Mr. Chandra

Model is usually a mini version of a physical entity, or a data/mathematical version of the entity. Simulation is usage of the Model to create outcomes under various scenarios. The above is in the context of Physical world: Example: We can create a model of a automobile gear, it can be to a scale say 1:10. This can give some insight into the time, effort resource etc to build. If we use the Model with some test scenarios (like speed, temperature)- we are in a way simulating and studying the performance... Having said that the most common paradigm in which Model and Simulation works today is in Digital world. Mathematical Models are built for a particular problem domain. and then Simulation is done on the Model with different data sets to predict the performance/outcomes. Example: If we have inflation data for past few periods- we can build a Model. We can then Simulate the inflation data for future periods using the model. In other words- Models represent the existing data- Simulation use the Model and predicts the future data.

We all are aware of the typical water fall charts/visualisation. It represents a series of bars in a hanging fashion that shows the movement/break up of a value. From my experience I shall describe one of the projects how waterfall is used: 1) For a particular business in professional services industry- we collected the head count data by various functional units. Waterfall is built showing the total headcount and the various functional units head counts that lead to this total head count. 2) We then did a T&M (Time and Motion study) and solutioned a certain head count reduction (by applying carious lean principles). The Waterfall is then enhanced to represent the headcount movement for each functional unit and the final total headcount of the business. 3) In my presentations- I even tried to put a time line on top of waterfall to showcase the headcount efficiencies that can be achieved Quarter over Quarter. Apart from this there are several other projects where waterfall model helped us to effectively present the data with leadership forums- Ex: 1) For a Global organisation- the regional/geography wise costs were represented with a split of fixed labor, contract labor and facility costs. 2) For an outsourcing project- the efforts for various sprints were represented in a waterfall to visually showcase effort comparison across sprints and the run up to the total project effort. To be honest earlier days making a waterfall used to be a tricky affair as I was trying to mimic MCK! Now a days with the Excel and other tools good plugins and readymade charts are available to make waterfall quickly!!

Cash with-drawl from a savings bank- Then-Earlier customer has to walk in to Bank, fill form, taken token from clerk, officer endorsement on form, cashier handover the cash after collecting token. Now- Customer either withdraws cash from ATM, alternately go to teller (single window) in bank and collect cash.

My view is Lean Six Sigma community (in current form) is good at working on Descriptive Analytics and Prescriptive Analytics. In simple words they analyze the past data and make meaning out of it (Descriptive Analytics) and also try to provide solutions for future business outcomes. (Prescriptive Analytics). However, when it comes to Predictive Analytics- I think both pure analysts and Lean community are doing limited value creation. Pure analysts (without appreciation of Business excellence) cannot really add value to the various possibilities that are thrown open by software. Lean Six sigma community by adapting to emerging technology can create tremendous value in the realm of predictive analytics.

I would prefer an error of False Alarm over an error of Missed alert. If an error happens in the form a Missed alert – we stand to lose out an opportunity to avoid a potential quality issue/ risk/service outage. An early warning signal is missed out. This could result in loss of credibility/warranty costs etc. If an error happens in the form a False Alarm- we tend to hit the emergency recovery procedures. We may end up evacuating buildings/stopping production lines/redoing the quality procedures etc. But we may end up with no actionable items. Fact is in both the cases- there would be cost implications. However, a Missed alert is a missed opportunity to fix something within our possible control/influence. A False alarm only end up in wasted resource/schedule and some frustration internal to an organization with no real damage to credibility/customer experience. Another way of looking at this is- A Missed alert has no immediate internal impact, but is a potential risk and can have external credibility impact. A False alarm is a short term internal inconvenience/internal impact and less/no external credibility impact.

We should not see two statements as mutually exclusive statements. Both of them are correct in different contexts. We need to measure things to manage them effectively. Especially we need measurements for trends, analysis and performance reporting. (Peter Drucker) However, if any business rely exclusively on their performance metrics and assume everything hunky dory some surprises are bound to happen. Let us consider an operations where the SLAs are always Green. If the operations head assume that the customer is super happy with the service and do a CSAT- it may throw some surprises. This we call it as watermelon effect. Where SLAs are Green and the CSAT is Red. The other example can be a product sales exceeding the targets for several quarters. If the business assumes too much of customer loyalty towards that product and doesn’t innovate enough- they are in for surprise when one fine day a better product enters the market. What we cannot measure in the above examples (and which is important for a Business) is Human behaviors, Relationships, Customer Requirements and Expectations. (Deming) So we should have good measurements in place to manage the Known- Knowns. At the same time we need to cognizant of the fact that some unknowns/unstated needs exist in the business which we have to manage even without Metrics.