Business Intelligence Capability

Key Points

  • Business intelligence capability will improve over time. Projects that are initially too complex to solve cost-effectively, will come into range as our knowledge of the data, technology and BI methods improves
  • Avoid expensive unpredictable projects at the start of the BI program because sponsorship will quickly evaporate if you fail to meet expectations

Business intelligence capability is useful because understanding how and why it changes over time should prevent us from picking projects that are too ambitious too early. Because it is so tempting to jump into the more complex projects, we may need several methods of communicating our business intelligence capability to impatient stakeholders

About CMBI

Business intelligence capability

The concept of business intelligence capability is useful because understanding how and why it changes over time should prevent us from picking projects that are too ambitious too early.
In the early stages of implementing a BI strategy we will be:

  • discovering a lot of information about key data sources
  • performing data extraction work for the first time
  • learning how best to apply the BI technology to the business problems

The combination of these factors throws up a plethora of risk, uncertainty, and inefficiency. The three pillars of the business intelligence strategy address the factors that initially limit our capability. Nevertheless, we need to appreciate the following three points from the very beginning.

  • When we start implementing the BI strategy our BI capability will be low and we should avoid long risky projects however compelling the need
  • Our BI capability will improve with BI project experience and at this point we will be able to take on the more complex projects but with a lower risk and more certain outcome
  • If early projects shoot for the stars and fail, sponsors will question the credibility of the entire BI program

.. Yes, we must deliver business value to the user community from the beginning of the strategy, but without taking on unreasonable risk. The following section introduces a BI capability chart I have created to communicate some of these concepts.

The BI capability chart

The Figure below is a simple way of visualising our ability to develop BI solutions of increasing complexity with the available BI tools and data.

i1
BI capability charts (Years 1-3)

The three charts describe the changes in our capability to build BI solutions as time passes and we complete our first few BI projects. The following sections consider the components of the charts and the implications for our BI program. 

Technology capability

Once we have completed the initial deployment of BI software we should get some quick wins from the capabilities of the technology with little additional effort or risk. The double-headed arrow represents the immediate benefits from the technology. As vendors release new versions of our chosen technology and we acquire other specialist tools, we will make further inroads into business process support; the double-headed arrow expands with the passing of time to capture more business processes as seen in the Year 2 and Year 3 versions of the chart.
It is worth noting that once you have implemented your business intelligence solution architecture you should not rely on further technology innovation to increase your BI capability significantly in the short to medium term. Game-changing technological advances are rare in BI and are generally the result of gradual product evolution, which has suddenly caught the imagination of the market, or been heavily promoted by a major vendor.

BI capability

The technology alone will only take us so far. After that point, our BI capability will determine the projects that are feasible now and those we should delay until our capability increases. The capability curve shows the relationship between the project complexity – encompassing the process and data complexity – and time it will take to execute the project.
In Year 1 (Figure above, top-left), the capability curve takes a steep and exponential turn upwards after the initial technology gains. This reflects the fact that we have an immature data strategy and will still be learning how to apply the technology to business problems. The practical implication is that many of our ideas for BI projects will be very unpredictable and potentially expensive to execute in the early stages of implementing our BI strategy.

Maximum acceptable project duration

We want to avoid expensive unpredictable projects at the start of the BI program because sponsorship will quickly evaporate if we fail to meet expectations. One practical way to reduce the risk of early disappointments is to set the maximum acceptable project duration.
In my experience of BI projects four weeks is a good initial maximum duration. The dotted line in the chart represents this constraint. Put simply, an estimate that greatly exceeds four weeks is likely to be very inaccurate. The steep curve above the dotted line represents the fact that small increases in project complexity can have an exponential impact on cost and duration. We should delay execution of any projects with estimates that exceed the maximum acceptable duration.

Increasing BI capability

If we prioritise projects that sit below the maximum acceptable project duration line, we will subsequently be in a better position to estimate projects of increasing complexity, whilst delivering tangible benefits to the business on these simpler projects. Each short project will increase our knowledge of the available data and BI tools whilst avoiding the risk of expensive project failures.
We see the impact of this increased capability in the charts for Year 2 and Year 3. In Year 2, the capability curve is not nearly as steep. This is because we will have identified and analysed a number of data sources, used the technology for real projects, and developed patterns for solving common business problems.
By Year 3, we will have worked with the majority of our data sources and become highly proficient in applying the technology to business problems. New projects will involve incremental improvements to existing solutions and applying proven patterns to new problems. The capability line becomes linear rather than exponential because we have all the information at hand to make informed estimates.

Using the BI capability chart

The chart is an informal tool that represents my experience of some of the dynamics of a BI program. We should not expect to provide exact cost estimates for each candidate project and prove the mathematical relationship between complexity and cost. Instead, we can use the chart to communicate why some business processes are initially better candidates for a BI solution. Equally, we may need to emphasise that other projects, however compelling the need, are simply too risky or expensive to tackle in the early stages of implementing our BI strategy.
Naturally, you will probably find that the most urgent business process improvements are initially above the maximum acceptable project duration. Do not be disheartened by finding your pet project on the wrong side of the line. Try to determine whether some part of the process can be supported using a shorter project that falls within the allowable project duration. The Business Intelligence Process Improvement Strategy provides some examples of how we might do this.

i2
Plotting candidate projects on the BI capability curve

The Figure above shows an example of how we might use the chart. At the beginning of the BI program, represented by Year 1(Figure above, left), we identify five candidate projects. We know that the capabilities of the BI technology are ideal for two of the candidate projects and feel confident that we can deliver them within a couple of weeks. We estimate that the three other projects are probably around six to eight weeks in duration, but acknowledge that BI projects of this duration must have some areas of risk and uncertainty.
By Year 2 (Figure above, right), several things have changed which alter the content of the chart. Firstly, a lot more candidate projects have appeared. This is because users have had a chance to use the technology and see the results of completed BI projects. The increased visibility of the BI program coupled with our efforts to identify suitable business problems will generate a large number of candidate projects.
Secondly, because the capability curve is not so steep, we can raise the bar on the maximum acceptable project duration. I suggest raising it to around six weeks. It is safe to do this now because we can be more confident in our estimates having gained familiarity with the available data and technology. The shallower curve represents the fact that a slight miscalculation in complexity will not have such an unpredictable impact on project duration as it would in Year 1 projects.

 

See Also