Last week I attended and made a presentation at the Chicago Predictive Analytics World (PAW) conference. Based on a combination of conversations, presentations, and observations, the message I came away with is that there is a pressing need to allow users with deep business understanding, but "human level" computer skills to develop Business Intelligence (BI) solutions that address real business decisions that are based on predictive analytics. I found this very reassuring, since this is exactly what we are striving for with the predictive analytics capabilities we are building into forthcoming Alteryx Strategic Analytic releases.
The case for enhanced usability was most strongly argued by Bruno Aziza of SiSense (who was formerly the BI director at Microsoft) when he declared that BI was broken, with companies spending an estimated $7 billion on business intelligence and predictive analysis tools, but with user adoption rates of only about 30% within organizations that have purchased these tools. Furthermore, he argued that most companies do not have a handle on their "small data," to say nothing of "big data" due to the low user adoption rate of BI and Predictive Analytic (PA) tools. Consequently, improving usability of these tools (a core strength of Alteryx) is a key issue moving forward.
The need to more closely tie BI and PA tools to the business decisions organizations face was most strongly argued by James Taylor of Decision Management Solutions. He makes a compelling argument that the way to do this is to focus on the business decision you are trying to improve, but linking key performance indicators (KPIs) to these decision, and then determine how predictive analytics can be used to better inform those decisions. Through the use of PA samples in the Alteryx platform, and vertical industry packages in subsequent releases, we are aiming to guide users through the process of going from KBIs to business problems to predictive solutions.
The Alteryx Western Region VP of Sales, Nicole Cachia, who has been involved with the PA industry for a number of years, observed that there has been a shift on the part of vendors away from BI/PA stacks and specialized PA appliances that address a very narrowly defined problem, and towards more user friendly, general tool sets. This is very consistent with the needs mentioned by presenters at PAW. The Alteryx intuitive, one-stop (what I like to call "tip-to-tail") approach that goes from initial data preparation to final solution implementation is well positioned to meet the needs of users with respect to ease of use and the ability to address specific business decisions.
Dr. Dan Putler
Data Artisan in Residence
