Business Intelligence Process Improvement and Requirements Strategy

Introduction to the process improvement strategy

The first ingredient of the business intelligence cocktail is the business goal/process that we want to support. The process improvement strategy provides guidance for identifying suitable business problems. 
In traditional application development, we start with the problem and then design and build a solution to address it. Business intelligence has a slightly different emphasis because we are trying to match the capabilities of the new BI technology with the existing business challenges. We may need to consider a number of candidate business processes before we find a good fit. Once we have matched technology and business process, we must also consider the steps within the process we can efficiently support. With this in mind, the four primary objectives of the process improvement pillar of the BI strategy are:

1. Define a structured approach for identifying process improvement opportunities
2. Define an approach for analysing an opportunity in terms of decision points and data
3. Create a framework for prioritising opportunities
4. Foster a culture where people question their own processes (as opposed to other peoples) and are rewarded for innovation

Business strategy alignment

Business intelligence works best when we align the project goals to the objectives of the organisation. Most successful companies will have a formalised business strategy in place. Managers at various levels will seek to implement it. A good starting point for a BI strategy is to take an organisation chart and augment it with these strategic objectives and supporting operational processes.

Aligning the BI strategy to business goals

It is helpful to see the whole organisation on a single chart. Strategic goals in different departments are often interrelated. Where interrelationships exist this is fertile ground for BI because it creates an incentive to share data and information across business units and departments.
We can use the organisation chart for reactive or proactive BI applications. For a reactive solution, we can map the hierarchy of business objectives and processes to a hierarchy of KPIs and metrics. The relationships between the different levels of the hierarchy will translate well into inactive reports or dashboards that give the user the ability to navigate from a high-level target to the underlying operational data and process level metrics. For proactive BI solutions, the organisation chart provides a list of candidate business processes with the additional assurance that the processes do actually support the primary objectives of the business.

Identifying a business intelligence opportunity

Having documented the business goals on the organisation chart, we can start to hone in on BI opportunities. Good BI solutions require those involved to think imaginatively about how they can improve the way they work. A brainstorming session involving people from different levels of the organisation is a very effective way to assess the feasibility of a project. Senior managers can provide the vision statement, access to resources, and clarity about objectives. Those in operational roles can provide guidance on the availability of data, nuances of business processes, and perceived roadblocks to change.

Business intelligence requirements elicitation

When I interview a manager or subject matter expert about requirements for a new BI project, I like to start by discussing the domain in broad terms. Often, the conversation will begin with the expression of an intractable problem that a dozen people have tried – and failed – to resolve. The stakeholder expresses the perceived solution to the problem in detailed terms, and with great optimism that finally – thanks to the miracle of BI – they will get a successful outcome. Alternatively, the stakeholder uses the problem as a challenge to BI as a product or methodology: ‘we have tried to solve this a dozen times, now let’s see if BI can do any better’.
It would take a bold ego or a bad case of hope over experience to think that there is an easy fix to something that has collectively baffled an organisation for a number of years. The best policy is to acknowledge the difficult nature of that problem and steer the conversation towards a high-level discussion of the business unit or department objectives.

Identifying opportunities for BI through an overview of business challenges

It is very likely that elements of the original problem will emerge during this more general discussion, but expressed in different terms and with far less emphasis on a solution (which you have already acknowledged is impractical). A BI opportunity is always easier to spot when we have sight of the full landscape for the business domain. When the opportunity comes into view, then it is time to have a more detailed discussion about potential solutions and the contribution they could make to the business. What is interesting is that you often get around to partially resolving the original problem. Typically, all problems encountered during a business process are in some way interlinked.

Removing defensive barriers to change

We all like to think we are doing the best job we can in the circumstances. Business process analysis requires us to step back from the natural instinct to say that our process can only improve if dependencies improve. Instead, we must focus solely on our own objectives.
A top down approach to process improvement analysis can help in these circumstances. If senior managers admit that they could do a better job, and how, then the barriers can start to come down at all levels throughout the organisation. If on the other hand, they start from the premise that others must improve first, BI adoption will be more difficult to sell throughout the organisation.

Decision point analysis

Once we have a candidate business process, we must identify decision points within that process. We may need to apply further decomposition to those decision points. Eventually, no matter how amorphous the original decision, we will get to a point where the decision will rely on some tangible data.

Mapping data to decisions

Having brainstormed the decision points, sub-decisions, and information requirements, we can now start to hone in on the relevant data. This is where we have to apply pragmatism to the solution requirements. Automating data acquisition is the most expensive part of a BI project. In the early days of your BI strategy, when you do not have the benefit of data integration from previous projects, the cost of data acquisition will be relatively higher.
For each piece of information required by the business process, we must determine the source and then approximate the effort of retrieving it. Source system experts and power users are normally the best people to guide our thinking here. With this analysis complete, we can determine which parts of the business process we can support.