Get Servitized

There is something telling – symbolic, even – about ARM’s former CEO, Warren East, taking the helm at Rolls-Royce. Going from one industry to another of course isn’t exactly anything unusual in the corporate leadership circles, but still, it’s difficult not to generalise on it, just a little bit. When such an iconic manufacturing group selects an alumnus of a computing powerhouse as its new chief executive, it surely has the Internet of Things, Industrial Internet, or, say, Industry 4.0, written somewhere on it?

Besides, it gives me a nice excuse to blog about something I’ve wanted to blog for a while now: servitization. See, if you have attended any IoT Innovation / World / Future [pick one] Congress / Conference / Summit [pick one] lately you have probably come across the concept during one of the Keynotes / Presentations / Panels / Roundtables [pick one] – and if you have, then my low-risk bet is that the Thought Leader / Sought-After Moderator / Charlatan [pick one] you were exposed happened to use Rolls-Royce as the classic example of servitization transforming business models in the real world. “We’ve all heard how Rolls-Royce no longer sells its jet engines, but leases them to the customers – and that’s the power of IoT! Kumbaya!”

After nodding knowledgeably through it a few times (and using it at least once), I eventually got that itchy feeling you get when a fact is sounding a little too conveniently inspiring to be fully true. So I took a look, and as far as I figured out (which, be warned, is not always far at all) it indeed was not. Sure, Rolls-Royce does lease aircraft equipment but that doesn’t mean that it has ceased the traditional sales model altogether.

The said revision then made me curious about the actual benefits of servitization in general. Is there actually that much hard evidence that it is working as well as often advocated? I didn’t find an awful lot, but this paper from INSEAD on performance-based contracting (PBC) for after-sales maintenance, which in fact relied on data specifically from Rolls-Royce, indicated that it improves product reliability by 25-40% against the traditional, time-and-materials contracting. Importantly, that appears to be only after accounting for the “endogeneity” of customers’ contract choice. Without accounting for it, there doesn’t seem to have any real difference between the two contracting models. In other words, customers who are likelier to require a lot of maintenance tend to gravitate towards PBC.

That finding highlights a downside that manufacturers embracing servitization have to deal with. If customers can choose between paying for uptime and paying for solving downtime, it’s natural that a certain self-selection bias will kick in. Blame it on a moral hazard or an ordinary cost-benefit analysis, but that's an area where the end user's interests are not easily aligned with those of the supplier.

This is where the topic gets genuinely interesting also from the IoT perspective. See, as a concept servitization after all doesn’t really require that the products being turned into services are either connected or smart; that's just conference speak. Rolls-Royce, for one, has been offering its Power-by-the-Hour PBC model since the 1980s. What is becoming different, though, is that manufacturers can finally have a much broader and deeper visibility over the way their products behave in the field, owing to the advances in sensing technology. The data captured by sensors is also more accessible, thanks to the improved connectivity. And last but not least, the same data is becoming more actionable, because of modern analytics.

In short: sensors, connectivity, and analytics make it possible for manufacturers to reliably monitor and analyze products they may be providing under a service-based contract. Previously such customer relationships were like a black box – with very little visibility over product usage, which inevitably lowered the potential upside. Now that this limitation can be mitigated, even significantly, quite exciting things could start happening.

For more thoughts about the relationship between IoT and analytics, you can refer to our recent research findings – this or that – as well as a blog we contributed to IEEE’s IoT newsletter on the subject.