The Fine Line Between BI And BS

February 2013 Integrated Solutions For Retailers

By Matt Pillar, editor in chief

I’ve been talking about retail BI (business intelligence) a lot lately, and I’ve been writing about it even more. Much of the conversation is with BI solutions providers. What a diverse cottage industry this big data conundrum has born.

On one hand, there’s the old-school analytics engine powerhouses like SAS, who have been around since the new school of hipster BI startups were wetting the crib. Of course, many of those young hipster BI CEOs probably cut their teeth on paychecks stroked by Dr. Goodnight, so there’s at least consistency among the “purist” BI folks, old and new.

Then there are the ERPs (which has become code acronym for we want to control your entire technology infrastructure) and, perhaps worse, the point solution providers who fancy themselves Big Data big shots.

Consider the last time you sat through a software product demo that didn’t either lead or conclude with a sweet-looking dashboard full of pretty pie charts and line graphs that promised to tell you everything you didn’t realize you needed to know. You can’t remember what a software sale felt like before dashboards, can you?

In all my discussions with analytics vendors, I’ve learned a couple of things that I’d like to pass off to you in the form of free advice.

1. If the BI/analytics package doesn’t have workflow tools, don’t buy it. When I say any dashboard package should have workflow tools built in, I don’t mean you should be able to right click on a particularly alarming pie chart, open up a new message in Outlook, drop it into the attachment field, then sit at your desk thinking for 20 minutes about who to send it to and what to tell them to do about it.
   
   I mean you should be able to share the data with a set of stakeholders, predetermined by role, within the application. In turn, the response of those stakeholders should be transparent within the application. And, if at all possible, integration with the very systems you’re measuring and analyzing should suggest — if not automate — responsive action to the data.
    
2. If you get bored within an hour of playing with it (and you will if it doesn’t have predictive modeling), don’t buy it. Unlike infants, we’d rather play with the toy than the packaging it came in. If static data presented in pretty colors can hold your attention for a little while, an interactive predictive modeling layer based on real data will enthrall you for days.
   
   More importantly, when you can slide the scale around and drop in your own KPIs and goals to see their impact on sales, inventory, labor, finance, etc., you come away with more than glassy eyes and drool on your chin (common symptoms of dashboard addiction). You come away with creative, fact-based ideas and the inspiration to initiate them.

I’ve shared these two simple suggestions with you because I feel strongly about them. I’ve shared them for free because I feel so strongly about them, they now seem obvious to me. And it just wouldn’t feel right to expect remuneration for something I think you should already know. Workflow tools and predictive analytics — two fine points that separate BI vendors from BS vendors.