To achieve each of these stages, it will be necessary to first gather sufficent information to act on - the specific techniques involved in data gathering and data modeling will be the subject of later sections.

First, we will outline the overall approach:

(i) Understanding the existing system

Most often, the system under development is intended to replace some existing system, so the first stage is to fully understand that system, including its strengths and weaknesses.

Ideally, after a data gathering stage, a full process model and data model can be constructed to describe the existing system. (The data gathering stage, process modeling, and data modeling will be discussed in the next several weeks.)

The clients/users of the system under development often focus their discussion on desired new features, and have many ingrained assumptions about what is "obvious" about the existing system. As such, creating a thorough understanding of what the existing system does is often critical for the systems analyst.

(ii) Identifying the improvement opportunities

Again, the process of information gathering, coupled with an accurate model of the existing system, should provide the analyst with the means to identify ways to improve the existing system.

Some of these will be based on client/user requests and suggestions, others may be proposed to the client by the analyst, based on their understanding of the system and what is currently technically feasible. Understanding of the business needs is critical if the analyst is to successfully advise the client, as such analysts with experience in a business domain tend to be highly in demand.

(iii) Develop and model the proposed system

The system concept will generally start out as a loose collection of improvement suggestions, and will gradually be revised and refined until a workable system concept is finally produced.

Ideally, analysis will produce several alternatives, along with recommendations and evaluations of their strengths and weaknesses. This will allow the client to select the approach they find most suitable, and proceed further from there.

In fact, for cost and time reasons, the collection of alternatives is often narrowed fairly early in the analysis process - focusing resources on the system/approach judged to be best for the client. (Again, this decision needs to be based on active consultation and interaction with the client.)

The final result will be a system proposal, which will outline the basic design of a system, a process model, a data model, and recommendations based on the perceived strengths and weaknesses.

The proposal will frequently be used as the basis for deciding whether or not the project should continue through design and implementation.

Since the overwhelming majority of analysis projects focus on business processes, we'll consider three specialized approaches to the three stages discussed above:

Business process automation

A business process is the way in which an organization operates, almost from an abstract point of view. For example, when you rent a video the process may include recording your customer id, recording the video identification, and deactivating the anti-shoplifting security on the video.

Business process automation refers to the cases where we are leaving the business processes essentially unchanged, but are introducing new systems to make the existing processes more efficient.

This could involve things such as:

• replacing manual steps with computerized ones ,
• adding computer support to processes, or
• improving existing computer systems.

For example, this could cover things such as replacing manual entry of id's with scanning cards or barcodes, replacing an old database system with a more efficient one, etc. This usually focuses on an improvement in efficiency for the users.

Strategies for analysis in business process automation:

(i) Understanding the existing system
This is critical for business process automation, since our goal is to improve the existing system without radically changing it. As such, the team will build very detailed process models documenting the activities and information involved in each stage of the business process.

(ii) Identifying the improvement opportunities
To identify how to improve the targetted process, there are two main strategies:

• Problem analysis: essentially asking the clients/users to identify the problems and potential solutions. This tends to lead to quick improvements in very specific features.
• Root Cause analysis: here we get the clients/users to identify the perceived problems, but we keep in mind that the feedback we get may describe the symptoms of the problem, rather than the underlying cause. Our goal is to improve the system by identifying the cause and addressing that.

  If your home furnace runs out of oil every month then a problem analysis solution might be "get a bigger oil tank", whereas the root cause analysis solution might be "fix the furnace".

(iii) Developing and modeling the proposed system
Generally in process automation the proposed system will be very similar to the "as-is" system, again with specific narrow aspects modified or replaced. For example, at a particular stage the data may be manipulated on the computer (new system) rather than on paper (old system).

As a result, often the model from the old system is simply copied and modified to express the new system.

Business process improvement

In business process automation we improved the efficiency of the process without making significant changes to the process itself. In business process improvement, we try to improve the process - usually by adding features or making a natural transition in some aspect of the system. For instance, in the video example, we might add a feature that allows the user to search for and reserve a video over the web.

Strategies for analysis in business process improvement:

(i) Understanding the existing system
Again, the proposed system must support most of the as-is system, but will be adding some specific enhancements or revisions. As such, understanding and modeling the existing system is again a critical stage, involving detailed process and data models.

(ii) Identifying the improvement opportunities
Since we are now trying to improve the business process (not just some computer support system) we need to actively seek out new business ideas and opportunities. This will involve the users, the clients, and the analysts.

We will consider four different analysis techniques:

• Duration analysis
  Here we examine how long it typically takes to perform each of the basic steps in the existing business process, and how long it takes for the process as a whole to be completed.

