In my previous post, I talked about how I thought the reason there has not been widespread adoption of BI/Analytic solutions is that they have not had much intelligence in them. Dennis Howlett did a post partially talking about how alerts and dashboards supported what I had talked about in terms of interrogating drill-down paths and finding "interesting results". This post will provide more color about finding "interesting results" as alerts are only a starting point.
The easiest way to understand what I mean by an "interesting result" is to use an example. Let's say that my business intelligence system had an alert set up for when Voluntary Turnover was greater than 20%. As the system interrogated the drill down paths, it found that in the Southeast Region in the U.S. the Voluntary Turnover rate was greater than 25% over the past 12 months. That is about as far as most BI/Analytics solutions go. If I manage the U.S operations, I would know that there could be a problem in the Southeast, but I still do not have a sense of what the problem might be. Again, I could drill down and around further to find the root cause, but I maintain most business leaders will not take the time to do this. It is more likely they will assign a business analyst, assuming they have one with the right skills, to do this work. They may or may not find the answer (or find the wrong answer because they did not look at enough of the possibilities).
Why cannot the business intelligence solution do some of the work for me? It could if it knew other related drill-down paths. For example, it could look at source of hire to see if there is any relationship between the source of hire and the people who have left (maybe I should look at hiring from different places). It could look to see if there are any other worker profile data similarities that might account for people leaving. It could look at the performance level to see if there was a difference between high performers and low performers in terms of retention (maybe high performers were frustated with non-differentiated treatment relative to low performers). It could look at internal mobility to see if people might be leaving because they did not have promotion opportunities (or if the manager of that region may be blocking their progress). It could look to see how much and how effective the training opportunities are for people in that region compared to other regions.
This is only a short list of possible drill-down paths that a more intelligent solution could explore. All of these examples are related to talent management, but there are paths to explore beyond HCM as well. Is there any relationship between this high voluntary turnover and sales or customer satisfaction in the region, for example (or vice versa, maybe we have some really bad customers that are causing our employees grief).
To understand the relationships between various metrics and their root causes requires significant domain expertise. This is something traditional BI providers have lacked, in my opinion. It also requires more sophisticated analytic techniques be applied where appropriate, some of which today are beyond the skills of typical business analysts. For example, training effectiveness comparisons require more than just simple data visualization. A good analysis technique in this case would be control group testing (like they do for pharma testing) to understand the effectiveness.
So, in my view, there is plenty of opportunity to put more intelligence in business intelligence solution to make them relevant to different business audiences.
Disclaimer: These are my personal views and not the views of my employer