Adopting a new demand planning philosophy
Right now, all CPG companies should be planning an honest review of their internal demand planning efforts. The status quo is no longer a viable option; assess and implement the right changes, or risk being left behind.
Using the lessons of recent events as guidance, organizations should proactively put into place the tools and processes that will enable them to remain competitive for years to come, as unforeseen variables will continue to drastically impact consumption and the global supply chain.
The first step begins with your outlook. Adopting a fresh perspective on demand planning means placing the utmost importance on data and your organization’s capabilities to actively utilize it.
It’s one thing to possess rich data; it’s another to be able to handle it, ingest it, clean it, analyze it, gain actionable insights from it, and then realize benefits across the business.
The cherry on top is to be able to automate this entire process, from start to finish, which will have a significant impact on the efficiency and efficacy on both the demand planning and S&OP processes.
Now, let's dig in.
An outline of what we'll cover
To help CPG companies think about integrating a more efficient, agile, and accurate demand forecasting process, we’ll review the following topics:
- Why it’s so important to adopt a fresh perspective on demand planning, and a step-by-step guide on how to do it.
- The importance of data quality, quantity, and usability, and why “clean” data is essential to forecast accuracy and measuring error;
- How automation and artificial intelligence (AI) can powerfully transform data into actionable demand planning insights and quantifiable results across the supply chain and business.
- The game-changing, company-wide benefits of automation, AI, and forecast accuracy on the S&OP process.
- Why the current moment provides a valuable lens into why it’s vital to implement a demand planning process built on automation and AI at your CPG company.
It’s all about clean, high-quality data, and lots of it
Size and availability are no longer a barrier
The sheer number of variables that can affect demand is both diverse and incredibly massive. But there should be no barrier between the data a CPG company possesses and its transformation of internal and external data into a gold mine (in the form of efficiencies across the business).
In the past, some organizations have been hindered by the amount of data they’re capable of handling and processing. Nowadays, with vital datasets widely available, organizations should not feel overwhelmed by the size of its data pool, nor their appropriately lofty goals of deepening it. Rather, CPG companies should possess the technological infrastructure needed to ingest and analyze scores of datasets—or partner with a demand planning firm that does.
Datasets essential to sales predictions
The most powerfully capable demand forecasting models should ingest a wide variety of datasets for a more complete and accurate picture of future demand. But vital, historical sales alone (shipments at the SKU-level by customer, ship-to, and ship-from locations) cannot account for rapidly changing consumer behavior. As a result, forecasting models should constantly analyze not only supply data but also consumption-level data. This includes:
- Point-of-sale (POS) data (at chain and store levels)
- Internal data, including all pertinent sales, marketing, and supply chain information.
- Geographical patterns, such Google’s Covid-19 Community Mobility Reports, which cross-references foot traffic with consumption at grocery stores and pharmacies.
- Category trends, such as surges in demand for various essential goods.
- Own and competitive marketing / trade investment (promotions, e.g.)
- Search engine data. Patterns in search queries related to the world of CPG are telling. This can cover products, brands, retailers, e-commerce, and other terms and phrases that provide insight into consumption.
- Cross-channel social media mentions and trends. Consumers are constantly on social media platforms. They interact with brands directly by posting their own content or by reacting to a brand they follow.
- Health and ingredient trends. Yesterday’s kale is today’s charcoal. Or meatless protein. Or gluten-free bread. This can help brands not only forecast demand but also develop or amend their own new product development or distribution strategies.
- Weather patterns
- Seasonality and other fluctuations related to time
It’s all about data quality
All CPG companies possess large amounts of data—some of them vast—but quantity is not a mark of richness or usability.
Simply put, demand forecasts are only as good as the data used to calculate them: 4,000,000,000 incomplete or mismatched data points is exactly that. As they say, “garbage in, garbage out.”
In short, it’s all about quality. So when data is “dirty,” it must be cleaned. And the more data there is, the higher the likelihood of errors existing in the data pool, which in turn can negatively impact the data's usability (to say nothing of forecast accuracy or its benefits).
Eliminating data errors
It deserves repeating: Every bit of data that’s used in a forecasting model should be “clean.” Clean data refers to all internal and external data—SKU numbers, BOM quantities, warehouse addresses, historical shipments, market trends, POS data, and any other piece of information essential to your supply chain and S&OP process—that’s organized, timely, precise, complete, consistent, and generally error-free.
It would take one human being (let alone a team of them) a very, very long time—a lifetime, even!—to sift through a company’s data and check, double-check, and triple-check the viability of every single piece of data. It’s a human impossibility, and this is before any sort of analysis is completed or insights are drawn. So what’s the solution?
How automation and artificial intelligence (AI) transform data into results
Automating quality: the data-cleaning process
Clean, rich data sets the stage for machine-learning analysis, which in turn drives actionable insights. As a result, CPG companies should possess substantial data computing capabilities that scientifically clean and process raw daily sales data. They can automate this process by applying cleaning algorithms that use a data mapping process to identify and correct various trends in error, such as human entry mistakes or those made by a third-party vendor whose data quality is not up to par with your needs and standards.
