Revenue and inventory have a love-hate relationship. ‘You can’t sell what you don’t have’ is the reality of physical product companies.
So naturally Boards, Executives and sales teams push to maximize stocking levels. This approach works great during boom times. But as the economy slows and interest rates rise, companies feel the painful cost of excess inventory. All of a sudden, operations and finance teams find themselves under tremendous pressure to resolve the risk and expense of too much inventory.
Excess inventory results from mismatches in supply and demand at the unit level. Every company that sells physical products must grapple with the challenge of planning uncertain demand and supply.
Typically, companies initially take the following approaches to deal with planning unpredictable products:
- Plan in Excel
- Why: Well of course - it’s the ‘lingua franca’ of analysis. Where everyone goes first to check out new data. Excel is awesome - but most companies quickly find that it’s not enough.
- Consequences: Excel has inherent limitations - formulas and models are prone to breaking (#REF!), collaboration is hard (23-May_Plan_v09.xlsx), compute power is limited so models take a looong time to run and have to work at aggregated or summary levels, and there is no notion of a ‘test’ vs. ‘production’ environment to do scenario planning.
- Aggregate data to plan at a ‘manageable’ level
- Why: Humans can’t process huge amounts of data
- Consequences: “Stock outs? We have 20 weeks’ supply for that category overall!”
- Use 80/20 to prioritize largest SKUs
- Why: There is only so much time in the day and so many SKUs to get through
- Consequences: 80% of SKUs are a mess, assortment breadth limited by team’s ability to manage it.
- Rerun models on a cadence (E.g., weekly or monthly)
- Why: It takes significant manual effort to collect data and run them through Excel models, and then more time to verify the results and inform members of the organization
- Consequences: The plan is only ‘right’ once a week, slow response to changes, most of the time no one knows what the plan is.
Fed up with the Excel based approach, we’ve seen most companies next try to solve this problem by investing in demand forecasting and ‘connected planning’ approaches. Both are helpful, but can only get you so far.
- Demand Forecasting. There is inherent uncertainty and variability in demand. At one company we worked with, the ‘perfect’ look-back demand model could only achieve 40% weekly accuracy. Yet we helped that company cut its finished goods inventory to ~1/3 of industry standard. As a result, working capital management became a strategic advantage for them, financing their growth.
- Connected Planning. It’s way better than fragmented planning. There are several very large software companies that offer these systems, which basically pick up the logic of the existing Excel models and put them into databases with an accessible UI. These systems make the team more efficient and effective day-to-day. But they are very expensive. The systems are implemented via time consuming and expensive consulting engagements. And they are not resilient to changes in business needs. When the business needs change too much, teams start re-building Excel workarounds until another expensive consulting engagement can get approved to rework the models.
In summary, we’ve found that although getting a good forecast and connected planning are very helpful, companies can quickly get better results by investing in a real-time, learning system to continuously match supply to demand.