Lynn Cinelli's Wake-Up Call: Pharma's Forecasting Makes CDMOs Sick
By Louis Garguilo, Chief Editor, Outsourced Pharma
Lynn Cinelli, Vice President Global Biologics Supply Chain, Bristol Myers Squibb, has years of pharma supply-chain experience, and attributes some recent epiphanies to Lora Cecere, Founder of Supply Chain Insights.
Cinelli has been strategizing on how – within recent variances in macro-economics, politics, and markets and trade – drug sponsors can better harness direct end-customer demand through the supply chain.
As currently practiced, she believes, drug sponsors are poor at forecasting and a major component of that are attempts to forecast the “unforecastable.”
“Our systems and planning, in terms of S&OP cycles or annual budgeting, run counter to the speed at which information is changing through that supply chain,” she explains.
“The group that feels the biggest bull whip is the early-stage CDMO.”
For example, sponsors are incentivized to optimize on longer runs to get more competitive pricing.
“We keep squeezing the CDMO, and they're trying to absorb all that variability through the supply chain,” Cinelli says.
“In addition, there are other externalities like tariffs, taxes, climate change, you name it. People are always jumping to something, and unfortunately the flow of information of true customer demand to the CDMO is warped and compounded by competing functional incentives.”
Much of this jumping around turns out to be unnecessary, or with unforeseeable impacts.
“BIOSECURE Act to me is one of those examples,” she replies.
“Until there's a true inflection point, many of these things resolve themselves. Especially in our industry, if it will hurt the patient, it's going to resolve itself or delay in implementation.”
Sponsors and external partners alike need to be cognizant of their macro- and microenvironment, but jumping into (premature) activities is not strategic thinking.
“To me,” says Cinelli, “it’s important to start with first principles and be patient oriented – how do patients take medications in general, that isn't changing dramatically,” she says.
Again, there is the need to do some “scenario planning and even find ways to take advantage of disruptions (anti-fragility), but we must balance analysis paralysis.”
Back to forecasting and regardless of the time period, Cinelli says drug sponsors continue to introduce that “bull whip of variability” from a demand and supply perspective throughout the supply chain.
Each group has different functional incentives, and that causes demand signal to shift dramatically.
“However, if I take as the first principle in our industry that we deeply understand the true burden of disease and provide full opportunity for sick patients to take their medicine in a predictable way, the true demand may not be as variable as we try to make it out to be.
Unfortunately, Cinelli says, sponsors whipsaw a CDMO with flawed forecasting.
Big Pharma may contend demand is going down 50%, when actual patient demand might have declined nominally but sponsors factored in plus/minus 20% because of all the variables coming from within and without the supply chain.
That’s no way to run a supply chain.
Trying To Operate Better?
One way is to create an understanding and convey to CDMOs more of that primary-demand signal.
“Internally within BMS, this manifests in helping all the groups understand the impact of their functional optimization, how they're detracting from that signal fidelity throughout the supply chain,” she says.
“Do all these pressures we exert – finance pressing for inventory and working-capital reduction, procurement looking for more volume and tier pricing – add value?
“My supply-chain planners, in the meantime, are balancing highly variable country regulatory approvals, and market variability, especially at that dynamic end of the product lifecycle. You have LOEs [loss of exclusivity], new product introductions.
One area you can forecast better, is in “the middle segment,” where you can employ new AI and machine learning capabilities. “You can forecast a good portion of your material,” Cinelli says, “but always be intention with your partner about managing those items.”
“Regrettably, we treat all products and all stages as forecastable. Yet if you really go back and analyze, they aren’t.” This drives frustration in the planning organization.
For Cinelli, this is a foundational flaw in our practice of supply chain as a discipline.
And it a reinforcing cycle. Everyone feels the effects of mis-forecasting, and the various functions then try harder to reduce errors by forecasting more. “New monthly projections are inserted; readjustments pile on readjustments.”
“We unintentionally cause a lot of non-value-add work, and actually make it worse,” says Cinelli.
The Final Forecast
And so … what Cinelli is striving for, and supply-chain gurus like Cecere at Supply Chain Insights are teaching, is “flip to understanding what is actually forecastable – and measuring forecast value-add – and then partnering with CDMOs in a more transparent way.”
We must improve the entire drug-cycle supply-chain, more accurately – and flexibly – determine quantities and the timing of production campaigns and deliveries.
“Sponsors simply need to better focus on the independent demand signals of the actual patients,” Cinelli says.
Independent demand signals, as in those not warped by distributors or warehouses looking to make a quarter end more profitably, or one’s own procurement and finance department’s strategies.
“What are skewing the real demand indicators,” says Cinelli.
For example, logistic operators have input into all Cinelli’s manufacturing plants – whether internal or external. “Add in things like signaling demand based on other products in our production mix.”
All this skewing can materially impact external partners.
After many years in the industry, Cinelli is amazed that although “in pockets within the company we have access to direct information from patient use, it’s not elemental to forecasting demand.”
“Sometimes we use our own orders as measurement of the forecast,” she says with incredulity.
“But after years of struggling with this challenge, I have reached that ‘A-ha moment,’” she says in recovery.
“Where aren’t we focusing supply and demand more on actual script data? This unadulterated signal would benefit everyone involved, including CDMOs.”