Fundamentally, there is a limit to how much efficiency can be gained from technology. An operator begins to see diminishing returns with each new technology simply because some costs can’t be eliminated.
More specifically, think about what percentage of an operator’s costs are actually “operating” costs. Usually, product cost accounts for somewhere around half, then add customer commissions and corporate overhead. However a company may perform, typically, it is fair to say somewhere around 30 to 40 percent of total cost is operating cost. So if an operator eliminated all operating cost (obviously not possible), the most that could be saved is 40 percent of total cost. Usually, technology can help reduce operating costs by about 30 percent or so — but that means the operator is really saving just 12 percent of total cost (30 percent of the 40 percent operating expense).
In effect, all the effort of installing a VMS, DEX, telemetry, dynamic scheduling, pre-kitting, and every other operating technology will likely save the operator around 12 percent of total cost. The industry has spent 20 years focusing on this. Don’t get me wrong, I do believe this 12 percent is valuable and certainly worth the investment. My point is simply any further returns will be capped.
On the other hand, using technology to improve sales has no natural limit.
Higher sales versus lower costs
There is another important thing to consider. Oftentimes, the two goals of maximizing sales and minimizing costs work against each other.
Again, taking it to the extreme for illustrative purposes, think about how often you would schedule a machine to be serviced if your only goal was to minimize servicing cost.
The answer is you would only service that machine when it is completely sold out of everything. When there are no items left in the machine, your revenue collected relative to the cost of that servicing is maximized.
However, as any operator knows, you cannot do that. You will be losing sales if the machine is not serviced until completely sold out (not to mention the customer service nightmare). But it is important to note: you have no idea what you did not sell. All you know is what you collected and what it cost you to service that machine.
This subtle, but powerful point is that the opportunity cost is not ever known to the operator.
Intuitively, the operator knows a completely empty machine has resulted in lost sales. But the magnitude is not known. Now let’s go to a more likely scenario with what an operator may strive for with dynamic scheduling — a system in which machines are scheduled to be serviced not based on a regular schedule but rather based on forecasting when the machine will need to be filled (using historical information), or based on real-time telemetry data when a certain level of inventory has sold out.
What is the ‘opportunity cost’?
When the vending machine is about one third sold out (about 15 out of selections), what is the opportunity cost of that? The operator may know he is collecting $250 per servicing and $20,000 per week per route, and that sounds great because the average collection per route has increased significantly as compared to a fixed schedule.
However, what is not known is what sales have been lost — the people who did not buy anything. My thesis is that by the time the machine has 15 out of stocks, the cost of lost sales starts exceeding the operational savings. Especially since the first items to sell out represent a disproportionate share of the consumers’ demand.
While the opportunity cost is inevitable, the heart of the issue is that the opportunity cost must be balanced with operating cost. This is a difficult balancing act because the true opportunity cost is never known. But through experimenting and refining the optimal point at which the machine should be serviced, profitability can be maximized.
It’s important to note the opportunity cost is not the same across all machines. A key factor to consider is the extent to which demand for an item can be substituted for another item. In some office accounts, if all the customer wants is Diet Coke, it is unlikely they will buy another drink. A lost sale will result. But at a school, if one item is sold out, it is likely the students will take an alternate item.