Scale, complexity and the hidden costs of CPUs

It was Leonardo da Vinci that said: “Simplicity is the ultimate sophistication”, inviting the reader to understand that often we must introduce sophistication in order to deliver simplicity. In our sector I find this to be particularly true when it comes to how we organise our food production.

In 1976 I started my first position in our wonderful industry as a chef, in a London hotel. Deep in the basement were the food stores, with goods-in managers and burly porters. Here were quarters of beef at 50 kilos or more, whole halibuts of over 20 kilos, and huge pallets of unpeeled potatoes filling every inch of storage, before being sent to the larder (preparation kitchen) which groaned under the weight of work to meet that day’s service needs.

These places don’t generally exist anymore, rightly made redundant by more integrated supply chains, and significant levels of pre-preparation by suppliers where specialist skills can be more efficiently applied.

But there is one potential anomaly – the Central Production Kitchen (CPU). These have mostly sprung up over the past 10-15 years – small operator-owned manufacturing businesses sitting between the inbound supply chain and their restaurant kitchens.

Generally, they have emerged from a relatively simple hypothesis, which goes something like: ‘we do this task in 35 separate kitchens, surely it will reduce cost and improve consistency by doing it in one place?’ This is often followed by two assumptions: ‘we can use an existing asset’ and ‘only we have the skills to hit the standards required’.

Ten-years ago these assumptions were often true. But site growth, menu changes, and supply market capability often mean that in just a few years things can look very different. Put simply, food manufacturing, in itself, is relatively easy – but only if things remain the same. If there is anything the past few years have taught us, it’s that they won’t!

Here are just a few of the circumstances we have encountered within operators with CPUs during the last decade:

• Site growth generating a need to open a new CPU generating major disruption and high/unexpected decommissioning costs
• Declining volumes resulting in a ‘forced decision’ to produce new SKUs within a CPU just to maintain capacity utilisation, requiring unexpected cash investment
• Poor operational efficiency leading to CPU staff costs nearly 20% higher than optimum
• Insufficient scale meaning that production runs of many SKUs were so small that the set-up/close-down processes made up over 50% of the product cost to the restaurant, making the cost per item ridiculously high
• Seasonal fluctuations in volumes leading to capacity stress in high season, and inefficient capacity gaps in low season (you can replace the word season with month/week/day as well)
• New entrants to supply markets that can produce same/better product at lower cost without operator investment
• In-bound raw material costs c.15% higher than the same materials sourced by the manufacturer
• High availability risk with production of all key SKUs in one building (no disaster recovery)

There are many more…

As businesses grow, what started out as a simple piece of ‘centralisation’ can so easily become a hub of hidden complexity, costing the business huge amounts of effort and seriously damaging margin.

I call it hidden complexity because many operators treat the CPU as a central cost, and do not evaluate it against other strategic options. It’s genuinely rare to see a clearly stated cost per kilo for each outbound SKU from a CPU, let alone a formal evaluation of CPU as a strategy – in terms of both its level of efficiency, and also its comparability with supplier sourced product.

Indeed, I often see the opposite – operators defending a single inefficient CPU SKU citing a further loss of efficiency if it were to be outsourced to external supply. When your boat is sinking it’s generally a good idea to look to see if there are other boats nearby.

Which leads me to my central point. It’s dangerous to think of your CPU as just another part of your operation. In my view a CPU should be regarded as a food manufacturing business competing against external supply. It has hungry resource and cash needs, often imposes unnecessary complexity on already stretched businesses, and unless closely measured and managed can easily become uncompetitive.

I feel certain that there are successful CPUs in our sector, but to be frank there is conspicuously little measurement to help us determine where they are.

Da Vinci said that we need sophistication to delivery simplicity. It’s time to raise the bar on our management of these critical assets.