Data Driven Analytics


Data Driven Analytics

… Are you Business Intelligent?

Are you making better decisions based on the data you’re collecting for your business?  Are you using data to re-enforce what you already knew about your choices? Are you able to predict change accurately with data?  Well, Then you’re probably using Data Driven Analysis and Business Intelligence.  Not every choice you make requires analysis. However, for the bigger decisions and the ones that will have an impact; including BI and DDA into process should have a dramatic positive impact on your enterprise. If you need help building a Data Warehouse, Business Intelligence or Advanced Analytics system, Please call or email us Today.

Do you measure the results?

One of the most over-looked part of Business Intelligence is measuring the results of your choices. Most of us can tell if we’ve made a good choice or a bad one.  However, How Good? or How Bad? is not something that is easily quantified. There are times when you make a choice that as an immediate bad result only to work out better in the long run. Sometimes, significantly better. So, Measuring results – right now and over time become very important and are often over looked.

Lets add one more. Ever re-construct a choice you’ve made in the past (maybe even distant past) to calculate if you could have made a better one or in an effort to perhaps understand the “root cause” of an event.  These are all part of “Data Driven Analysis”.

In my mind, Answering one questions usually results in several more questions and branch points.  That why it is so important when your building a data warehouse or analytics systems that you consider someone else will be asking these same questions.

Can you Predict Change … Accurately?

Another Big challenge in Data Driven Analytics is predicting change accurately.  There is a math axiom that states our ability to predict the future diminishes in accuracy over the time interval.  I disagree.  If you do allot of time based analytics, you will notice patterns in your data. Modalities – places where the data changes based on time of day, day in week, week in month, week in year, month in year, quarter in year ..etc.

These state changes are be used as a map to adjust focused predictions. Example:  Its Q4 and your in retail. You have established a Year over Year increase in your sales. Your Models predict an increase in sales BUT, for some reason you experience a significant “Dip”.  Yikes, Didn’t see that coming…  The first thing most people do is blame data quality. Of course, Quality will have an effect on magnitude (most likely volume). However, The model should have predicted the state change from positive to negative. right? Maybe not…

The model predicts an increase in sales but some where the data lost focus and wasn’t able to predict the state change.  If you were able to see the modalities (“Ice Bergs”) and make adjustments – would that have value? Again, Measuring the outcome of your choices would allow you to evaluate the decision ex-post-facto (after the fact) .. And, will help you with adjustments..

Hidden Fly-Wheel?

Sometimes the Model is spot on and you have identified all “Effect” except for the ones the exist out-side your system. What? This often over-looked when considering the micro ecology that is your business.  We sometimes forget that we are apart of a larger system.  We take for granted that things like power, compute and system steady state are constant. They are not. So, When your considering Data Driven Analytics – make sure you have accounted for the variability of external forces.  Hint: There are Bayes Networks and HMM models that will help you, just to name a few..

For me, Having Great (Accurate, Timely, Repeatable) real-time metrics are paramount.  I’ve had the honor of working at some amazing data driven orgs like Google, Amazon and Blueskymetrics.  Its fun, hard and rewarding work.  If you want to learn more about my approach to Data Driven Analysis, Please let me know. Have Fun! – Gus

Please Contact US or Subscribe to our Blog if you found this interesting or would like more information.

Subscribe : Blueskymetrics Blog

* indicates required,  Managed By Mail-chimp – Please check your Spam Folder and Confirm Subscription.