Recently, Mike Ryan put out an interesting post at Uproxx where he dared anyone to explain the Return of the Jedi’s “rescue Han Solo” plan that happens in the first act of the film. Basically he was sitting around with a bunch of friends talking Star Wars, and he tossed out the question: if Luke’s plan had gone perfectly, then what was that plan. Which no one at that table of Star Wars aficionados could answer. It was an especially acute question, because what actually did play out on screen didn’t really make sense, when you come to think about it, as a plan. Mike Ryan said, “So let’s not try to pretend that what happened was “the plan.” There are way too many variables.”
Too many variables.
Modern business contends with too many variables, but too little data per variable. Big data may mean more information, but it can also mean more false information.
The challenge is in uncovering the unknowable.
A person that is attentively monitoring a dashboard may work for just one or a few KPIs. This manual approach, however is not scalable to thousands or millions of metrics, while still allowing for real-time responsiveness. Beyond the mere number of metrics in many businesses, is the complexity of each individual metric: different metrics have different patterns (or no patterns at all) and different amounts of variability in the values of the sampled data. In addition, the metrics themselves are often changing, often exhibiting different patterns as the data adjusts to set a new “normal.”
Nonetheless, some of the most valuable metrics to measure are also the most variable, if they can be properly addressed.
What Was Luke’s Plan?
Let’s go back to that post for a moment. Just to put this in perspective. In short, Return of the Jedi started off with C-3PO and R2-D2 showing up at Jabba’s palace and are promptly offered up in exchange for the release of Han Solo. Obviously, Jabba doesn’t take the deal, but then keeps the droids anyway. Then, our old buddy Lando Calrissian steps forward and turns out he has been hiding in plain sight as a skiff guard doing … something. Then, Leia arrives disguised as a bounty hunter, and hands over Chewie. That night, she unfreezes Han, only to immediately be caught by Jabba and his cronies. Finally, Luke shows up, fails to mind trick Jabba, kills a rancor, gets taken prisoner, and ultimately retrieves his lightsaber from R2-D2 so he can save the day, over the Sarlacc pit.
Piece of cake?
This series of missteps, failures, and dumb-luck coincidences forces us to wonder: Just what was the plan here? What if Jabba had just accepted the droids as payment? Would Luke leave them there? How was Leia planning on freeing Chewie after she rescued Han? Was Luke’s only good idea to try to Jedi mind trick Jabba, and after that didn’t work he would just improvise? What was Lando even doing?
There were far too many variables at work for one to consider all the myriad avenues of action that could occur.
Likewise, for today’s data-driven companies, millions of metrics may be collected and measured. Companies may end up with quite an impressive dataset to explore how their business is performing. Humans supervising this data typically will prioritize which metrics to track, and which to ignore. Even if there was the human resources to continuously monitor those dashboards by skilled data analysts, they would still not achieve real-time actionable insights.
Within this dataset are data patterns that represent, basically, business as usual. An unexpected change within these data patterns, or an event that does not conform to the expected data pattern, is considered an anomaly, a deviation from business as usual.
Can’t Rely on Your Gut Instincts
Companies and executives often make a big mistake when it comes to dealing with their business data, trying to ignore the data and rely on gut instincts to make decisions.
Lacking proper insights, managers and executives facing the sheer volume of data, and many more variables of possibilities, fall back on their experience rather than data. Without the data-driven business tools geared to today’s pace of business, these decision-makers risk missing key business issues that can impact revenue. Big data won’t replace human intuition, but it can complement it.
Overwhelming Variables in Creeping Thresholds and Alert Storms
Let’s look at how video advertising company, Eyeview dealt with these deviations, constantly updating static thresholds as traffic increased, then variability due to seasonality continuously made those thresholds obsolete. Stretched analyst time, that could have been spent on uncovering important business events, was instead diverted to updating thresholds and sifting through the constant flood of alerts around these thresholds – some relevant but many not.
Unable to correlate anomalies, they were unable to distinguish between a primary anomaly from another onslaught of anomalies coming in an alert storms.
Anomaly detection powered by machine learning can process multiple variables from across a range of sources, training algorithms to identify regular patterns within the datasets based on a statistical understanding of their performance. This approach can identify issues that human observation and threshold monitoring will normally miss or identify too late, due to the complexity of the signal interactions. It is good for catching both the system-wide issues as well as identifying the more subtle, local failures before they cause damage.
Not Ready for Many Metrics
Beyond the mere number of metrics in many businesses, is the complexity of each individual metric: different metrics have different patterns (or no patterns at all) and different amounts of variability in the values of the sampled data.
Identifying dirty data, catching glitches, and leveraging data assets in hyper-complex, multifaceted ecosystems is a task that stretches far beyond human capabilities. Even if multiple dashboards with hundreds of different signals could be provided, the human brain is simply not equipped to process all of them and definitely not equipped to correlate the different signals to find the root cause of an issue.
Analyzing all the Variables with AI and Machine Learning
The only way to address this issue and get insights on a granular level is by embracing new technologies with machine learning powering anomaly detection. This can process huge amounts of data in real time and surface the anomalies that matter for an indefinite number of dimensions.
Grouping and correlating multiple anomalies by design, AI-powered analytics elevates the most important insights first. An AI-powered analytics solution with anomaly detection capabilities can monitor millions of metrics to a granular level, giving both the detail and scale needed to be able to identify the business incidents that matter, including the most subtle ones, that big data dashboards would overlook and obscure. Automated anomaly detection frees talented data analysts from the futile task of trying to manually spot critical anomalies while your business is moving forward, sometimes at breakneck speed.
So whatever Luke’s plan was, and if The Force gave him the ability to consider every possible variable, today’s data driven businesses need to rely on the next best thing – AI-powered analytics.