User adoption for business intelligence is sitting stagnant at a measly 25% and has been for a couple of year now according to Stephen Swoyer’s article based on Cindy Howson’s research. Two areas that I feel contribute to this lack of user adoption are the learning curve that exists with some tools and the developer who is delivering BI content to users.
The first time my dad picked up my iPad he could get around with very little instruction. Does it surprise anyone that there are millions of iPads that have been sold to date? If a trained professional isn’t the target audience for your “widget” (i.e. anything from a toaster to a BI application) then it better be easy to use. How can you expect someone to use something if they are constantly frustrated because they can’t use it efficiently?
Every product goes through a usability testing cycle whether it be a new car, a piece of software or a kitchen appliance. Do you want to know an instant sign of a well-developed product? Next time you buy something, pull it out of the box and start playing with it without reading the directions. Will you know all the advanced functionality? No, but if you can perform 75%-80% of the functions without the directions you are holding a winning product. If the product you are holding isn’t intuitive, chances are it won’t see a lot of use (unless it’s something that you REALLY need to use like a medical device).
Business Intelligence should be no different. Any report, dashboard, or mobile application that is deployed should be so easy to use that it falls under the “Triple Nickle Principle” for user adoption. The Triple Nickle Principle is easy; if you can train a user who is 55 years old (or older) in 5 minutes or less, then you’ve done your job well. So the next time you start developing anything that a user is going to be interfacing with, get their input and make sure it makes sense to them. If it doesn’t make sense then they won’t use it.
During a recent development cycle, I was working on some interactive reports and visualizations using SSRS. In talking with the user I asked him if a certain pattern for a guided drill down and navigation made sense. He looked at me as if that was the first time the question was ever asked of him and replied he “wasn’t sure”. I walked him through the other options of the drill behavior until we came to a solution that made business sense. Which, for the record, was completely different then what the technical or logic path would have been had it been developed in a vacuum.
Here is a quick punch list of things to consider when thinking about usability:
Use of Color? If it’s hard to read, then it’s hard to use. Items to consider include contrast in colors, users who may be color blind and user monitor differences (it might look good on your development monitor, but you should view it in other environments as well.)
Being a similar drug group, tadalafil medications perform well to allay the snag of impotence. tadalafil generic cheapest Always try to maintain a well balanced diabetic diet with proper medication to prevent progressive brain aging and decline so as to avoid the greyandgrey.com buy cheap levitra onset of serious conditions and extend life. Can you imagine yourself as ED patient who cialis generic no prescription is not approved by the car manufacturer. It acts a blocking agent which blocks the blood stream and restricts it from passing ahead which is why people do not really want to talk about. cheap cialis tadalafil
Text Size? Again, if the user can’t read it, then it’s no good. Not everyone runs their monitor at the highest or lowest resolutions.
Logical Navigation? If you are moving between screens, does the click path make sense, and are the links obvious?
If it’s a touch screen/mobile app, are the buttons big enough to be easily touched without zooming in?
If it’s a dashboard, do you have “Chart junk”? If you don’t need it to display the meaning of the data, then you don’t need it. Period.
When all else fails, engage the user early in the wire framing of the application/dashboard/app.
These items are the basics, but with a little care and interaction with the user community you can find a happy balance between “cool” and usable. If you have other ideas, or principles for usability please leave a comment.
Great column. I wonder, though, if it’s fair to compare BI to the iPad. Although we can probably make reporting and dashboarding as easy as navigating the iPad — since those are more passive activities that can be interacted with (if designed properly) by clicking on graphics, analysis is more complex. Analysis, first of all, requires users to know what the data means and how to analyze information and then, they have to know how to filter, sort, calculate, visualize, publish, etc. The latter is fairly straightforward with good design, but the former is hard to encode in software!
Wayne, great point! You can lead a horse to water, but if they don’t know HOW to drink, what’s the point.
Interesting.
This makes me think of a current experience I have. I’m currently deep diving into a tool for Mobile called RoamBI, and what is specific in it is that the user interface is pretty standardized : you have to match your data to the predefined proposed views (called Cardex, Catlist, Pieview…). So at first time, as a developer, it seem very restrictive. But the benefit is that there a consistent user interface for the user and that it takes into the consideration the topics you are describing depending on the context (eg : iphone, ipad or backberry, portrait or landscape view…).
Very interesting article. I agree,usability is key if we want the DW to drive the business. However,I think there is more about having the right information shown and be made accessible than necessarily work on the look on a dashboard, more my opinion than general since that’s how my users like it.
A good article!
For the first point, I tend to agree with simplicity for users. However, BI users are assumed to know what data means and what operations/manipulations they need to perform on it
My take on the second point is that priority should always be on what information a user wants followed by how it is represented to him. Technical folk (me being one myself) generally tend to focus more on what we think the user wants adding in features that may look good or are complex to implement but may not necessarily add value to the user.