<img height="1" width="1" style="display:none;" alt="" src="https://dc.ads.linkedin.com/collect/?pid=358169&amp;fmt=gif">

Quick Response Forecasting: A Blueprint for Faster and More Efficient Planning by Larry Lapide

Books on desk in library at the elementary school

With an increased demand fluctuation which could be caused by social media, this article discusses the advanced concept in demand forecasting, which IBF is calling quick response forecasting (QRF). Dive into examples of downstream demand signals, the idea of quick response supply teams and QRF forecasts, as you read this article from Dr. Larry Lapide

EXECUTIVE SUMMARY

This column is written to start a dialog of an advanced concept in demand forecasting, which the IBF is calling Quick Response Forecasting (QRF). It is predicated on the premise that in order to take advantage of “predictive” downstream demand signals (such as from social media) that might portend an impending demand spike or other rapid change in demand; today’s forecasting and planning processes are too slow to detect and respond to them. Quick response supply teams, put in place on an ad hoc basis, would need QRF forecasts to capitalize on revenue opportunities that “predictive” demand signals offer. In the QRF process, monitoring signals on a real-time basis will be needed to alert these teams, as a call to action. Support of Business Continuity Management (BCM) teams put in place for severe hurricanes, as well as other short-cycle time processes are discussed to illustrate some processes that might benefit from QRF.

Dr. Lapide writes:

Some time ago I was on an IBF Advisory Board conference call when members were discussing situations where social media data were signaling rapid changes in demand for products. These data might, for example, include some favorable online reviews of products, as well as a rise in the consumer intent to try or buy them. I listened for a while as members stated these were great “predictive” downstream demand signals, but since they were signaling rapid changes in demand, members believed their current forecasting processes could not incorporate the information quickly enough to act upon it. I blurted out: “We need to start thinking about Quick Response Forecasting (QRF) techniques,” making up the term. While all agreed that the term captured what was being discussed, everyone pondered: What did it really mean? The Board decided to include the topic for discussion at the IBF’s 2017 Leadership Business Planning, Forecasting, and S&OP Forum that was held this past October in Orlando, FL. I anticipated that the demand planning panel that I was moderating would address two issues regarding QRF. The first issue was to deal with defining QRF and starting with the concept of “updating forecasts in line with ‘real’ and rapid changes in demand, both during and between planning cycles.”  The issue was that if most companies were on monthly planning cycles, how long would it take for forecasting processes to incorporate a rapid change or a short-term spike in demand? For example, at a previous Forum, one participant described a case where, at one of her concerts, Lady Gaga wore a unique color of nail polish that his company sold. Shortly thereafter, social media lit up and sales “went through the roof.” The company and its supply chain partners ran out of the product, as well as a key ingredient that went into making it. The company felt it was not prepared to take full advantage of the rapid change of demand for this nail polish, and likely lost a significant revenue opportunity. The second issue was to deal with “enabling QRF by leveraging predictive analytics, social media information, and other Big Data.” This relates to the explosion in digital data and the enormous amount of information available about customers and product users on the World Wide Web. As data get bigger, companies are looking for techniques and methods to both find and incorporate a few key demand signals among Big Data’s noisy information deluge. Following this year’s Forum, the plan was for me to use the discussion points as input to write this JBF Winter 2017-2018 column. Unfortunately, my demand planning session ran out of time before it could adequately address QRF. Thus, I am using this column to give my initial thoughts on the topic, in hopes of having others build on it in future JBF articles and IBF presentations. In this regard, Eric Wilson, one of the IBF’s thought leaders has agreed to make some preliminary comments, which are at the end of this column.

