Eyes on the road, not dashboards

Car dashboards are simple visual indicators of a complex machine with many parts that performs a high-stakes task in a context of many overlapping, conflicting rules and goals: personal convenience, safety, minimum travel time, courtesy to other drivers and so on. The speedometer, perhaps one of the most important indicators, serves one key purpose: to show how fast you’re going, mainly so you avoid getting a ticket. Often the biggest indicator on the display, its job is not to teach the physics of motion or of many-body collisions (although if you stare only at the speedometer, you will certainly gain firsthand experience in the latter), but to give you just enough information to remain compliant with one of the main rules of the road.

Vehicle dashboards are not in-depth analytic tools that beckon deep, prolonged interaction. Their main job is quick indication through simplicity, because what really requires your attention is the road, weather and other drivers around you. When something’s wrong and your car is at the shop, that’s when in-depth diagnostics and analysis can occur.

The Devil is in the Details

There are many similarities between traditional business intelligence (BI) dashboards and car dashboards. BI dashboards are simple to a fault since they provide overall summaries of important metrics. By simplifying the massive complexities of companies – especially large ones – they provide a manageable stand-in for its performance. Yet running and growing a company is also a complex, high-stakes endeavor with many conflicting rules, goals and requirements. These include complying with labor laws, reducing costs, increasing profits, maintaining capital, investing for future success, responding to competitors’ moves in the market, among others.

By design, dashboards are meant to be checked occasionally. However, frequent checking is useless since the data visualizations on them (especially KPIs) are not updated in real time. Your real attention should be on the finer details of your business: developing and launching new goods/services, attracting top talent to fill your open positions and more.

KPIs, like cash on hand or monthly total revenue, mainly serve to prevent you from running out of the money necessary to pay your taxes, vendors and employees. The finer details of day-to-day business operations require analytics that can find business incidents in the minutiae of your data.

Specific Decisions Require Specific Data

As real-time analytics and anomaly detection company Anodot believes, big data dashboards are not up to the task of providing the real-time actionable insights that businesses need to stay competitive. Their main flaws include being so general that they can overlook isolated but important business incidents, they don’t actually detect anything (unless static thresholds are applied and constantly maintained), and the business incidents they do (indirectly) detect are delayed.

Business incident detection requires anomaly detection, which in turn requires conciseness in order to be usable and effective at scale. And modern machine-learning-powered anomaly detection systems can finally meet the needs that traditional BI dashboards never will. By correlating and condensing anomalies in related metrics, these solutions can point to the problems behind the alerts, and do so in a way that doesn’t overload your analysts.

Machine learning can reveal empowering insights hidden in your data. Just like any other discovery tool, you may not like what it reveals, such as when an image-recognition software built by a college CS professor discovered and then greatly amplified the gender bias in the large collection of images it was being trained on. Since artificial intelligence must learn from natural intelligence, the screen often becomes a mirror as we see our own flaws exaggerated.

Focusing on What Matters Most

When it comes to all the metrics you’re collecting, each detected anomaly is an opportunity to save or earn more money. Each anomaly is an opportunity to cut losses and even outmaneuver a competitor. Data can drive decisions. Bad data, or data that is badly analyzed, can drive you right into a crash. The real-time actionable insights provided by AI analytics can give you the skill and speed of a race car driver, while allowing you to see the path ahead as the majority of your attention is focused on the road. Coordination, timing, planning: This is what driving really consists of, not staring at dashboards and trying to make sense of it. The same applies to big data dashboards.