I have come across many occasions where IT Professionals have pushed really hard to use all of the BI stuff in SQL Server and have got nowhere fast. Not because they don’t know what they are doing nor because they didn’t give it their best. So before you have a go it might be worth exploring why business intelligence projects go wrong and what the warning signs are.
BI is just like BI tools that come with SQL Server. You probably bought SQL server or it came as part of another solution to run your business, and so the database engine is essential to what you do and you will get it working however hard it is. However you may not even have noticed all the other tools that came with SQL server, or you feel you don’t need it. BI is the same; EPoS and ERP systems are seen as essential,`but many businesses think they can survive without BI. So the most important parts of the readiness test for BI are :
- Compelling business need. What return on investment will we get from this?
- Strong business ownership. Who is the champion on the board to push it through and make the tough decisions?
You need both before you can even start. Without business ownership, the need might be recognised but not properly presented to the board, so the business will put up with the pain or not even know the extent of the problem. The other way around is that a visionary manager will not be able to convince the finance director to put up the cash for the project or get the rest of the business to use the system. If a BI system is to be successful the business will have change in response to what the it tells them and changing business culture is far harder than MDX, the CLR or understanding software licensing!
The next two are directly in your area:
- Sound relationship between IT and the business. Do they trust you to deliver and do you trust them not to keep changing their minds? another issue here is that the users are doing it for themselves – you might see the mother of all spreadsheets on your servers or access databases sprinkled all of the place with linked tables back to your source system.
- Suitable Infrastructure. This means the IT team as well as the network and the servers
Finally there needs to be a culture of analysis in the business. This means users relying on the data they use to make decisions rather than ‘using the force’.
Does it matter? You could simply argue everyone is doing it so it must be useful – Microsoft have spent a huge amount of time on developing BI for the masses, Oracle acquired Hyperion to have a stronger position in this space, and the niche players are still growing. You could be fed up of writing reports time and time again for different users who keep asking for changes to them or your line of business system is simply groaning under the weight of badly written end user queries.
The harder question is what do you do? I would quote Douglas Adams here and say you should make it Somebody Else’s Problem. Find a business sponsor , identify the information issues arising from the data – it’s your data so you can see who’s running the big queries , where are the big spreadsheets etc? Try and understand the business priorities and how Bi can address them.
thanks again to Hugh McLeod