A PAPER-CUP MANUFACTURER  MIGHT BENEFIT FROM QRF

Since 1998 I have given numerous tutorials on “Designing a Business Forecasting Process” at the IBF’s academies attended by practitioners, and during MIT course lectures that were attended by many graduate students. The tutorials involve an initial presentation on process design principles, which is then followed by breakout sessions for students to work on, and report back on, a fictitious business case involving a manufacturing company, Cups4U—which makes and markets a simple product, paper cups. I wrote the case to exaggerate many of the things that I had heard could go wrong with a business forecasting process, since many companies acted as if forecasting only involves statistical forecasting techniques. Thus, when implementing a forecasting process, too often little consideration goes into the particulars of the process itself. Cups4U sells cups through three sales channels: 1) mass retailers and warehouse clubs with a private label, 2) fast-food restaurants that buy cups with their branding on them, and 3) supermarkets, and mom-and-pop grocery stores that sell the product with Cups4U’s name on them. It turns out that the company had installed a forecasting computer system the year before, and shortly thereafter, inventories went through the roof. A consultant’s report pointed to a variety of issues with how the forecasting process was working. In fact, it was working poorly because the process hadn’t changed at all during the system’s implementation. The students address designing a new process, including what types of forecasting methods might be appropriate to use. Generally, they propose methods that vary by channel. The third channel is the simplest because using time-series methods that identify trend and seasonal patterns is sufficient, given the insignificant amount of papercup promotions run in this channel. The first and second channels would additionally need to put in place coinventory management programs, such as Collaborative Planning, Forecasting and Replenishment (CPFR), because both channels buy private-labeled products from Cups4U. However, what students usually miss is the fact that fast-food customers sometimes act like “fashion” businesses, subject to unexpected spikes in demand, or a lack thereof, for branded cups. I point out that fast-food chains, such as McDonalds, co-promote youth movies, such as Star Wars, Legos, and Transformers, by displaying moviethemed figures on beverage cups during movie launches. Thus, the sales of these cups depend on the success, or lack thereof, of the movie itself. Recently I’ve been pointing out that tracking social media information can help predict sales of these types of cups, as well. A “hot” movie launch means that a fastfood company will need more cups than expected, while a movie launch that is a bust signals they will need much less. Interestingly, this type of information can be obtained starting from the first day of the launch. Moreover, it includes movie attendance data and what is trending on social media about how the movie has been received by critics (such as on Rottentomatoes.com) and audiences. Thus, with these “early predictive” demand signals, the papercup forecasts can be updated in time to significantly modify production plans, either up or down.

QRF CAN SUPPORT SHORT-CYCLE PLANNING

However, like in the nail-polish demand spike mentioned above, often supply chains cannot act quickly enough to take full advantage of the opportunity that the demand signal offers in generating extra revenue. In my Fall 2007 JBF column, “How Often to Forecast,” I pointed out that more frequent forecasting has the potential to increase forecast accuracy in terms of identifying rapid changes in demand. However, supply chain responsiveness may be too sluggish to take full advantage of a forecast that includes them. For example, manufacturing managers might complain about getting whipsawed by demand forecasts that change rapidly—despite their increased accuracy. One quote I often use with regard to planning responsiveness came from a manager (I’d interviewed) who ran the S&OP process at a high-tech company. Generally, these types of companies operate “responsive,” in contrast to,  “efficient” supply chains. “Responsive” supply chains handle high-margin, high-value, and possibly fashion products. Their major goals are less about minimizing operating costs and inventories, and more about maximizing inventory availability at the point-of-sale/consumption in order to capture potential upside revenue. The latter “efficient” supply chains handle mature and lower-margin products, and thus are goaled on minimizing operating costs and inventories, and worry less about lost sales for these types of products. This high-tech manager was planning a “responsive” supply chain. When I asked him what kind of S&OP process he needed, he simply stated: “I need a process that is able to chase demand or supply quickly.” His goal was to strive constantly for a highly “responsive” planning process. If his company’s commercial (i.e., demand) side identified a significant revenue opportunity not previously planned for, the supply-side of the house was required to quickly do its best to source the supply needed in a timely fashion. Similarly, if his company’s supply side identified a supply issue, such as a sudden and unexpected excess in the inventory of a product, the demand side of the house needed to quickly put sales programs in place to sell more of it. While the manager ran a fairly responsive monthly S&OP process, I suspect it would be too slow to take full advantage of QRF. The planning processes that might take such an advantage would have shorter cycle times than monthly, and possibly even shorter than weekly. In the heydays of Dell, it had a best practice that I used to refer as a “mini S&OP process” with a daily planning cycle. It would have been a process that might have benefited from QRF. At the beginning of each day, managers would conduct a supply-demand matching meeting to align “vendor-consigned” inventories with items being promoted and ordered on the Dell website. The major input to this meeting was the on hand and on-order inventory positions of components used to assemble systems for customer shipment. If a component had a surplus of inventory that day, the managers would heavily promote configured systems that included it—such as by drastically reducing the price on them. On the other hand, if a component had little to no inventory left, the managers would no longer promote configured systems that included the component—such as by drastically increasing the price of these systems, as well increasing their order lead times by pushing back expected delivery dates. The reason why this process could be done daily was because Dell could quickly change what was being promoted on its website that day, virtually immediately following the mini-S&OP meeting. Thus, this process might have benefited from QRF. HP is another high-tech company that also runs a “responsive” supply chain, and it too might have benefited from QRF. It had developed innovative “flexible” procurement contracts for sourcing scarce components embedded in new computers being launched. Based on pre-launch-range forecasts of demand for the components, its procurement managers would set up two types of formal contracts with suppliers. “Fixed Quantity Contracts” would involve a set amount of components to be purchased, with long lead times and the lowest cost to HP. The fixed quantity would be based on a low-demand scenario, representing an amount HP was certain to sell. It would purchase these components well before the computer’s launch. “Flexible Quantity Contracts” involved a commitment to purchase up to a specific amount of components with shorter lead times and higher pricing, because the supplier was not guaranteed HP’s business. The up-to quantities were set based on the difference between a base-demand (e.g., most likely) scenario and the low-demand scenario used to set up the “Fixed Quantity Contracts.” In essence, the suppliers on these flexible contracts would only get business if a new computer’s sales exceeded the low-demand scenario. This is why HP paid more for the components, and got shorter lead times on them. Lastly, HP made no commitment to suppliers for any amount of demand exceeding the base-demand scenario. Should it be fortunate to have demand exceed the base-demand forecast, it was willing to buy components in the spot market; paying much higher prices, but getting components fast enough to help meet the upside-revenue opportunities afforded. Procurement managers responsible for purchasing on the expensive spot markets would need accurate QRF, to keep pricing to a minimal. For example, if a QRF demand forecast triggered needing to buy on the spot market too late in the launch, components would be needed sooner, and would be more expensive, than if a QRF forecast triggered it much earlier during the launch period.