  If the sum of the individual stages is substantially smaller than the average completion time for the whole process, then there is probably an opportunity for improvement.

  For example: suppose a loan application involves four steps, with each individual step typically requiring an hour, but on average it takes 5 days to get an approval response to the client. Clearly there is a bottleneck in the process somewhere (the loan documents sit on different desks for hours or days) and an opportunity for improvement.

  The opportunities for improvement may involve integrating some steps to improve processing time, or parallelizing some steps to improve processing time.

• Activity-based costing
  This is very similar to duration analysis, except that we focus on the cost of each step instead of the time required. Effort is then focused on the most costly steps, in an effort to reduce overall cost.

• Informal benchmarking
  This involves studying how other organizations carry out a business process, to obtain ideas on how to improve your own. This is often applied for processes that are visible to the customer, since the customer is highly likely to encounter the "superior" processes at other organizations and make unfavourable comparisons to the client organization.

• Formal benchmarking
  This typically involves a formal sharing of ideas and process information with other organizations (not usually a competitor). The goal is to get management and staff to interact between the two organizations, study each others processes, and obtain ideas on how your own organizations' processes can be improved.

(iii) Developing and modeling the proposed system
Again, most of the business process is retained intact, and the model of the new process and data will reflect that. Typically only those sections relevant to the "improved" part of the process are remodeled.

Business process re-engineering

Unlike the automation and improvement strategies, in business process re-engineering we actually make substantial changes to the processes the business follows. This may go as far as a complete "from-scratch" re-design of the system.

Naturally, this is the most risky, time-consuming, and expensive of the three strategies, but it also offers the greatest potential for system improvement.

Strategies for analysis in business process re-engineering:

(i) Understanding the existing system
Because the goal here is to completely eliminate and replace the existing business system, relatively little time and effort may be assigned to understanding and modeling the existing system.

The existing system may be used as a tool for helping to extract requirements from the stakeholders, but the analysts want to avoid becoming entrenched in the design ideas and methodologies surrounding the old process.

(ii) Identifying the improvement opportunities
The key goal is the radical improvement of the business process, and as such requires a dedicated effort to get the clients and users to think creatively about the business itself.

Here are some approaches to stimulating that form of thinking:

• Outcome analysis
  Effectively, pretend you are the customer, and try to identify what the desired outcomes of the process are - completely ignoring how those outcomes are to be achieved. Once the outcome list is established, there may be alternative viewpoints on a process to attain them.

• Breaking assumptions
  Get the stakeholders to provide a list of "fundamental business rules" for their organization, and one-by-one try to challenge each of them. Systematically try to generate ideas to break each of these fundamental assumptions. Again, the result may be a radical new way of doing business.

• Technology analysis
  Develop a list of important/interesting technologies, and systematically try to determine how each could be applied to your organization.

• Activity elimination
  Look at each and every activity in the business process, and try to identify a way in which it could be eliminated.

• Proxy benchmarking
  Develop a list of industries that have similar structures to the business processes you are considering. Study what the impact (positive and negative) has been on their organization, and what they could have done differently to improve their process.

• Process simplification
  Processes often become needlessly complicated over the years because new stages and rules are added, and steps taken to make them "fit it" with the existing process. This is often more cumbersome than if the process could be redesigned from scratch. The goal is to eliminate complexities from normal day-to-day operations, and isolate them in small special cases.

(iii) Developing and modeling the proposed system
This involves the most extensive of revisions. Frequently the proposed system has little, if anything, in common with the original process model, and as such the new model is generally developed from scratch after extensive interviews and information gathering involving all stakeholders in the new system.

The analysis plan

Identifying the potential business value

Business process automation has the lowest potential return of the three strategies, since it generally focuses on improving efficiency in specific/localized target areas.

Business process improvement has a greater potential for return, being able to target both efficiency and effectiveness on a larger scale.

Business process re-engineering has the greatest potential for improved business value, since it seeks to radically change and improve the nature of the business.

Identifying the project cost

As expected, due to the scale of the approaches process automation tends to have substantially lower costs than process improvement, which in turn tends to have substantially lower costs than process re-engineering.

Identifying the breadth of analysis

The breadth of analysis is the extent to which processes must be examined during the analysis phase. While this is most heavily focused on business function, it may also need to include related functions and processes within the business, and even customer and supplier processes.

Automation tends to take a very narrow focus, dealing only with interactions around specific tasks or functions. Improvement must cover a wider breadth, usually focusd in one part or area of a business process. Re-engineering is highly likely to cover several major business processes and related areas/processes.