BI – What’s the Problem?
People mostly like the idea of Business Intelligence. BI provides the process and platforms from which critical data is aggregated, searched, presented and analysed. Up-to-date information is the lifeblood of business, with almost everyone wanting plenty of it with sugar on top, but there are some catches.....
BI can be complicated. To satisfy the full information requirements of users often large volumes of data need to be pulled from disparate sources and multiple technology platforms. As a result there are lots of technology considerations and also timing issues. Users often need to accept some degree of latency (out of datedness, as it were) with some types of information. For it to be of any use they also need confidence. “Does it all add up? Can I trust it?”.
These complexities can end up moving projects to a point whereby they start to become expensive, open ended and with the deliverables getting fuzzy as compromises are made. Because of this they often start with the best intention but end up being pulled down to the lowest common denominator of understanding, thereby addressing only ‘reporting’ requirements rather than outputs that help derive business understanding. Sure, this is a great way to burn cash and have the IT folks more toys to play with, but did you really set out just to re-implement the legacy on a different platform?
BI projects are often technically led, with a discussion and scoping sessions with users being the precursor to a large technical project where technical folks disappear into a development black hole and then come out the other end with ‘the solution’. Often users don’t like it as it doesn’t give them ‘their report’, and thus more work ensues.....cost overrun, scope creeps etc, etc. Nightmare projects with the end result is a further dependency on IT. Just why did we bother again?
And then there are the entrenched practices. The spreadsheets, the macros, those myopia inducing printouts.....BI was supposed to move you away from all these and leave users more self sufficient with better info........but the projects often just plain fail or only deliver a modest improvement disproportionate to the cost.
BI - A Sensible Approach
There is good news, after all, why would I be writing this article? You can deliver the benefits of BI whilst overcoming these problems following a few simple guidelines:
1. Projects should focus initially on understanding the types of question and problems a business faces, ascertaining what the best mechanism is to answer those questions based on audience, frequency, secondary questions, etc. i.e. not just ‘where is all the data’ and ‘let’s design a monolithic data warehouse’. To gain the correct understanding a lot of assumptions need to be challenged and creative ways of thinking explored.
2. Once you know where you want to go, project deliverables need to be prioritized and broken up into bite size chunks. Focusing on high profile wins initially gives the project good visibility, internal momentum and can secure continued sponsorship. Continued sponsorship has always been critical but it’s even more important in the current climate. These projects ultimately should represent value for money in moving an organization forwards.
3. The project should allow for plenty of education in BI basics and the end user tools (whatever they are) to ensure self sufficiency to draw outputs in the future e.g. reports, metrics, analysis and alerts. It is my personal belief that you just can’t deliver good BI without a basic understanding of the different types of outputs that BI can deliver, and what is good where. There is lots of BI functionality, for some very good reasons. It is best to start explaining this very early in a project to avoid a project getting too dependent upon one output type.
4. Get used to, and get users used to, accepting practical compromises. The number of times I hear ‘I need it real time’ in situations where it just isn’t relevant is, er, lot’s. It is derived from an expectation that good BI is ‘right now’, when really good BI is about providing understanding and answers. I always ask, ‘If you need it now, how now is now?’, with most people conceding that as long as they know when the data was prepared and that they force a data refresh then that is perfectly adequate. This is a great way of unnecessary IT infrastructure expenditure.
Project Approach
In ThinkTree projects, we focus on 3 key components in order to successfully deliver BI:
1. A structured project - This includes the normal project planning, scope, requirements, resources, documentation etc. with agreed deliverables to move you towards the BI thing. The project should be initiated with a requirements workshop within which we get the principles of BI across coupled with the dimensionality of your business and an understanding of data sources.
2. Data – We are going to need well managed and reliable data. Whilst you may well already have this it isn’t always a given. An understanding of data structures and platforms is required in order to overcome issues around the timing, volume and general trustworthiness of data. That understanding can be applied and held within some ‘dimensionally modeled metadata’, making it much simpler to take the information later as and when needed. It also makes your solution easier to manage. Please don’t make me explain that here!
3. Presentation - A way of authoring and consuming a variety of well presented outputs including metrics, analysis models and reports from the data above. As described earlier we like to focus on ensuring you are self sufficient in this area, and well managed data makes this as simple as possible.
With all of the above in place any businesses whether small or large can cost-effectively implement BI, either for small or large requirements. Increasingly, we see a need for organizations to start small and extend once a project has proven to deliver business value. At ThinkTree this is our preferred approach, quickly getting you moving and then extending to full-scale enterprise wide implementation as and when required. The aim of the project is to provide you with a BI platform over which you are the master, leaving you able to develop and roll out outputs as requirements change easily. Self sufficiency for your company and self service for your users.
To find out more visit www.thinktree.co.uk or contact us
