Plug and Play Business Intelligence?

Really? Do I have to revisit this topic? Again? 2 years ago, I wrote that there is no such thing as plug and play business intelligence. Here we are and has anything changed? The answer is a resounding NO! 

Now, let me clarify this. Analytics tools have come a long way. There are several good analytics tools out there that allow you to do quite a bit of reporting and analysis yourself.

Then why am I insisting that there is no plug and play in BI? Unless you are the Data Warehouse Ninja , ETL(extract transform, load) Samurai and you have a team of 47 Ronin who, through the years of wisdom in the scrolls given to you by the smartest kid, I would venture to say that you don't have a handle on all the data that the business is looking for. And even when you do, the requirements are ever changing. Get my point? It's the data!!!

Data integration and standardization across the enterprise is going to be your biggest challenge. Unless you have that under control, you can put the next Alien ("look it also makes me a sandwich") dashboarding tool on top of it and you will still get bad data. Ever heard the saying, "all that glitters isn't gold"? Well, all that creates dashboards isn't BI either.

Omnition partners with United States Preventive Medicine

Omnition is pleased to announce that we are partnered with United States Preventive Medicine.
http://www.prweb.com/releases/2013/11/prweb11306930.htm

Big Data

Recently, I copy-pasted a link to an article saying that Big Data is dead. That got me thinking. Is it really dead? Serves me right for copy-pasting a link! Here are my two cents on Big Data. Most people don't have Big Data. Google has big data. Yahoo has big data. But if you are a regular business, and even if you have large amounts of data, you might not need to apply big data principles. Let me clarify. If you have large amounts of "unstructured" data and you have a need to mine them, then you might have a need for applying big data principles (Dr's notes, for example) . If most of your data is "structured", no matter the volume, you may never have to apply big data principles. If your data is mostly structured, even with large volumes, it is possible to extract information in a timely manner and you may never have to use big data principles. 

To predict or not to predict?

Let's face it, if you have been to any healthcare industry conferences lately, you have heard about analytics and more importantly, "predictive analytics". Wow! Can you imagine the power of "prediction" at your fingertips? What if you could tell the number of people who are going walk through your door any given day before it happened? What if, you could tell the number of patients who are going to be cancer patients in the next few years? That would be awesome, wouldn't it? The question is, would it?

Let's take a scenario....
Let's say that I now have the predictive capability of knowing that 3 out of my 10 patients will get cancer in the next three months. Great! Which 3? What type of cancer? What can I do to prevent that from happening? The answer is, probably nothing. It is not specific enough. While there are some general changes you can make to your treatment plan, the results will likely be a hit or a miss.

So why spend so much time, effort and money on prediction? To me, a better scenario would be a retrospective inspection of what caused something to happen. In this scenario, you can drill down to the specific root cause of what went wrong and it would be specific enough for you to be able to fix. Much better, "measurable" ROI, in my opinion.

So, when choosing to go after predictive capabilities, understand your goal for it. Look to see if there is a measurable ROI for it. If not, don't try to predict, analyze.