How to build integrated global forecasting teams
Miriam Stache, senior European demand forecasting manager at Eli Lilly, on how integrated teams can provide more reliable forecasts
Global forecasting units are rarely airtight operations. Most have local, regional, and global forecasting components, and each is responsible for different subjects and beholden to different stakeholders with varying expectations. Additionally, communication from one market to the next is often infrequent, and a lack of common standards, methods, and models makes it hard to evaluate comparative effectiveness from a global perspective.
Miriam Stache, senior European demand forecasting manager at Eli Lilly, says these gaps can lead to a “black box” feeling where individual forecasters are in the dark as to what other units are doing and how their work collectively fits into the larger puzzle. “Even though we are one dedicated forecasting unit, [individual] forecasters are torn apart between different expectations,” Stache said at eyeforpharma’s Pharma Forecasting Excellence conference in Berlin. “And then the question always comes, How do you prioritize?”
The “Last Best Estimate”
The gaps and inconsistencies among local, regional, and global forecasting functions proved problematic enough that Lilly decided to overhaul its approach to global forecasting. The goal of its new, integrated global forecasting team is to increase transparency and visibility, improve teamwork via shared learning and better tools, and ultimately provide quicker and more reliable forecasts for improved decision making.
Lilly tried to meet these goals by effecting change on a micro and macro scale. On the micro side, they assigned a local forecasting team to each affiliate. That forecasting team is then tied to a regional forecasting lead, who in turn is tied to a global forecasting lead. The local forecasting team is the owner of the “Last Best Estimate,” or LPE, which for each country gets uploaded into a central forecasting repository.
On the macro side, Lilly created the central forecasting repository as a shared place where the company could store all of its forecasting data with all of its different forecasting components. Inputs in this repository are assessed on a quarterly basis, and the company creates a consolidated report each quarter that contains all the key performance indicators that the company now tracks on a local, regional and global level. “They all have the same view, and this helps tremendously,” Stache said.
The consolidated report, built from short-term inputs, then serves as the foundation for a Long Range Plan, or LRP, which forecasts the long-term outlook, risk, and assumptions of products under the Lilly brand. The LRP is reviewed on a quarterly basis, which Stache believes is the right balance between monthly and annual reviews. “This is a very good compromise because you don’t wait too long to update your assumptions,” she said, “but on the other hand, it’s not an overload of work [where you’re] doing the same exercise every month, where nothing really happens.”
Making the model sustainable
Stache reports that after some initial hard work developing the new structure and central repository, the global forecasting overhaul was successfully implemented in 2011 and was actually “quite easy to achieve.” Much of that is because the new model is intuitive to forecasters, and therefore their tolerance for change is higher. “Now everybody is looking at this one central report and there’s not this overflow of different reports of different colors,” she said.
At the same time, in addition to a new model, forecasters have had to unify in new ways to harness the model’s power. For instance, structured teamwork has become a focus. If numerous units functioning in different markets and with different products are to work in lockstep, they need to open and maintain better lines of communication. Local, regional, and global units now schedule 90-minute Web team meetings every month. In addition, local teams schedule live, week-long team meetings four times a year, and regional and global teams assemble for live core team meetings twice a year.
Stache cautions that global and regional forecasters might not be aware of local workloads, so it’s important to communicate early and discuss feasible timelines and deadlines. She also says forecasting units across the country need to become more consistent in the way they hire external support. External resources at Lilly are now brought on for specific tasks, like development and creation of standard forecasting models, consolidation of forecasting outputs in the central forecasting repository, and the creation and validation of forecasting outputs.
While Lilly outsources standard work on forecasting infrastructure and reporting, its forecasters still perform all calculations in-house. These functions are solidified and consistent among all forecasting units across the company. “Now within Lilly we are one forecasting team,” Stache said. “We are spread across [numerous regions and countries], but we are one team.”
For more on forecasting, join the sector’s other leaders atForecasting USA in October in Boston.
For an overview of eyeforpharma’s forecasting coverage, seeHighlights from eyeforpharma’s Forecasting coverage.
Since you're here...
... and value our content, you should sign-up to our newsletter. Sign up here