The internet of things has gone from a much-hyped new area within technology several years ago to a far more routine, mature sector.
For evidence of that, look no further than startups like OpenSensors. It was launched in 2013 by developers Yodit Stanton and Malcom Sparks. Since then it has gone from a two-person operation to a company with 10 staff handling 15 million sensor messages every single day.
Where OpenSensors differs to other IoT companies, which may specialise in consultancy or hardware, is its focus on open data.
“We’ve seen both the IoT and open data market mature while we mature,” says Stanton. “It’s definitely stuck.”
The idea for the company first spawned from Stanton’s care for her oldest child who has asthma, a condition that can be badly aggravated by poor air quality.
At the time, Stanton says, there was plenty of macro-level data about air quality but local information wasn't available.
The company set up an air monitoring project called ‘Breathe Heathrow’, which promised to ‘democratise data’ by getting local residents to create sensors around the airport to collect data on air quality and noise impacts. It’s a serious issue: air pollution kills 9,500 Londoners every year, according to King’s College London.
The scheme provided useful lessons in how to foster bottom-up, open data-driven projects within local authorities and communities and a repeatable data-driven methodology for such initiatives.
“The things you create out of your own frustration can often seem to work the best,” says Stanton.
“Now there’s worldwide air quality data from various publishers in Europe and North America, and that’s starting to turn into content,” she explains. “Some local newspapers are even putting that data onto their sites.”
The cofounders initially bootstrapped for the first few months, and have managed to grow reasonably sustainably so far, according to Stanton.
“We took some angel funding, but very little,” she says. “We were in a lucky position of having enough revenue to pay salaries. So that kind of organic growth is working well so far. It’s early days still obviously so we’re still trying to figure out what we’re doing, what people want.”
Stanton has mixed feelings about whether the company should take venture capital cash; or at least too complex to be reduced to a pithy quote. There are advantages and disadvantages in engaging with VCs and they vary enormously by company or even individual.
"In the early days, everything you do is unproven. In my opinion the focus for the early days of a startup should be about scaling the learning curve and figuring out what your customers want as quickly as possible before you run out of cash," she says.
OpenSensors now has over 10,000 users, with each of its customers managing from 2,000 to 20,000 sensors every year. They include big names like Cisco, TripAdvisor, Reuters and Arup.
Most of OpenSensors' clients tend to work within the commercial building, engineering, infrastructure and manufacturing sector. Any sectors where there are “lots of moving parts” and goods to be tracked tend to be suitable for OpenSensors, according to Stanton.
“We help companies manage their sensor network, but at the same time you can have a load of open data overlaying your private organisation’s data, which gives you much better analytics,” Stanton explains.
For example, an art gallery could overlay visitor footfall data, picked up by sensors, with open data on the weather and traffic on given days, allowing them to predict what future visitor numbers will be like.
However, the most valuable deployments tend to be slightly more routine and focused on efficiency gains, according to Stanton: “IoT tends to be most useful if you have physical infrastructure costing you money, energy firms for example. Deploying sensors is like having 1 million extra people working there: you have eyes on stuff and if it breaks you know about it immediately.”
One benefit is that this helps to save money by avoiding pointlessly replacing systems or machinery, Stanton adds.
OpenSensors has not seen much traction within the local authorities, perhaps suggesting the ‘smart city’ is still more of a buzzword than reality.
“I don’t think many councils can afford smart city deployments,” she says. “They don’t have any money and what’s being sold are usually multi-million pound massive five-year enterprise deals. Plus, for the most part it’s very hard to navigate the procurement process.”
Instead, Stanton believes the ‘smart cities’ ideal will come about from community-level deployments – people “doing things they care about” – rather than top-down schemes.
There have been some IoT projects within local government but they tend to focus on areas like parking, which help to raise revenue, according to Stanton.
The current team at OpenSensors has been drawn from a variety of backgrounds, but many of them have worked in banking in some way or another. Stanton herself worked in Lehman Brothers’ tech department before the investment bank went bust.
And she admits that for startups in competitive industries, it's not often enough just to focus on products: “We are very engineering-led. We just focused on products initially but in the last year or so we’ve had to do more sales. There's nothing wrong with that but we aren’t very used to it. If you’re crap at sales, no one just says ‘never mind, you’re a startup, have some money anyway’. You just fail.”
OpenSensors graduated from the Open Data Institute’s startup incubator last year and is now loosely based around a working space within Google Campus in Shoreditch, east London.
Rather than renting a full-time office or headquarters, the business has a distributed working structure where people work wherever is most convenient.
The real tie that binds the team together is that they all have experience working with a programming language called Clojure.
“It’s a niche language we all happened to learn that we love,” Stanton says. “We use Clojure for OpenSensors’ systems too.”
It’s particularly good for dealing with big data as it allows programming on multicore processors, according to Stanton.
One of the biggest milestones so far was when the company hit 10 million daily messages, she says.
“We realised this isn’t just us and a few of our mates. This is real. Then large content providers started asking if they can use our data. That was cool. It’s when it transitions from a project between you and people you know to an international thing. It’s like okay, this is serious,” Stanton adds.
Running a startup
It is clear that running a startup is a far cry from the more stable environment of working in a big corporate (albeit one in banking). Stanton seems to be thriving from the pressure – but she admits it is tough.
“It is very difficult running a startup versus running a job,” she admits. “You have so many different things to do and you have to be fairly good at all of them. You also constantly worry about whether you are making the right choices.”
On the plus side, Stanton says cofounding a startup has given her so many skills she wouldn’t have had otherwise.
“I’ve moved up about 10 years of experience,” she says. “In a corporate environment it would take ages to get to this level of responsibility. It is hugely stressful, however, and you can’t switch off!”
Stanton splits her time half and half between the customers and product development, and then does admin on weekends. Most of her time is spent ‘firefighting’: fixing problems then handing them onto staff members once solved.
“I guess I’m probably completely unemployable now,” Stanton laughs. She doesn’t sound remotely unhappy about it.