Supply chain planning analytics is a strategic and critical component of supply chain management. It drives aspects like raw material sourcing, manufacturing, goods delivery and return.

Need to Go 'Beyond Reporting'

Traditional supply chain planning is supported by reports generated by an enterprise resource planning system. The best result an ERP system can offer is historical transactional data and standard algorithms. ERP systems lack the ability to provide insights that optimize supply chain planning decisions and stretched lead times to understand interdependency among key performance indicators, which results in experience and gut-feel based optimized decision-making. Standard algorithms like demand forecasting techniques, do not accommodate the ever-changing product behavior during its product lifecycle (sudden growth, seasonality, demand stability, etc.) and the result can be an absence of future outlook and prediction features.

Supply chain planners are keen to link planning-related metrics to business critical KPIs such as On-Time In-Full (OTIF), sales target achievement, excess/low inventories, and actual raw material and logistics cost. Analytics can address these traditional supply chain planning challenges by providing solutions for future outlook generation, optimization, inter-KPI dynamics, quantification of impact of supply chain management metrics, and accurate forecasting. This leads to more informed decisions.

SCM (Planning) Analytics as a Concept

The supply chain management (SCM) planning department primarily targets accurate forecasts regarding product demands for the near future. Demand forecast figures further drive production, distribution, freight cost, and budgeting. Like any other business activity, inefficient SCM planning stumbles upon varied issues.

Sometimes, demand volatility manifests continuous over- or under-forecasting trends. A constantly biased forecast is detrimental for the supply chain.

Sometimes, there is low market serviceability. For example, a particular market demand for 1000 tons within a week was not met because only 750 tons could be dispatched on time, and the remaining only after 2 weeks. In other words, On-Time-In-Full (OTIF) is approximately 75 percent below the market threshold of 90 percent.

Sometimes, there is insufficient or excess inventory.  For example, a new medicinal soap was developed and launched via an intensive advertising campaign. The development and advertisement cost millions. Even though customers are excited about the new product, demand planning was under-forecasted. So, the estimated demand was less than the actual orders, causing insufficient inventory at distribution centers.  On the other hand, when demand is over-forecasted and there are actually fewer sales, the enterprise could end up with non-moving inventory, resulting in a sudden decrease in profits because of the expense of carrying this excess inventory.

Sometimes there is a misguided future outlook on the supply chain efficiency. For example, 100-percent order fulfillment is expected in the coming month, but the production plan shows insufficient planned production to meet the forecasted demand.

Finally, there are often inefficiencies in an organized and unorganized offline trade environment with multi-layered sales and interdependencies: Multi-layered sales implies a combination of primary sales, to private distributors, and secondary sales, to retail outlets.

Data analytics is a combination of business analysis supported by analytics techniques, such as exploratory data analysis, forecasting, regression, correlation and software tools. Dependency on different components of analytics varies from case to case.

Analytical capability can help to mitigate supply chain planning issues by way of analyzing sales data vis-à-vis supply chain planning KPIs and interpretation of trends and patterns. Few of the notable examples are analyzing different aspects of supply chain production efficiency, analysis of stocking norms and distribution efficiency, reviewing variations and inefficiencies in logistics costs incurred at various legs of stock movement until they reach the end consumer, utilizing statistical techniques to identify controlling parameters towards supply chain planning efficiency KPIs and quantifying the impact of individual parameters. For example, quantifying impact of demand drivers & adjustments on forecast bias using a regression model like forecast bias (positive / negative) is a function of base demand, demand drivers, adjustments, etc. Statistics-based customized forecasting approach for individual products, categorized into different scenarios like strong or weak seasonality, stability, and small packs also work.

SCM (Planning) Analytical Construct

The SCM planning analytical construct is comprised of hindsight, insight and foresight. It can be better understood through a consumer promotion scenario; thus the analyzing impact of supply chain planning numbers on a particular promotional event (such as the Temporary Price Reduction (TPR) on the promotion calendar that is typically scheduled during January and February). A 200-percent uplift is expected. For the desired outcome, supply chain planning should be aligned to promotion-related operations.

The following scenarios can be envisioned with some hindsight, insight and foresight. Hindsight, comprised of first-level analysis based on SCM cube slice-n-dice, e.g., visualization of supply chain KPIs during past instances of TPR. Insight, i.e., hypothesis testing of causal factors, e.g., does market serviceability support a positive uplift during TPR promotion type?  Finally, foresight, which involves developing an advance analytics model, e.g., what-if scenario to arrive at an approximate uplift based on SCM variables such as forecast bias, OTIF, and delivery compliance (%).

Best Practices for Supply Chain Planning Analytics Capability

The best practices for such capability include conceptualization of a SCM analytics cell for smooth churn-out of analysis requests via faster data extraction and cleaning methodology, documented analytics frameworks and analysis reports, and project governance.

Regular interactions between analysts and supply chain planners, in terms of the existing SCM process, SCM reports and KPIs, and business logic are important, as well as SLA-driven processing of analytics requests.

It is mandatory to measure performance of analytics capability investment in terms of incremental order fulfillment - volume and value sales. For this purpose, a business value logic standard needs to be developed jointly by planning and analytics team.  Business value logic will translate recommendations and insights into incremental sales/order fulfillment.

Conclusion

Analytics plays the role of support system for a supply chain planner by providing business specific insights in identifying and mitigating issues. It not only provides a means to act in the present based on the past but also provides a future outlook. Supported by data-driven insights, a supply chain planner will be able to improve forecasting accuracy, understand patterns, trim inventory fat, reduce stock-outs, and optimize raw material and logistics costs.

Source: Mindtree


Keywords: supply chain management, supply chain planning, supply chain solutions, supply chain management IT, forecasting accuracy, demand variability