns days before Netflix dropped its bombshell, and anyone monitoring with Meltwater’s platform would have seen that there were worrying trends for Netflix all over the place.
And yet the majority of investors failed to notice the warning signs.
Hugh O’Connor, Director of Data Sourcing & Partnerships at data broker Eagle Alpha, notes that “demand for alternative data continues to rise across the industry. Increasingly Eagle Alpha are fielding requests from discretionary fundamental clients which want to gain new or differentiated insight into key stock drivers for a small concentrated portfolio of companies.”
To translate for non-finance types: Hugh is now starting to see an interest in this kind of data from investors who in the past would have relied mostly on “gut feel” when choosing which companies to invest in.
But it is clear that there is still a long way to go for this kind of analysis to become mainstream.
In the meantime, those who embrace data-driven decision-making – or at least, who take the time to make sure that their instincts are backed up by external data – have a distinct advantage over those who don’t.
Hardly surprising, when you think about it.