Introducing Claire Rowland
Currently based in London, Claire is an independent UX and product strategy consultant, and one of the leading figures in designing for the Internet of Things. Her background as Director of Service Design at AlertMe.com, a connected home platform provider, and Head of Research at Fjord, gives her a unique insight into how computing and the consumer world collide. She is also lead author of Designing Connected Products: UX for the Consumer Internet of Things, recently published by O'Reilly.
Could you give us a bit of background into how you found yourself designing for connected products?
Like a lot of people in design and HCI, I’d always thought ubiquitous computing was interesting. Back in 2010 I was working for Fjord on cross-platform digital services, when we became involved in a European consortium called Smarcos researching the 'interusability of interconnected embedded devices and services'. That was entirely R&D work, but it was a great opportunity to get a theoretical grounding in the space. In 2012 I went to work for AlertMe, a connected home platform company in the UK (now part of British Gas). That was where I got to learn about the practical challenges of making real, connected consumer products.
Your book focuses on the consumer application of connected devices. In many of these markets, the technology is ready, but still are yet to crack the consumer market. What are the breakout use cases going to be?
That’s a bit like asking what the breakout cases for the internet are going to be! It’s becoming so cheap to put extra computing power and connectivity in devices that we’ll see more and more of the things around us gaining processing power and becoming connected in the next few years. Figuring out what to do with it, and doing that well, are big challenges.
The technology still has a long way to go in terms of interoperability between different products, although that’s improving. Reliability is also an issue: connectivity introduces new ways for things to break. Consumers don’t want things in their homes that only work 90% of the time, so we have to get better at making systems robust. A more complex challenge is figuring out how, once we have lots and lots of connected things, we can get them to coordinate in sensible ways. We are in the very early stages there.
To be honest, although my experience is in consumer systems we are seeing faster uptake in commercial and industrial systems because the business case and value is stronger. I think the consumer market for a long time to come will be about incremental improvements to existing products (heating, security, doorbells, energy monitoring, health monitoring) rather than a few big bangs. And that’s probably a good thing, as there are a ton of things we haven’t worked out yet, like what good privacy looks like and how users’ data should be handled.
When bringing these products to market, what are the biggest challenges facing companies?
The combination of hardware and software is complex. Software enables you to change the functioning of your system over time. There is more upfront cost in developing hardware and it’s harder to iterate in hardware. Hardware design and engineering decisions may need to take into account how the functionality of the system might change in future. So you need to be really sure you’re building something people want, and understand how it should be designed, before you get too far down the process of making the hardware. And you have a duty to users to maintain the internet service that supports your system: customers don’t want to think they’ll be left with a useless brick if you go out of business.
Making things that work as reliably as people expect can also be is an issue: we accept delays and failures on the internet as part of normal usage but they will feel strange to users as experience through the real world. Connected versions of mature consumer products, like lighting, can feel in some respects like a worse experience than their non-connected equivalent when they are sluggish to respond or fail to do what we ask them to.
It’s hard to engineer that out completely as latency and reliability are often issues with networking. And finding the balance of end user value and cost – in terms of effort to get the thing doing what you want, as well as money, is always a challenge. We’re still seeing lots of things that demand a lot from users in relation to the value they provide. Early adopters don’t mind that but cracking the mass consumer market is hard.
What are the biggest challenges facing designers moving from disciplines such as UX and interaction design into designing for IOT?
One of the key challenges is understanding some of the technology of IoT, in particular designing for distributed systems. Partly this is about the figuring out how to distribute functionality between devices – do you put this feature in the app or on the device or both – and then designing for a coherent experience across all the devices. It’s also about designing for the properties of networks.
When we’re designing for PCs and mobile devices we can generally assume that the devices are usually connected. Embedded devices – the specialized devices that make up many of the Things in the IoT – often connect only intermittently. Many of them run on batteries, and maintaining open network connections uses a lot of power. So they spend a lot of time offline and only check into the network intermittently. That means that parts of the system can be out of sync with each other, so in the case of a heating controller you might have one device telling you that the heating is set to 21C and the other telling you it’s set to 19C. And you have to consider the delays that may occur between one device sending a command and another one acting on it, and how you handle those in the UI.
A lot of IoT design is about handling delays and discontinuities gracefully. And then there’s the complexity. If you’re working on a system with lots of devices that interrelate in complicated ways, like a connected home system, keeping track of all the ways things can be grouped and automated gets really complex, really quickly. And of course, Things are situated in the real world, so the context of use can be complex to understand.