How does artificial intelligence analyze data?
And at a time when data exists in near-infinite amounts, artificial intelligence is a vital tool for organizations in the CPG industry. When specialized data is incorporated into machine-learning models, the more they learn and the more accurate they become. These models form a neural network whose speed, dynamism, and computing capabilities trump any human’s ability to ascertain such complex relationships between datasets. Eventually, these trained algorithms behave like a powerful demand planning AI tool.
Converting raw data into actionable CPG datasets
Behind the scenes of automation, predictive analytics models are calculating dynamic relationships between humongous datasets. Expert data engineers tweak and iterate models in search of locating clear-cut demand drivers, based on consumer behavior and market signals, while giving weight to top-selling SKUs.
Between good data and commercial benefits is the process of understanding how data translates into dynamic insights about your supply chain, business model, and the CPG industry at large. In order to open the "black box" on demand and eliminate blind spots, it’s vital to be able to reveal the factors influencing past and predicted sales. This is the job of artificial intelligence: to analyze data and find dynamic relationships between datasets across various time horizons. (Even better is a platform that presents these insights to you in a centralized and ranked fashion, available immediately at the click of a button. That’s the definition of truly accessible automation.)
Datasets that provide dynamic insights
- Historical sales & seasonality data. There’s an immense wealth of data contained in raw day-to-day sales patterns, including the relationship between seasonal trends and long-term growth.
- Historical inventory data. This information can include quantity on hand, quantity on order, weeks of supply, service levels, and more. For example, the analysis of quantities ordered vs. quantities shipped is vital to understanding sales that could have occurred vs. sales that actually occurred—due to insufficient inventory.
- Product transitions / maturity data. SKU numbers and/or UPC codes often change during a product’s lifetime. This informational trail provides vital data about a product’s history, especially during the process of establishing and analyzing time relationships between SKUs and UPCs. Other aspects include the size of a label or package, or any other bit of data that’s usable. If it has the potential to impact the business, it should be incorporated.
- Obsolescence and discontinued product data. Data detailing product sunsets is vital to understanding the demand of comparable products. Running conversely with product (or component) obsolescence is the pace of innovation and speed to market for many CPG products and product categories. Gaining a clear basis of understanding as to the reason a product is discontinued—such as innovation, new claims, a product enhancement, or poor performance—is essential to better predictions.
- New products and distribution data. Machine-learning models should be programmed to analyze information related to planned (future) product launches, expected increases in pipeline shipments, and distribution gains (versus “regular” consumption levels). It's difficult to produce accurate predictions for new products unless it is known what there is to predict. Similarly, it’s important to remove this data for future forecast calculations. Distribution change models can better predict the sales impact of distribution gains in the base business (pipelines from everyday consumption), distribution losses, and new product distribution gains. It's also vital to understand in-store dynamics, such as the number of facings, shelf placement, multiple placements, shopper marketing activities, competition placement, etc.
- Marketing investment data. It's vital to establish the dynamic numerical relationship between promos and changes in sales volume and velocity (lift/drop). Promotional data is important for your own promotions as well as competitive promotions. Considering promotions and other market spend allows us to enhance our models by incorporating another dataset.
It bears repeating: Only artificial intelligence designed specifically for nuances the of CPG can get the job done right.
What underlies data that drives growth and efficiencies is the ability of demand planning AI to translate data into CPG- and supply-chain-related insights. It’s the work of a team of data engineers to build and iterate predictive models to ascertain these company-specific insights and segment them in a way so that they can be applied across the business.
The game-changing, company-wide benefits of Demand Planning AI
Engineered for agility, efficiency, and accuracy
The value of demand planning AI extends well beyond its ability to ingest, handle, and analyze multitudes of data. If recent events are any indication, organizations must be able to rapidly recover from an outlier event (like Covid-19), which can quickly throw off their commercial plans and disrupt supply chain operations (let alone an entire industry). Artificial intelligence enables demand planners to automatically account for new market variables and regain accuracy levels. It adjusts—or, rather, corrects itself—with each passing millisecond; that’s the power of automation. Demand planning AI is also a tool that’s trained to be agile and extremely efficient.
AI and accuracy at work: Realizing benefits across the supply chain
The best zero-touch demand planning AI platforms also offer highly accurate predictions. They provide actionable SKU-level insights that identify where risks and opportunities may lie in the future as well as strategic recommendations for immediately realizing efficiencies across the supply chain. Benefits include working capital improvements (such as a reduction in safety stock levels), improved service levels, reduced transfer costs by moving inventory from warehouses with lower demand, and reductions in logistics overhead. And that’s only the beginning. The following are some cross-functional benefits of forecast accuracy:
- Accurate sales forecasts help VPs, senior managers, and other executives to communicate consensus forecasts—as well as their department’s capabilities and/or restraints in hitting those targets—to executives and other stakeholders.
