Process, as per dictionary is "a systematic series of actions directed to some end". A set of questions arise here. What are these set of actions? What triggers these actions? Do they need some ingredients? Should these actions be done in a defined sequence or is any sequence ok? etc. To address these, Process is represented in a structured way as having:
- Entry Criteria (E) - The conditions that trigger the process to run. These are typically time-driven (e.g. every Monday 9am) or event-driven (e.g. whenever a new customer arrives)
- Inputs - Ingredients consumed by the process e.g. Description of service request
- Tasks (T) - Set of actions. To be defined in right sequence where necessary
- Verification Tasks (V) - A sub-set of tasks needed to verify if the steps have been conducted well
- Output - The product (can be a software work-product) obtained by executing the above steps
- Exit Criteria (X) - Set of conditions based on which the process can be considered 'Complete' e.g. When the Customer leaves the shop
- Measurements (M) - Base measures needed to analyze a process e.g. Effort taken, Cycle Time Taken to execute the process or a step etc. Note that the tasks needed to measure or record these measures also need to be defined as a part of Tasks
While reviewing a process, one can ascertain which of the the above elements are available. ETVX, ETVX-M are well-known abbreviations in the industry. Some common (sense) questions to review a Process:
- Are all the inputs listed, consumed by one step or the other
- Is each of the step playing a role in transforming the input to an output
- Are the tasks in right sequence, wherever it matters? For e.g. in a sanitization procedure, it is important to clean the hands with disinfectant solution before touching a masks. Reversing the step, one will wear a mask touched by hands not sanitized which defeats the purpose.
- Are there verification steps optimized? Theoretical best is "First Time Right" i.e. "Zero Testing and review and Zero defects". However in a practical scenario, verification tasks do exist but are optimal. "Under-doing" testing may risk the Quality of output and 'Overdoing' will result in excessive cost of process. (But is Zero testing a possibility?)
- Are all the products coming out of the tasks, listed as Outputs?
- Are the Measures collected needed and relevant? Note that "Collecting these measures" itself is a task, which consumes effort and can be seen as "Overhead" by practitioners if not utilized properly
No comments:
Post a Comment