QUICK-RESPONSE SUPPLY TEAMS REQUIRE QRF

Generally, QRF would not be needed in support of S&OP processes put in place to “tactically” plan for full product-line demand. Since S&OP processes are disciplined, routine, and have monthly/weekly planning cycles, they would be hampered by having to incorporate a demand signal that portends a very significant and rapid change in demand for a small portion of a product line. QRF would be better suited to support a planning process put in place to manage unexpected and non-routine episodes of rapid demand changes and spikes. This type of quick-response supply planning process would be “management by exception” in nature. The process should be managed by ad hoc teams formed to act quickly to garner as much of the revenue opportunity afforded. An example of a process that might need support from QRF is the Business Continuity Management (BCM) process. This type of process is not routine, as it is initiated from a major disruption that has significantly affected business operations. This disruption might be caused by natural disasters, such as extreme weather-related events, earthquakes, and wildfires. A disruption of this type only impacts a portion of business operations, rendered inoperable for a period of time. Often companies assemble a cross-functional “response team” on an ad hoc basis that is charged with bringing operations back to normal business levels. As such, these teams are goaled with getting operations back to normal as quickly as possible, and with minimal revenue loss. Having come off an extreme hurricane season in the United States, let me consider the processes put in place to help the recovery of business operations and living conditions in the affected areas. First, these processes are supported by a good QRF organization, the National Weather Service—it has gotten very good at forecasting the severity and the path of hurricanes. Once forecasted, companies start assembling their response teams. Some response teams might be focused exclusively on BCM, and start to project and plan for the recovery of operational disruptions caused by the impending hurricane—such as its impact on plants, warehouses, transportation, and workers living in the affected areas. After the hurricane, these plans are updated to account for the actual damages that took place. Recovery operations are then initiated and monitored over time until business operations reach stability. QRF would be needed by these BCM teams to update recovery plans. Other response teams, involving companies not expecting to experience any business disruptions from the hurricane itself, are focused on projecting and planning to sell the goods that will be needed by customers during recovery operations. For example, companies (such as Wal-Mart and Home Depot) that sell basic products such as food, shelter, clothing, and medical goods, start planning to provide emergency help by pre-positioning inventories of these items close to the affected geographical areas. Their teams also need to update supply chain plans throughout the recovery period to ensure help is provided throughout recovery. QRF, incorporating “predictive” demand signals based on the real-time conditions on the grounds of the affected areas (e.g., from daily news reports), is needed by these teams to update supply plans.

