Typical Daily Scenario – The SCM (supply-chain) Team has crunched its numbers using several excel sheets, pulled data from multiple sources and pivoted & re-pivoted to get all the scenarios. The management has now reviewed it, then approves the next steps and the company moves on.
But did the review happen on numbers which are reliable & consistent with each other or did each member approximate each other’s calculations, or assumptions.
In other words, if there are five persons involved, who are each very good & who individually prepared their worksheet, what is the error rate of the final aggregate numbers, based on which the decision was taken.
The five supply-chain persons here are representative of the Sales Team (Demand), Procurement (Supply), Finance (Cash-flow), Production (Capacity) & Logistics (Lead-Time).
If each team (or department) in this cycle are completing their task at 99% accuracy, the supply chain data network is performing at 0.99 ^ 5 = 0.95 or at 95% accuracy. In other words, there is a 5% probable error creeping in, when taken as a whole.
If the individual accuracy is now revised & taken on an average as 95%, the aggregate accuracy comes in at 0.95 ^ 5 = 0.77, and the aggregate probable error comes in at 23%. Given 10 projects each of value 100,000, this may result in approximately a value of 230,000 which may not be accurately reported or managed. Cause for concern.
The actual error rate scenario in many companies is similar to the second example, and usually only the Demand is reasonably accurate. The rest are moving variables all the time and require constant evaluation, re-planning. The complexity only increases with size, matrix of suppliers, credit terms, environmental parameters and all the rest of the variables.
It is pertinent to note here that most errors creep in due to outdated data, inconsistent reference points, limitations on using more parameters or simply wrong assumptions. The SCM-Team therefore is spending more time in fighting with the data, then in planning further or taking corrective actions.
This realization now leads to obvious next steps in figuring solutions. How may this be resolved or minimized.
In my opinion & based on solutions deployed, Process Automation techniques using machine learning variations is extremely useful, with immediate RoI.
Offloading such procedures of constant data gathering, interacting, verification & preparation & analysis of various events are best fed into data pipelines, validated & referenced as one single source of data. This is acted upon in an automated manner by Digital Assistants who are now part of the workflow.
Today’s technology can easily read from multiple data sournces (e.g. excel, pdf, emails, databases, web-portals, scans etc.) and very quickly crunch numbers, identifying patterns, discrepancies and generating the required report. Importantly, outputs can be compared with each other to figure out inconsistencies, assumptions and provide the best-fit recommendations.
Easy Automation Targets in Supply-Chain process include Quotations, Cash-Flow Forecast, AP/AR, Demand/Supply Gaps, MRP/BOM Sourcing, Freight Schedules etc.
Lastly & more importantly, these solutions are very cost-effective to customize, quick-to-deploy and the RoI is immediate. Definitely recommended if you are dealing with repetitive tasks, large scale data crunching all the time.
Interested on how to implement the solutions, pl. do contact us !!
References & Further Reading