The BI Guy
-
The BI GuyI've certainly not been accused of being the "BI Guy", a BI expert, or even an advocate. In fact I tend to think that BI won't really catch on in the mainstream. My view has been that there are some small places where BI is well implemented, provides a wonderful ROI, and makes sense. But the OLAP based versions of it are just too hard and complicated for most of us to implement. I think that data warehouses will slowly grow in companies, but BI is a fad and won't really catch on, despite all the press it gets. And I'm not alone as the Battle for BI talks about.
However I'm willing to admit I could be wrong and maybe BI is the next big thing. So I'm asking.
Are you going to implement an OLAP based BI solution?
I'm looking for any concrete plans, whether next week or next year, on you putting a BI system, servers, clients, and software on an end-user's desk in your company. I guess if you're building a BI type solution off of relational data or a warehouse that should count as well. I'm more interesting in the OLAP side, which is how I see BI, but let us know if you're doing BI.
Maybe it will help us decide if it's worth it for us.
Steve Jones
-
In 2004 I participated in a product that had a layer of BI (not part of what I was responsible for), and I had a chance to observe BI expertise in action. The big challenge was evident from the start, which is that BI ends up being an ongoing process, because businesses have a lot of difficulty defining what they want the BI components to address. In this case it was a shipping company, and they simply couldn't push past the bad data to the good information. They ended up with the BI component confusing more than otherwise.
I also saw a lot of misunderstanding of how BI generates value (I'm sure it does in many scenarioes). Managers simply thought of it as some kind of static reporting element on steroids, and seemed unhappy that it took actual effort to generate the information they thought they wanted. In the end, they returned to pouring over printed lists, because it required no new skills to do that, and the data was less confusing (in their opinion).
Is BI a fad? Probably not. Does it have widespread application? Certainly not yet, because businesses, even most large ones, are not mature enough to make it return value to them.
But then, if I was BI guy, maybe I would feel differently, too.
-
There is no doubt BI is difficult to achieve, but it would be unwise to dismiss it or minimalise its impact because many have failed to implement it successfully. The concept of being able to both measure success and make key decisons by analysing a few screens of summarised relevant information is highly compelling to most decision makers. In most instances it fails because you need to merge key business knowledge about what are the key facts and KPIs we need to measure along with the difficult process of collecting data from different operational systems and bringing it all together in a system which is easy to comprehend visually. This often means selecting resources from various parts of the organisation putting them onto a BI project, but the problem is none of them actually have a good feel for the whole business either as a business or from a systems point of view. Consequently too many issue fall through the cracks and eventually undermine the value of the final solution. So in the end its a people problem not a software problem.
Start small with Excel and the right people and you will move on to success. Construct a BI team with representatives from all areas, have a software beauty parade to select a BI suite and you will be mopping up tears and saying BI is too difficult.
-
I have built many prototype solutions for companies like Crystal Sugar, Ag Co Corp, 3M, Rockwell, etc... back in 2004 in support to sell an analytical client call OLAP Hub.
The common problem I ran across was that the results were perceived as a simple report writer and, as others who have responded have stated the client did not understand why it took so much work.
While I don’t think BI is mainstream yet, I think it is picking up significant momentum in the market and that will only grow over time. I believe that the BI tools are still maturing and evolving and that is why it is difficult for both the client and technologist right now.
I am currently working on a project with Rockwell Automation to implement a BI dashboard to help a plant manager evaluate the performance and events that occurred on the machines on his/her shop floor. The original solution was developed in 2003-2004 for SQL 2000 and we are refitting it now for SQL 2005. Their customers were so hot for this solution because it showed them things they couldn’t see before that we are adding more capability as well.
I have another company I am working with that is looking for the capability that BI provides but does not know that the answers are found in BI. So, I submit that business wants what BI provides but does not necessarily know it resides in BI.
Fundamentally BI provides a way to aggregate or roll up data at three intersection points to description information that is grouped. This is past information that provides the "trending" information. From this we can move to predictive analytics and provide a calculated guess as to how the "trend" will occur in the future by studying how it trends in the recent past. We also can "mine" the data for meaningful patterns that would otherwise not be visible under normalization review.
All of the methods described above can and do provide strategic advantage to a business to be competitive and make no mistakes, technology implemented in a business has only one real purpose, assist in adding to the bottom line because in the end, business only survives if it makes money. These kinds of tools help by providing “decision support” for spending and makes you successes and failures based on historical data more visible to guide your business decisions.
-
Being part of IT in a fairly large healthcare provider, we are currently in a position where we are investigating the pros/cons of implementing a BI product. I'm a "lower-level" guy in the chain and will be supporting whatever product our company decides to go with. The BI demos that I have attended over the last few weeks have left me feeling excited about seeing any one of the products in action. BUT... at the same time I can foresee the extreme difficulties trying to put together any kind of warehousing or metadata scheme that will make any one of the products function as they have been presented. The final decision will be made within a few months and begin implementation within a year. For me personally, I would rather be working with the new SAS products.
It will be very interesting to see how we combine data from hundreds of different sources to supply an OLAP data warehouse. We have data types ranging from SQL data marts, Oracle db's, Cognos cubes, mainframe, homegrown MS Access db's, and the ever-popular MS Excel workbooks. This is just to name a few.
In my opinion, BI appears to be a very nice solution for those companies that have the desire to collaborate and analyze overall company performance. But at what price to the IT folks??
Will the fad fade? Well... that depends on how many companies work through the growing pains and share their success stories. Hopefully, ours will be one of those success stories
-
IMO the waters are a little muddied when it comes to describing what BI actually is. I tend to think of BI as just knowing what's going on in your business so that you can make correct decisions. If that is what it is then BI has been going on since long before we all had a workstation on our desks.
