Quick! Answer this question: Can you do the work of an accountant?
Now, before I tell you why I ask and why you answered the way you did, humor me and ask yourself a few more questions: What made you think that you could or couldn’t handle that effort? Do you have any facts to back that opinion up? And, how certain are you of your opinion?
Another great question to ask is this: If I presented information to you that showed your opinion to be incorrect, would you question my data or would you question your conclusion?
As it turns out, in my oh so scientific survey, most people see accounting as a necessary evil which is fairly simple at its core and can be done by anyone who just sits down long enough. Now, why would we think that about accounting, but not say, programming? Well, it boils down to the depth of your understanding of accounting!
There is an interesting cognitive bias in play here, it is called the Dunning-Kruger Effect, which basically states that the less you know about something, the more you think you know about it and the stronger your feelings about that item’s unimportance. There is also a correlating theory that outlines the four stages of competence and looks like this:
1. Unconscious incompetence – I don’t know it but think I do, and so, it can’t be hard and therefore it isn’t all that important anyway.
2. Conscious incompetence – I don’t know it but I also realize that it is bigger than I am.
3. Conscious competence – I know it, but it takes work to get it done.
4. Unconscious competence – I know it and I don’t have to think about it.
I put these things out here because I have an interest and a concern about how they play a part in the application of business intelligence.
We use BI to help people make better decisions and that almost always requires overcoming their biases and preconceived notions, and their decision making models. Biases are difficult to overcome especially when someone is at a level “1” (unconscious incompetence) and they are making decisions thinking that they have everything that they need.
In my accounting example, if I don’t know anything about accounting, then it is easy to say, “just enter the invoices into the computer, pay the bills and stop pretending that it is so hard. You don’t need more resources or a new system because the old one worked and not much has changed anyway.” The reason is because in my mental model, unencumbered by any understanding of the complexities and vagaries surrounding it, accounting is just an administrative event that anyone can do. For those of us in the technology field, we can close our eyes and hear many managers say things like, “just put it on my iPad, how hard can it be?”
My concern is that while we fight biases all the time, this one is particularly difficult because the successful use of Business Intelligence requires an active consumer of the data in order to be relevant. So to fight recency bias (the bias for more recent data over older data) for instance, we give larger time frames of information so that the consumer can see the trend and gain perspective by noticing that historical data is relevant as well. The bias is addressed because the consumer can’t avoid looking at the whole chart.
The Dunning-Kruger Effect is more insidious because people aren’t looking for the data and what they see they discount as either irrelevant, or even incorrect. I’m looking for suggestions here, but here is my basic thought process.
In order to get this bias addressed, we must first create an environment or tool where the decision maker will seek out the data. This can be done in a number of ways but I’m partial to gamification of the data. The basic description of this process is that we have to turn the data access and analysis into a game of sorts. Basically make the use of the tools “fun”. Now this is pretty hard to do with old green bar reports if you read my last article, but today with dashboards and really awesome looking mobile applications you can take a step towards making visualizing and accessing the data fun… not quite golf, but better than a root canal.
A scoring mechanism is difficult of course (all games need a scoreboard after all), but you can add that be providing collaboration opportunities where executives can share their insights and the data that they have found to be of value with others. This gives them a way to show that not only are they engaged, but can also alert others that may need to address (or act upon) what has been discovered.
The second piece of the cognitive bias is to fight the urge to ignore things that we don’t understand or that don’t fit in our mental model. Basically we are trying to get the decision makers into the 2nd stage of competence. To do this we need to make the data simple and easily applied and compared to other things which we already know and believe in. This is obviously way easier to say than to do, but the process is straight forward: start by making your data as simple as possible and then simplify it.
An example might be, if you are trying to show the MPG (miles per gallon) impact of maintenance on your vehicles and your VP of operations isn’t a big maintenance fan, break down the numbers into just preventative maintenance and MPG. This way, the data can be consumed in a way that will drive appreciation of the relationships that you’ve found. Don’t put a chart out there with 100 different types of maintenance as that will drive away the consumer. We like to make really complex dashboards because that’s what we do for a living; just don’t let the desire to make something really cool override the desire to make someone do something really cool.
I would love to hear your thoughts and ideas relating to this. leave me a comment below.