Are you a production planner drowning in a sea of sticky notes and spreadsheets? We understand how difficult it is to make hard, fast decisions within the chaos of ERP system alerts, inventory notifications, and a constant stream of last-minute change orders via texts, emails, and voicemails. Manufacturing companies, commodity manufacturing companies in particular, universally struggle with planning daily and weekly production. Data science and AI software can help you.
What are some challenges for production and demand planning in commodities manufacturing?
Commodity manufacturing companies like dairy, meat, fruit, grain, and vegetables are what we consider “deconstruction industries.” They use a non-uniform supply where each unit (for example, a unit could be a gallon of milk) has different characteristics than every other unit around it, creating supply chain characteristic uncertainty.
Within that supply chain there are a myriad of products that can be produced from that unit with varying levels of profitability based on supply. Additionally, the customer demand must be taken into account – it does you no good to turn your milk into cheddar cheese if there is no demand for incremental cheddar, but there is demand for Parmesan cheese, or dried powder and cream, or liquid milk. Production constraints and the concept of incremental margin (breakeven values) are also factors to take into account, as are external factors that impact your supply chain, like weather (or, lately, a global pandemic).
Demand variability fluctuates both weekly and seasonally, as well. This makes planning into the future challenging without the help of better tools like data science but so many planners are still stuck with analog solutions like sticky notes and clunky, complex, expensive and time consuming, spreadsheets. For example, if your company is looking 8 weeks into the future you must consider both supply and demand variability. Think about that for a second. Ask yourself if your team has any clue what to sell in the near future. My guess is that they don’t have confidence in the forecast of remaining product to sell at all.
How can data science help you manage these levels of supply chain complexity?
To get to a point where your team is ready to move beyond the spreadsheets, sticky notes, and magic eight balls you need an integrated system configured to bring all the supply and demand elements together to support their best work and – more importantly – make their work easier. This system would:
- Forecast supply and demand for your team using advanced data science
- Offer the flexibility to accommodate unexpected changes in the supply and demand forecast (such as when a retail customer buys a large front-page ad quantity)
- Break down a basic supply unit, like a whole hog, into component parts – without any work from your team
- Create near real time demand planning adjustments as orders flow in, eliminating guesswork and saving time
- Connect the demand signal through to the production plan based on business rules, but also be “smart” enough to know when not to change the production flow as well
- Constantly calculate what your sales team has left to sell at any point in the future based on the supply forecast, demand forecast, actual orders, external variability, and internal demand
If a system does all of these things you can shift the majority of the burden of fundamental demand and supply planning to it. This will save your team enormous amounts of time and reduce errors significantly. During my years scheduling and selling meat I found that most of the major issues around forward planning were caused by human error – our system removes that challenge.
What do we mean by human error? One example would be a salesperson selling something and forgetting to tell the scheduler for a day or two. They then put the order in and now you are shorting orders due to that human error. Or the inverse, the sale is reported to the scheduling team and they lose the information for whatever reason, causing you to once again short orders. These are two examples from a long list of possible errors.
Working with our software, if your salesperson takes their sale and inputs it, then that input immediately adjusts the demand plan (instead of sending an email that gets buried in hundreds of other emails). That automatic demand plan change then cascades through the entire scheduling and product availability systems. This has eliminated at least five potential failure points caused by human error, just from one sales data input.
Improving the supply and demand planning process so it is integrated with production planning and product availability has shown our clients improved returns (in several cases up to 4% in the first two months of deploying and using our software). What kind of profit increase would your company see if they could reduce or eliminate shortages and stop “fire sales?” We’d love to help you find out.