OLAP is a technique that can help you achieve business intelligence (note small 'b' & 'i') but it is one of many and its wrong to think of OLAP and BI as being the same thing.
So...will BI catch on? Yes, it already has done. Businesses have been doing it for years and years.
Will OLAP catch on? Who knows - I hope so. From a technical perspective it fascinates me.
And let's be honest - the majority of the effort and pain in implementing a modern BI system comes in the back-end - i.e. sourcing the data! It isn't usually spent in implementing the presentation mechanism!
Just my two-penneth!
Oh, and in answer to Steve's question: "Are you going to implement an OLAP based BI solution?" my answer is "Yes, I'm doing it now on Analysis Services. And I'm loving it
"-Jamie
Jamie Thomson
http://sqlblog.com/blogs/jamie_thomson -
I'm actively researching a BI solution now. Or a MIS solution, choose the buzzword of your choice. Products I've looked at are:
- Business Objects
- Cognos
- QueryObject
- QlikTech
I also looked at SQL 2005 Analysis Services, but dismissed it when I read MS had bought ProClarity and was repackaging it for release next year. So their current implementation is not good enough?
I've done most everything in SQL apart from the OLAP sphere, although at the last place we had a Business Objects looking at a Data Warehouse hosted on a piece of hardware called Netezza, which was cool but I was at the 'pump data in' end of the process. Others looked at the 'TL' part of the ETL process and managed the MIS process.
BTW, if anyone has got any familiarity of the above products, I would value your feedback on the experience.
My two penn'orth, BI is a tool that my business customers are asking me for, and will be a valuable tool to my end user community.
Dave J
-
Hi David,
"I also looked at SQL 2005 Analysis Services, but dismissed it when I read MS had bought ProClarity and was repackaging it for release next year. So their current implementation is not good enough?"
What are you referring to when you say "their current implementation"?
ProClarity is a front-end to Analysis Services which means ProClarity is useless without AS. Microsoft bought ProClarity to enhance their product set - it isn't replacing anything!
-Jamie
Jamie Thomson
http://sqlblog.com/blogs/jamie_thomson -
Guys,
BI implies presense of both - business AND intelligence...
Decision Support Systems - BI in new speak - is as old a concept as computerised data itself. There are plenty of very good tools today on a market that make implementation of Analytical systems' a no brainer... when actual INTELLIGENCE is available on both levels - business and IT.
Unfortunately (or fortunately), IT departments are run (and consist of) people who think they are very intelligent, but actually are donkeys(!) And most of business people are sick and tired of IT BS. That's why so many IT jobs end up in Bangalore.
I can sense from your questions that you are more interested in DEPLOYING servers, PCs...
It's good to remember an old saying - someone who speaks, does not know. Someone who knows, does not speak...
Good luck with STM (Someone Told Me) education...
-
Apparently, you are not a person who works in Information Technology and according to you're own post, you do not know.
-
BI is a general concept, but its usually confused with metrics, KPIs, predictive analytics and a lot of other incredibly promising but usually nebulous technologies. Ultimately a BI system succeeds or fails based on the reporting foundation you've already got in place. If you have a way to deliver tactical reports to the people who need them, you've got the social basis for building a BI system. Its the social basis that's the hardest to get. That doesn't mean it isn't hard to get metrics and KPIs and a display that shows red, yellow or green depending on how things are going, but if you can't deliver tactical reports and you don't know what data people are using to run the business, you can't guide management and your developers in the right directions.
As far as the specific technologies go, you need a way to deliver the output and a way to run reports without impacting your transaction systems. That's it. Right now I use an ODS, a warehouse (dual fact- and aggregation schemas), and Reporting Services to report on data from three major ERP systems, a dozen online colloboration tools, Excel spreadsheets, flat files, and a partridge in a pear tree. But I've also used Access, Business Objects, Analysis Services, cute little VB6 reporting apps I built myself, whatever. A little ingenuity goes a long way.
So don't worry about the technology. BI is very useful. It makes business people very happy to get the data they need, and it can help your team a lot to engage with people in the business - its good all round. Its the analysis that really makes a difference. Go slow, make sure everyone knows its a launch-and-learn process, build in lots of reviews, and put your most socially-adept people in the middle of things. You'll have BI before you know it.
Oh, and skip Analysis Services as a starting point. Write reports against a db. That's the technical foundation.
-
David Grover,
Well said! I couldn't agree more! Its what you do that counts, not how you do it!
-Jamie
Jamie Thomson
http://sqlblog.com/blogs/jamie_thomson -
Jamie
"I also looked at SQL 2005 Analysis Services, but dismissed it when I read MS had bought ProClarity and was repackaging it for release next year. So their current implementation is not good enough?"
The question mark means I was asking a question
Explicitly, Is the current implementation good enough?
What have you built with it?
Any issues in implementation?
How was the support from your reseller?
Was it MS? Or a third Party?
I could go on...
Dave J
-
David,
I still don't quite understand why you seemed to think that the purchase of ProClarity reflected badly on Analysis Services. If my interpretation of what you said is wrong then pardon me!
This is a bit of a digression from the main point of this thread so if you contact me offline via the link below I'd be happy to try to answer some of those questions !
-Jamie
Jamie Thomson
http://sqlblog.com/blogs/jamie_thomson -
No worries. The smiley indicated I'd not taken offense
I'll click on the link as you suggest & we'll discuss further.
Dave J
Viewing 15 posts - 1 through 15 (of 23 total)
You must be logged in to reply to this topic. Login to reply