AI and Product Cost – Better Insights?

It’s an uncomfortable truth, that many healthcare company manufacturing finance models fail to offer a “true” insight into the product cost of goods (COGs) at the SKU level across the entire portfolio. While, over the last few decades, most companies have done some great work to establish better activity-based costing models, our experience is that whilst these typically provide some sort of defendable position on “true” brand or line COGs, in many cases they fail to provide the accessible, comprehensive and accurate levels of understanding at the individual product level, required for high-quality, agile decision-making. While significant progress was made with the cost models available at the time, emerging AI and Machine Learning (ML) technology can now provide for a step change in increasing understanding and transparency in costing with limited effort.

Several years ago, I had the pleasure of talking to a company outside the healthcare industry as part of an external benchmarking exercise on how “best-in-class” companies – in this case, a manufacturer of printing equipment – manage portfolio complexity. Frankly I was staggered at the degree to which these “masters” had a complete view, and hence complete control, of their product portfolio. And what was the most powerful tool in their toolbox? In my view, their 100% understanding of manufacturing costs, which they could use to make truly informed decisions about portfolio and cost allocation. In one example we were shown three replacement parts for printing machine installations. To my amateur eye, these all looked essentially identical – same design, same size, same construction – but our hosts demonstrated that these three items had very different costs. Their hugely impressive cost models were extremely detailed, including not only highly granular breakdowns of the materials (down to the individual bolt level) and labor usage (in minute-level detail), but amazing consideration of things like “cost of carry” for the slow-moving versus fast-turning SKUs and the cost of return and restocking items ordered in error.

Back at work a couple of weeks after our visit, I was reflecting on the printing discussions whilst meeting with a group of senior colleagues to debate some remediation work to improve the robustness of a small, single-market product. The obvious question all of us were pondering was whether the additional work required by technical services plus the costs of the external laboratory testing being proposed, along with the collective salaries across the conference room had effectively nullified the SKU’s margin contribution for the next three years. For this particular product, a meeting of this type was not uncommon, (one generally happened every couple of years) and the question in the room was why weren’t we making good decisions with our commercial partners on the future of this particular SKU?

So it’s clear that in cases like these, the responsibility for imperfect product costings must lie squarely at the door of the Finance department! Of course not – that would be completely unfair. While the almost complete mastery of cost demonstrated in our printing company example is impressive, the work required to assemble these models have historically come at significant cost themselves. Our printing friends had employed a team to undertake this work and the cost analysis was so detailed, it took many months of effort to be done for each item in their comparably compact portfolio. At the time, their approach involved countless hours of detailed manual data collation from a variety of structured and unstructured sources across all functions of their supply organization, which were were then used to develop and test their sophisticated cost models. To scale this approach up to the full SKU count for, say a large international consumer healthcare company, and then factor in the more dynamic nature of a CPG environment, this mastery has historically been a challenge to achieve at the required scale and speed to return value.

But perhaps, there are now technologies emerging which can deliver healthcare companies a step-change in their ability to determine true COGs, perhaps even on a dynamic basis, factoring in real materials usage, real direct labor and accurately attributing indirect overhead allocations at the SKU level. Over the past few months, Impact Supply Chain Partners has been looking at a number of knowledge-based AI and ML systems offering the capability to start building and processing large pan-operational data sets from previously disparate sources such as enterprise systems, manufacturing execution systems, operations management systems, timecard reporting, right the way through to automating the inclusion of unstructured data from manual batch records or excel files. The capabilities of some of these tools could be truly transformational and from what we’ve seen, open a gateway to better COGs understanding and transparency.

So will these new AI tools herald a dawn where your supply team will actually allocate the real cost of each individual batch of product to your sales team? I guess while it could do so, it’s probably an unlikely change in the short-term. While such an approach obviously sounds very appealing to the manufacturing arm of your organization, it would undoubtedly be far too much of a shock to your commercial colleagues who crave the certainty of a standard cost in their P&L for the next 12 months, irrelevant of what happens over that period. However, what these new tools have the undoubted potential to offer, is a step-change in the quality and timeliness of the conversations about true cost and profit contribution of products in the portfolio. They will allow manufacturing organizations to become more effective roadmap in setting budgets, identifying cost savings, making faster decisions, and improve prioritization. Ultimately it is like they will offer insights for the transition to more profitable product portfolios, better resource planning and even the design of manufacturing organizations themselves.

We truly believe that the future deployment of AI tools in healthcare manufacturing which deliver a better understanding of COGs will have a range of significant impact from the purely diagnostic, right through to the transformational. If you would like to have a chat about how Impact Supply Chain Partners could help create value in your supply chain through some of these tools, please contact us at contact@impactsupplychainpartners.com or see our website at www.impactsupplychainpartners.com

Keywords: #SupplyChain, #ConsumerHealth #Costmodels #AI #ArtificialIntelligence #MachineLearning

© IMPACT SUPPLY CHAIN PARTNERS, SEPTEMBER 2024