What’s the cost of complexity?

I was catching up on my weekly Autocar last week – which I have been reading since I was a young teenager when my father used to bring it back from his office – and picked up on an article related to the latest Vauxhall/Opel Astra.  The driving impressions were quite positive, but the bit that I noted related to the product range.  There are apparently 84% fewer derivatives on the new Astra compared to the old model, although that seems to exclude paint colour.  The choices are limited to five power plants, two transmissions, three trims and six options, so 180 build combinations excluding colour if there are no restrictions on how you combine your choices, still just over 1,000 with the seven paint colours included.  As the Autocar article stated ‘two minutes on the configurator and you’ll be done.’

In contrast, in the same issue, there was a road test of the Audi SQ5, where one of the many options is Matrix LED adaptive headlights which are bundled with OLED tail-lights.  That seems to be a sensible approach, but there are actually a further three options on the tail-light illumination pattern which apparently need to be specified at the time of order.  I hope that these are actually software-configured rather than a physically different light unit, but even so, it is yet another choice that the buyer has to make, and where they will want to go and check what the different patterns are, and which one they think will suit their tastes.  To understand and select all the various combinations through the Audi configurator would more likely take two hours rather than two minutes, and the buyer would almost certainly still have doubts and questions that they would need to resolve before finalising their specification.

This is perhaps one reason why in the UK today I could carry on from the configurator on the Vauxhall Online Store, and place an order for home delivery with a 14 day money back guarantee (and for some reason a £500 discount code despite a 21 week lead time).  I am sure that the number of takers remains extremely low as it is for almost all online car channels, but it does look like a feasible option, whereas Audi obviously recognise that they are a long way from a customer being able to make an informed choice online, so steer you to a dealership, armed with the configurator code that you generate online.  The simplicity of offer has been a core feature of Tesla from the beginning, and was the unseen differentiator ten years ago when everyone was focused on the powertrain.  Tesla drives a higher level of online sales than any other OEM through the combination of a simple offer and a much leaner physical network.

However, the benefits of low complexity are not limited to the point of sale.  Many years ago when I was with consulting firm A.T. Kearney, we had a service offering which we called ‘deproliferation’ and a number of OEM clients commissioned studies of this type.  We were focused on the purchasing, manufacturing and supply chain impact of complexity, and in particular on low volume derivatives and options.  Low volume tends to be accompanied by high volatility in demand, resulting in batch production of any unique parts and higher safety stocks if lead times are to be kept at a reasonable level.  All the effort involved in creating and supplying the low volume variants has to be amortised over the smaller volume, and the effort itself may also be higher than for what we would call a ‘runner’ specification – one that is stable and predictable and is therefore low maintenance.  An activity-based costing approach shows that where low volume specifications have been costed applying standard overhead rates, this may be a fraction of the true cost of offering that option.  Even with a premium price attached to the option, it may well be loss-making in real terms.

This is the reason why Tesla chose to build functionality into their cars which is then software enabled – as supply chain complexity is removed, whilst still giving the customer some choices.  There are clearly limits to how much redundant hardware can be included in a vehicle where only a percentage of the customers will pay an additional sum at point of sale or through an in-car or app-based purchase at a later date to activate that functionality, but it is a better approach in supply chain terms than building the complexity in right from the start of the supply chain.  I’m reminded of the impact that my late friend and colleague Martin Leach and the late Richard Parry-Jones had on the driving dynamics of Fords in the 1990s and 2000s.  The customer was paying for a car that could provide a great driving experience on the right road, but as they crawled through urban traffic they were totally unaware of the redundant excellence that had been included in the car.  Are software-enabled features so different?  As I mentioned in my blog a few weeks ago, some customer education is required, but the economic argument is potentially sound.

If we are going to get cost out of distribution and give the customer a realistic option of buying online, then we need to follow the Astra or Tesla route.  We need less complexity, with more flexibility around how and when a customer can create a car that fully reflects their needs and tastes.  Where a brand chooses not to do this, then they need to fully understand the cost of complexity and configure their supply chain and sales processes and networks in a way that truly generates an incremental profit.