Appendix 1 – POSitive impactSM - Inside The System-wide Business Dynamic Model: POSitive impactSM is a custom, holistic approach to process optimization. With appropriate and intensive analysis of business constraints, combined with a clear definition of a business primary goal, we apply the mathematically sound principals of constrained optimization to deliver the best possible allocation of resources available to you. This optimization approach is based on a sound process model; some of the salient activities of this methodology that we believe are applicable to you are:
• Look for correlations and patterns in model variables; establish causality links where possible
• Look for feedback loops in system modules as they are the basis for change
• Look for uncertainty, quantify it with probability measure where possible; there is always an irreducible uncertainty in a complex business model
• Analyze data flows and data life cycles; establish approval and promotion rules for critical product artifacts
• Collect volumes of business data including 6-sigma data; normalize and turn it into a valid metric; where quantification is unfeasible, use terse expert consensus measures
• Establish the model time horizon and forward visibility, probably 5-6 years out; foresight is always imperfect and typically becomes useless at some point forward; hindsight maybe perfect but its applicability is limited to the future
We expect that you will assist LCS™ in absorbing a significant volume of specific knowledge such as existing practices, requirements, constraints, process metrics, and resources. Special recognition will have to be given to the non-deterministic; for example, the unpredictable nature of many your processes like approval cycles. They will be modeled using statistical techniques where only a partial confidence may be attached to the outcome. Such cases will require extra efforts to collect additional stochastic metrics (mean, variance, distribution etc) and to validate statistical hypotheses.
Once the model is complete, reviewed and approved, it will serve as the foundation for a set of solutions aimed at optimizing the your organization's global performance. In order to accomplish such optimization, a single optimization criterion will have to be selected and quantified. For example, one might attempt to (a) minimize the total average time to production, (b) minimize the total average time to sale, or (c) maximize profit over a certain time period. Once the optimization criterion, also referred to as the performance index, is selected, all other considerations become subservient to achieving its optimal value. These other aspects of system behavior will be modeled as system constraints. LCS™ will carefully identify and classify the system constraints, according to their nature, importance, origin, and relative durability. Some system constraint examples include product quantity; quality per region; risk of non-delivery; unacceptable lateness; budget; manpower and contractual commitments.
LCS™ will identify, and where possible, quantify all resources available in order to reach the optimal solution satisfying the set of stated constraints. Examples include: system structure; dependencies (both human and process); capacity allocation among modules; controlled redundancy and early warning to reduce risk and mitigate disasters; outsource and resource utilization.
Naturally, the wider the leverage provided by the resource set, the closer one could expect to get to the optimum point. Similarly, the more relaxed a constraint set, the better the solution.
Organizations producing product involve the interaction of human resources, internal and external, throughout the supply chain, and their computers systems. Given the likely high complexity and uncertainty of an organization's model, LCS™ will not attempt to achieve a mathematically precise, absolute optimality for a selected criterion. Instead, we will try to devise a heuristic algorithm (perhaps, a set of algorithms) that will iteratively (incrementally) approach the optimal solution thus providing an acceptable finite approximation at some point determined by a stopping rule. We expect to achieve the biggest performance gains in the beginning of such an algorithm and come up with a reasonable and effective stopping rule.
In the end, the quality of the ultimate solution presented to you will be a function of the following four major factors:
(a) stringency of constraints (b) availability of resources
(c) validity and accuracy of the model (d) validity and effectiveness of the optimization algorithm
LCS™ realizes that a project of such magnitude and complexity undertaken within a limited timeframe and budget will include some degree of irreducible risk. Every effort will be made to mitigate the risk and provide you with timely status reports on the optimization progress so that early corrective action may be taken.
