SFaaS: Boost Your Team’s Performance With Statistical Forecasting as a Service

  • March 30, 2022
Businessman walking in front of graphs on screen

Outsourcing of supply chain operations has been a common practice for years. Companies seeking to improve supply chain performance have frequently chosen specialized service providers to perform entire operational processes. Now, supply chain planning services offer an alternative to planning done solely in-house.

Statistical Forecasting as a Service (SFaaS), an outsourcing service offered by NTT DATA, provides an option for companies to quickly set up a team, processes and IT tools for statistical forecasting backed with comprehensive demand analytics.

You probably know how accurate your statistical forecast is. Have you ever wondered how good it could be?

Most companies measure and are well aware of the accuracy of their statistical forecasts. Regular accuracy numbers are often automatically generated at the SKU level in an Enterprise Resource Planning (ERP) or an Advanced Planning and Scheduling (APS) system. However, what organizations can’t measure is the size of the opportunity to improve statistical forecast accuracy. Uncovering that opportunity is the first step of the SFaaS service.

The first step in SFaaS is to conduct a forecastability study. Through many simulations — across planning hierarchy levels, market and product attributes and other relevant business factors — data analysts on the SFaaS team segment the portfolio by demand behavior. Based on this analysis, they identify the best forecasting hierarchy levels and methods for each segment. This work makes up the initial setup for the statistical forecasting process.

Are you able to explain your forecast?

Sometimes the implications of a statistical analysis aren’t fully clear to sales or demand planners. Random variation, exceptional ups or downs and other unexpected behaviors raise questions. Finding answers through the ability to drill down to the root causes needs to be an element of a statistical forecasting process.

A forecasting system shouldn’t be a “black box” that works its magic but leaves an audience mystified. In the SFaaS concept, exception detection and visualization are essential components. Besides the statistical forecast itself, each SFaaS run provides reports highlighting exceptions based on defined business rules. This capability helps to maintain continuing confidence in your statistical forecasts.

Configure once, run forever or improve continuously?

Letting an ERP or APS system generate a statistical forecast in an auto-pilot mode is a common practice. Many organizations trust that, by detecting historical demand patterns, systems can determine the best algorithms and generate a statistical forecast of the highest possible accuracy going forward. With stable market-demand-portfolio conditions, that practice can work well. Unfortunately, such situations aren’t the reality most businesses face right now.

Changes like product ramp-ups and ramp-downs, market disruptions and changes in customers’ buying behavior shake the demand of individual products and bring turmoil to the entire portfolio. From the statistical forecasting point of view, such dynamics call for repetitive analyses to assess changes and identify opportunities for tuning the forecasting model accordingly.

Continuous improvement is an integral element of SFaaS that ensures the forecasting model remains precise and in alignment with changes in the business. Clients also enjoy continuing service upgrades as part of NTT DATA’s ongoing product development efforts.

Do your data analysts actually have time for analysis?

It’s common that job descriptions comprise multiple roles, one being the data analyst role. This is adequate in jobs where employees can balance the workload between their roles while maintaining high overall performance. However, this rarely applies to demand analysts.

Advancing your understanding of market demand and improving forecast accuracy is a journey without end. Given how critical this journey is, you shouldn’t interrupt it every now and then by daily operations, which understandably have priority except in times of crisis. Another challenge these days is keeping those staffers that have grown into skilled analysts vis-à-vis current massive market demand for data analysts.

One primary benefit of outsourcing is the great relief that results from assured availability of data analyst talent and capacity. In fact, professionals focused completely on — and with a passion for — supply chain data analytics develop and deliver NTT DATA’s SFaaS. Our people serve organizations within multiple industries. They face a wide variety of business problems and continually find innovative solutions for them. The best of these become features of our SFaaS service and benefit all SFaaS clients.

How much are you prepared to invest in IT for better forecasting?

IT software and tools for demand forecasting come with distinct features and functions, technical characteristics and, of course, costs. They range from a standalone Excel spreadsheet to Artificial Intelligence/Machine Learning (AI/ML) digital planning and decision-support platforms. Statistical forecasting’s need for enhanced IT capabilities is often underestimated, as is the expense.

Another common problem with ERP and APS systems is that they don’t scale to support additional data needs as the company’s planning maturity level increases. In addition, integrated planning systems rely on existing data hierarchies and attribute structures that are designed from a functional integration standpoint and often driven by financial or operational processes. These structures rarely cover specific data elements that are essential for investment-grade demand forecasting.

Outsourced SFaaS comes built in with the IT capabilities needed by a statistical forecasting process. IT tools for data transformation, analytics, forecast generation, calculation of KPIs, data management and visualization and reporting are all included. AI/ML are also available. We design our SFaaS platform to support the entire journey, from the initial forecastability and segmentation study to the continuously developed intelligent forecasting process — all this with pay-as-you-go pricing with zero investment in IT hardware. The only investment required is in the interface software that extracts input data from its respective sources into SFaaS, then imports statistical forecasts from SFaaS to the demand planning system.

SFaaS frees your team by automating high-confidence forecasting so it can fulfill its primary responsibility of managing your supply chain. With a lightened workload, team members can dedicate more time to building Sales & Operations Execution (S&OE) and Sales & Operations Planning (S&OP) maturity while developing ongoing strategic plans.

Subscribe to our blog


Related Blog Posts