What is Embedded BI?

As I glanced at facebook I saw a status from Deborah Lilley interesting enough get me thinking at 7am in the morning, she was talking about embedded UI.

This morning I went out of the house at exactly 6:59, ran down the hill and quickly across the road to arrive at the bus stop just about the same time as the bus.   I jumped on catching my breath, sat down and switched on my iPhone (this is a great thing about taking the bus to work, I get 12 minutes sitting reading, checking email or whatever rather than 12 minutes driving the car).  As I glanced at facebook I saw a status from Debra Lilley that was interesting enough get me thinking, even at 7am in the morning.


I applied the embedded BI concept to my habitual tardiness and I liked the analogy, so rather than write 500 words on a facebook status, I decided to put my thoughts in a blog post.  Here goes…

My operational activity is getting out of bed, ready for work, maybe doing some work/checking mails, maybe having breakfast, maybe getting kids dressed and finally I walk out of the door to jump on a bus to the office.

My traditional BI world is looking back on a week and wondering why I missed the bus three times, what impact that had on my life and how I might take some corrective action.  I may determine it takes 6 minutes to walk to the bus stop or 4 minutes if I run a little.  Maybe I notice that on two occasions the bus left 1 minute early and that was why I missed it.  I might analyze decide that for the next week I will set all the clocks in the house 1 minute fast, that way I should arrive at the bus stop 1 minute early and that ought to prevent me missing two buses a week that arrive early and make a huge difference to my punctuality and effectiveness in morning meetings.

Now as anybody who has tried setting clocks forward in order to trick themselves into leaving earlier and so preventing occurrences of tardiness will testify, it doesn’t work.  You look at the clock showing 6:50am and mentally subtract a minute and figure you have time for one more slice of toast if you eat it fast and then run a little.

So how is embedded BI going to help me more?  (I am getting to the point soon, I hope you’re still reading) Well the BI I get needs to embedded into my operational world and help me make better decisions in real time.  In order to do that I would need it to be relevant, timely and actionable.  Let’s look at these three three items in teh context of an example – I am considering eating one more slice of toast at 6:50, my next bus is at 7:05, the one after that at 7:40, the bus journey takes 12 minutes, then I have a 5 minute walk and I need to be at my desk at 8am to host a Web Conference, the fare is $2 and I only have a $20 bill in my pocket.

Timely – This is easy, I need information before I put the toast in the toaster.  If it comes after I put the toast in the toaster but before I start eating it then I might still get my bus but have to leave the toast behind and wasted.  If I get it even later I may miss my bus.  As you can see the later in the task I get the information the worse the result is.

Relevant – I don’t want generic information on my bus catching/missing history over the last 12 months.  I want to know how often does having a slice of toast 15 minutes before my bus is due causes me to miss my bus. Some information about how often the 7:05 bus arrives early would help too.

Actionable – In my bus example the action is pretty much down to me, I get up from the table and head out of the door, the action is in my own hands.  One action I might need to take is to break my $20 bill to make sure I have the correct change for the bus fare, embedded BI would remind me of this and might even present the alternative option of borrowing $2 in change from my 5 year old’s money box (A further action might be to remind me when I get home to pay it back)

The example I am using is fairly trivial and most people would tell me to just arrive 5 minutes early all the time to be sure of catching the bus.  However in business the decisions are much more important and the results translate into operational efficiencies, less mistakes, higher productivity, higher sales, customer satisfaction, employee satisfaction better profits all those good things.

So that was my attempt to explain how I see embedded BI.  Agree, disagree or otherwise express your thoughts in the comments section below.