- Accurate distribution forecasts enable distribution teams to help sales teams. When sales teams are provided with visibility into customer inventory levels, they’re able to incrementally upsell the buying and planning teams. The best distribution predictions come with a Customer Scorecard, which enables distribution teams to predict their entire distribution footprint and adjust distribution schemes accordingly. They can optimize service levels, lead times, and DC splits.
- Accurate inventory forecasts reduce costs associated with working capital, transfers, obsolescence, and more. Accurate Inventory predictions help manufacturing teams plan while enabling procurement teams to negotiate better prices, MOQs (minimum order quantities), and delivery dates with suppliers. They also help teams manage working capital, in part by negotiating volume cost breaks for raw materials, components, and finished goods. When inventory levels are optimized, stock-outs, dead stock, and spoilage can potentially be avoided, and warehousing costs can be lowered. Supply and inventory management planning are key elements of effective demand planning.
- Accurate manufacturing forecasts help to maximize the efficiency of line capacity and Capex (capital expenditure), reduce obsolescence, take a stronger position in negotiating with suppliers, and more. Manufacturing forecasts enable plants managers to schedule production runs efficiently across all locations, and determine potential bottlenecks (in real-time) that could impact their ability to reach target production metrics, including throughput, capacity, and utilization rates.
How automation can transform your S&OP process
Reviewing the S&OP process as we know it
Though no two S&OP processes are the same, the end goal is clear: to formulate a plan that maximizes sales and profits. But the process of integrating demand, supply, sales, marketing, product, distribution, and financial teams can become bogged down by the amount of time, resources, and overhead it takes to align on a demand forecast. This process is further complicated when team benchmarks are misaligned data, perhaps tied to departmental KPIs rather than total company; when data is siloed amongst teams; or when systems are varied and individualized. This can cause confusion and informational discrepancies, not to mention integration challenges. But it doesn’t have to be this way.
The benefits of automating demand forecasts
When demand forecasts are automated, and sales data is visible across an organization, S&OP meetings can begin from a place of alignment rather than difference. One source of truth, if you will: a singular, 360-degree view of business across all areas of the supply chain in real time. Furthermore, forecasts are produced with the click of a button (rather than the grind of a spreadsheet and alignment meetings). Here are some ways automation can benefit the S&OP process:
- Increased efficiencies. S&OP is a melding of the minds: Multiple stakeholders from various departments come to the table to align on a demand forecast. But data between business units can be varied or siloed. With an automated demand planning process, brand data is universal and agnostic. There is one single source of truth. This makes cross-functional team dynamics stronger and more efficient, pushing the decision-making process in the right direction, faster.
- Full visibility. Having a complete picture of sales and inventory data is essential for creating a demand forecast. Automation enables that.
- Refocused sales teams, increased employee efficiency. Less time spent on demand planning-related tasks frees up staff to put their energies into other growth efforts, enabling sales teams to focus on growing the business rather than spending their time manually tabulating forecasts .
- Assurance during uncertain times. Because leadership can trust Unioncrate’s demand forecasts, they can confidently allocate resources in areas that would maximize profits and operational efficiencies, such as investment in products with higher profit margins.
- Economies of scale. Among other benefits, forecast accuracy helps brands prepare to take advantage of new sales opportunities. As companies scale, profits should increase relative to increases in sales volume. The larger the output, the lesser the costs. Meanwhile, supply, R&D, and distribution costs will decrease per unit sold, while profit margins increase.
Conclusion: Consumers are evolving. Your demand planning process should, too.
There's no time like the future
It's no secret that the CPG industry is experiencing a once-in-a-generation shift. The impact of Covid-19, for example, has touched nearly every facet of the CPG supply chain while accelerating drastic changes to consumer behavior, including further adoption of eCommerce across the board.
An entire generation of online-native shoppers with enormous logistical expectations wasn’t created out of a vacuum. By all accounts, a consumer evolution was inevitable; Covid-19 sped it up.
For many organizations, it was mayhem. Retailers, distributors, e-Commerce platforms, logistics companies, warehousing firms—you name it—were more or less shooting in the dark. For many manufacturers, inventory simply fell short and countless companies missed out on sales; demand predictions were typically way off, inaccurate to a fault. And as service levels dropped, stress levels rose—from the production floor to S&OP meetings, where forecasting demand accurately was a serious challenge.
The pandemic laid bare just how quickly consumer behavior can change, further exposing a pressing need for organizations to be able to recover rapidly and adapt accordingly when faced with unforeseen circumstances and supply chain disruption.
In short, when the future meant more than ever, it was increasingly difficult to sift through a jungle of data, analyze it, and accurately forecast sales. During the most challenging times it's vital to have the tool in place to overcome obstacles like "too much data to handle" let alone an inability to gain actionable insights from it.
As you navigate these new and often unprecedented market realities, it’s important to have that perspective, and we hope that this guide will continue to provide answers and illuminate an actionable path forward for years to come. That path is paved by automation and artificial intelligence built specifically for CPG. Both are essential to an effective demand planning and S&OP process—now and for the foreseeable future.