FINAL INITIAL THOUGHTS

Since the early 1990s, when WalMart launched its Retail Link System to share point-of-sale (POS) sales data with its suppliers, I’ve been researching downstream demand signals, in the context of improving a supplier’s forecasting via multi-tiered forecasting and planning. Over 25 year later, I’ve come to the conclusion that the industry believes that POS data was always too detailed and cumbersome to work with, especially for the lion’s share of a company’s product line. It appears to have only been useful for products that experience significant changes in demand from time to time, such as during promotions and new product launches. Thus, POS information was rarely, if ever, meaningfully incorporated in a formal way into demand forecasting and S&OP processes. It was incorporated on an exception, ad hoc basis. Some of this is due to the fact that POS data is not especially “predictive” anyhow, since it comes from actual sales at the retail store-tier of supply chains. I have hopes that social media data will be more “predictive” because it comes from the consumer-tier, and includes consumers’ intent-to-purchase as well as satisfaction with product use and the consumption of products. This type of information is not just about actual purchases, but future purchases. Thus, I think QRF using social media information is more promising than POS data. Below are some of my concluding thoughts about QRF supporting quick response supply planning processes, such as the hurricane-related processes discussed. First and foremost, we need to recognize that forecasts are only useful if there are planning processes that use them. A typical S&OP process is a routine planning process that would be too disrupted by having to incorporate QRF, so it is not a good candidate process. QRF is needed to support teams that are specially put in place, on an ad hoc basis, to manage significant event based and substantial demand changes when financially justified. These teams should be cross functional and be supplemental to the S&OP process. The teams need to be quickly assembled once an on-going QRF forecasting organization detects that a demand spike or significant demand change has, or is likely to occur. Once the team is put in place, QRF forecasts for the event need to be continuously provided to the quick response supply team. If forecasters ask managers whether they need QRF today, they will likely say no. They don’t want their operations to be whipsawed by frequently changing forecasts. This is why a separate and special quick-response supply process will be needed to handle each event. If you are on the lookout for an organization to partner with to develop QRF and supply response teams, your sales organization is the best bet. The teams are focused on going after significant revenue opportunities, for which current processes are too slow to take advantage of them. A supply response team will be goaled on taking full advantage of an opportunity, in terms of squeezing as much revenue from it as possible. Isn’t it the very type of opportunity that sales is responsible for chasing after?

COMMENTS FROM LUMINARY ERIC WILSON

For some time, return on investment (ROI) has traditionally been measured by cost savings. More and more these ROI models are beginning to be replaced with newer agile ROI models. In every changing business environment and even more competitive omni or multi-channel landscapes, companies are looking for a competitive advantage. They are focusing on what opportunities and sales they may be missing out on adapting faster than the competition. The new ROI model looks at the responsiveness, flexibility, and speed to manage the change on a daily basis. Unlike the lean supply chain, the agile supply chain uses real-time data and updated information, and is the perfect opportunity and need for QRF. Quick Response Forecasting is responsive with the ability to sense changes and translate chaos into usable demand signals. QRF is flexible enough to look at major disruptions and do ad hoc analysis to better understand the impact and effect of events when they occur. QRF is, by its definition, quick and helps manage the response to evitable but unforeseen changes. And QRF most importantly can help companies quantify what money they may be leaving on the table, and place a value on risk by understanding the changes as they occur. As Larry mentions, while these may be part of a quick response team or done ad hoc to handle operational disruptions, QRF may be a daily process more than a once-in-a-disaster at some point in the future. There already has been a lot of work to try to remove latency in the demand planning process. We need to continue to automate what may be automated, and reduce our internal process lead times. Secondly our roles migrate from advanced analytics to advanced decision making. We need to use our process to provide more data than information. Quick response forecasting may be able to do both, not just providing a forecast number but a response at the speed of business. 

Re-posted with permission from the author and the IBF - originally appearing in the 2018 Journal of Business Forecasting  | Winter 2017-2018 | www.ibf.org

Additional Resources

About Dr. Larry Lapide, Ph.D.

Dr. Larry Lapide is currently Research Affiliate at the MIT Center for Transportation & Logistics, as well as a Lecturer at the University of Massachusetts: Boston. Dr. Lapide has over 30 years of experience in industry, consulting, business research, and academia. Recently he was the Director of Demand Management at the MIT Center for Transportation & Logistics (CTL). He also managed the launch of MIT's Supply Chain 2020 Project and is responsible for CTL's Strategy Alignment Workshop. He concurrently served as the Research Director for the Demand Management Solutions Group, a consortium of companies that sponsored a multi-year research project focused developing advanced strategies, principles, and methods to optimally match supply and demand. Dr. Lapide is the recipient of the inaugural Lifetime Achievement in Business Forecasting & Planning Award.

About John Galt Solutions

Since its founding in 1996, John Galt Solutions has built a proven track record of providing affordable, automated forecasting and inventory management services for consumer-driven supply chains. We have an unmatched ability to configure tailored solutions for customers, regardless of size or business challenge, that save both time and money by compressing implementation periods and delivering intelligent information that positively impact your bottom line.




Tags:     PR and News

Subscribe