“Business Intelligence” made simple
June 20, 2009 9 Comments

dilbert on business intelligence
Business Intelligence (also known as BI) is probably one of the most oversold areas of IT.
There is no doubt you need information to make decisions but too often a mystic, voodoo magic blanket is drawn over what should be an easy to understand topic. Consequently the big vendors of BI charge through the roof for their solutions.
You are convinced, rightly, that you need information but in most cases you do not need a hugely complex and expensive system to provide and analyse that information.
The important starting point is to have database systems in place to manage your core functions. You might already be using systems for CRM, HR, Finance and so on. If you are using spreadsheets to manage core parts of you business you are going to struggle. Spreadsheets are a good tool but they are no match for a properly designed database system.
So, what are the goals of business intelligence?
- Todays Information
- The first goal is to accurately provide information on what is happening across your business today. This is the hardest part as you have to make sure your systems are in place and are producing the correct type of information. If you want to track your sales per regional office for example then you have to make sure your systems are capable of capturing that type of information.
- Trending the Information
- Now as your data builds up in your systems you can start to trend it over time and compare it. There are some very specific issues to overcome if you wish to do this accurately. Your systems need to be able to accommodate for changes over time. For example, an employee may move from one team to another over a 2 year period and you need to ensure that all their related sales information is categorised across their old and new team correctly.
- Benchmark Comparison
- This is where the value of business intelligence begins to grow rapidly. By setting goals, targets, and benchmarks you can track the performance of your company against expected results. How are sales doing against target this month?
- Tomorrows Information
- The holy grail of business intelligence is to make inferences and predictions based on historical data. Depending on your type of business this is either going to be a reasonably valuable tool or completely useless. And for the vast majority of companies it is the latter. It is dependent on having very large amounts of data that is very well defined and very accurate. I am sure the likes of large telecoms companies and scientific institutions are playing around with this aspect of business intelligence.
- External Comparison
- A hugely beneficial but rarely achievable side of business intelligence is the comparison of your information to other companies in your industry and other industries. The problem here is the lack of reliable information that you can compare too.
If you have achieved the first two stages above you are in a good position and ahead of a lot of companies. If you are hitting stage 3 then you are in a stronger position. Getting to even the first stage requires planning and you have to implement the right systems in the right way. The “garbage in, garbage out” moto is one to take seriously at a very early stage.
Just remember that you do NOT have to blow the IT budget to achieve this.
Pingback: Powerful use of Data « Niall Hannon
Niall,
I was wondering what led you to saying:
“The holy grail of business intelligence is to make inferences and predictions based on historical data. Depending on your type of business this is either going to be a reasonably valuable tool or completely useless. And for the vast majority of companies it is the latter.”
Particularly the last sentence.
You mention making predictions about the future based on history as maybe being relevant to Telcos. To offer just one other example, how about Insurance?
Some thoughts on this are here:
http://peterthomas.wordpress.com/2009/03/24/the-specific-benefits-of-business-intelligence-in-insurance/
I would be interested in your feedback.
Peter
Hi Peter, thanks for the comment and your link (interesting read).
I do think there is a place for predictive BI but it is only suitable to a minority of businesses. I think you need a very well defined business model to be able to apply prediction to it. I would think this suits the insurance industry. My assumption here is that financial models are quite broadly used, that it is all about mathematical models?
Other industries, e.g. the software industry, don’t have the amount of low level data required in the use of predictive BI and other things are just difficult to pin down e.g. quality. But that is not the root of the problem.
The main issue is that for most businesses you simply cannot accurately predict the future. There are too many unknowns, the boundaries are too vague and results are subject to a random set of events that have no relationship to past events.
One good book that explains this better than I do is “Ubiquity: Why Catastrophes Happen” by Mark Buchanan.
I guess it depends on what you mean by prediction – having spent the first 8 years of my life working in a software house, then I’m thinking that advance notice that the number of leads you are getting for Product X is dropping off might lead you to predict that sales are going to dip and that a product refresh is required. Prediction doesn’t have to be via sophisticated statistical models – sometimes just a graph suffices.
A difference between Business Intelligence (its subset) Business Analytics maybe as per:
http://peterthomas.wordpress.com/2009/04/14/the-dictatorship-of-the-analysts/
Peter
I suppose the difference is that it is a human who is then predicting based on a historical trend rather than any BI system generating predictions. Maybe they both give the same result anyways in all simple cases.
I think a lot of companies would be very happy just to get to the point where they had accurate reporting showing the drop in leads. Having that accurate data to report on is the bigger issue.
I agree that statistical modelling may be over-sold in some industries – at least at this moment in time.
Peter
Hi Peter,
Any exciting new trends in the area of BI? Where are things going?
Niall
My take on trends from back in March:
http://peterthomas.wordpress.com/2009/03/09/trends-in-business-intelligence/
Peter
Pingback: Spreadsheets – Dr. Jekyll & Mr. Hyde « Niall Hannon