Some good points here

Business Intelligence Programs for Healthcare

Some really good nuggets in this one.  The one aspect that I know that we have struggled with over time is the Data Governance aspect of the data warehouse?  Who "owns" the data, who says what it means, who gets access to it?  It is a slippery slope and always turns political, especially where there is not enough guidance by the executives. 

I really like the idea of the ecosystem as pictured here as well.  The users are not the only part of the system that you need to worry about, their managers (assuming that they are the executives) need to have bought in to the whole program so that they can direct people to it.  Without that, it is really had to achieve any sort of real ROI. 

User adoption is the fuel of your BI program. Trust, ease of use and training are the trinity of user adoption. Users need support and training.

User adoption is the mother milk of a BI system, you can have the best, cleanest most efficient data warehouse, but if  it is difficult to use, hard to understand, or, not what the users want, there is little chance for success.

Simplifying your user interface is also very helpful. Research has shown that users are sensitive to the number of mouse clicks they have to make – anything beyond ten clicks and users will complain. 

I have seen this almost on a daily basis.  We have built some of the most robust and flexible reports for our users with an insane amount of options, but what I find, is that for a majority of the time, they choose the higher level data to run, and then they cut it up themselves in Excel (I contend Excel is both friend and foe to BI).  A good analytics tool would help to eliminate this from happening, but the learning curve on some of these products (especially anything being put out by the large players) makes it a really difficult sell to the users.  I believe the best tool out there for ad-hoc analysis has to be Tableau, everyone I have shown it too has taken to it.  

Anyone else out there run into issues like this over the years?

Time series comparison

Baseball again..

Lee v Holiday 

Stephen Few would be proud.  Excellent time series comparison over 2 very good pitchers, but who is the best consistently?  Seems to be Roy in this case, which really does not surprise me, since he dominated the Red Sox for so long in Toronto (but at least those games were pretty short).  The question is who do you want to go in a game 7, some one who is consistent, or someone who, when hot, it better than everyone (Verlander needs to get out of D-Town).

However, I would have like to seen something that compares year over year to see if there is some sort of trend as the season goes along.  Seems to me as well that in 2008 they were essentially the same pitcher.

You know those "rewards" cards?

We all know (or at least should know) that these rewards cards that are given out by any and every retail store these days are used by their marketing departments to create really rich data.  As this post illustrates this data is very important especially now, where the average customer (is there an average customer anymore?) shops multiple channels (brick and mortar, online, phone, mobile), so what these cards allow is to attach more demographic information to the actual sale as well allow for aggregation across channels.

Read about it here: Retail and BI

Again this data is very rich, but needs to be used very carefully.  Yes, you now know that the mother of a 10 year old likes to buy jeans at Old Navy, but you do not need tell them that.  It is a very slippery slope between good direct marketing and "big brother" looking into your home and wallet.

On point

Don't let BI make you dumb


Spot on here.  I especially like the comment on sitting with the sales VP and only after going through 9 reports, were they actually able to see one that was of any worth to them.  This happens all of the time when going through a project.  Users are brought in, they say I would want this and would want that.  What happens is that multiple reports are made for a fragmented user base, where no one report is used as much as one would like.  This makes the ROI almost impossible to reach (see my previous post here where I expound on this more) and then possibly making the project a failure. 

Know your audience is also very important, people do not like what they think of as Big Brother, especially in an economy that has people nervous for their jobs as it is.  On top of that, there was always this feeling that I got when dealing with certain users where they looked at BI as the enemy, that BI was going to replace them and then they would be out of a job.  While in some cases this is true, in many, it would only enhance their jobs, getting rid of day to day drivel and allowing them to concentrate on improving their own personal ROI. 

Back to Baseball

I guess the division series had an affect?

Pretty neat visual here, shows the how the expansion of baseball (first geographically, then the expansion of the leagues, then the expansion of the playoffs) affected the length of the season.

The travel times back in the day were pretty insane though (there were not team charters that flew them out one hour after games time), so it is interesting to see that the length of the series in days, including off days was not that horrible.

I guess the real culprit here is TV.  The expansion of the playoffs is one thing (and now we are going to have another round.. if the rumors are true), but EVERY game is on at this point, so they have to spread them out to avoid as much overlap as possible.  Good news this year though, barring weather, its over before Halloween.