Analyzing the Internet of Things: Mnubo, data science and connected devices

Montreal startup Mnubo (the “m” is silent) has its fingers in a lot of pies – from industrial irrigation optimization, to smart thermostats, to a connected fleet of beer fridges that monitors its inventory and sends a message to owners when they’re understocked before a Habs game.

But back in 2012, when Adi Pendyala and his three co-founders started Mnubo, it was hard to get companies manufacturing smart or connected devices to see the value in paying someone to analyze their data.

“We must have been crazy,” Pendyala told MTLinTech at Mnubo’s office in Pointe-St-Charles’ Nordelec building. “The industry was still early.”

But they knew that if you were going to launch a mobile app, you’d need analytics.

Now, Pendyala says the Montreal startup ecosystem is big into analytics, likening the boom to the recent surge in VR and AI.

As more devices became connected – and Mnubo began demonstrating that IoT data analysis could make a company more profitable – the contracts came. And so did the Series A funding lead by White Star Capital.

According to White Star Capital’s Managing Partner Jean-François Marcoux, what Mnubo does is like Mixpanel for connected product manufacturers.

“Rather than back one gadget-maker, we wanted to back the engine that could power them all,” he says.

At first, Mnubo had more success reaching out to smaller, B2C companies. Decisions could be made faster, there was less data to crunch, and results were obvious. One of their first customers was a smart watch company in Hong Kong who wanted to know how their customers used the watch, when they took them off, when they registered them (usually only when they had a problem), when they put them in a drawer and when they were running out of battery. By analyzing the data, they could improve their customers’ experience with the watch. Happier customers meant more watch sales.

But as Mnubo grew, they were able to take on (larger, though less sexy) industrial and agricultural companies as clients, like when they started analyzing sensor data for automating irrigation systems on California avocado farms and improving maintenance schedule for HVAC systems in large office buildings.

Heating, ventilation and air conditioning systems traditionally use a calendar maintenance system, Pendyala explains.

“The first of every month they’ll go down, check, check, check, fill out a card and leave. What analytics allows you to prioritize the servicing of an asset based on its score and performance. You’re able to give them a dashboard showing the 100 assets in the field, the ones that have exhibited a low performance behaviour and how the maintenance should be applied. So they can prioritize their servicing. The cost of a person is optimized where it’s needed most. And customer experience is improved because you’re preventing machines from breaking down.”

Now Mnubo has customers on four continents and its approximately 45 employees aren’t sitting on their laurels (or taking too much advantage of endless beers in their self-restocking fridge). The only challenge in working with new customers, says Pendyala, is when a customer doesn’t know what problem they need to solve.

“In a typical engagement there are two scenarios,” says Pendyala. “One, a customer has a problem, they have an algorithm and they want to bring that algorithm to a platform. The other scenario is when you have data, you have no algorithm and you’re relying on data scientists to create that algorithm. That we’ve also done. The area we fail is when you’re like, ‘Show me something amazing.’ It’s very hard for us to do that because we can’t solve problems that the customer isn’t aware of when we don’t understand the domain.”

Pendyala says in these situations, their customer success engineer (each project is assigned one) can work with the customer to figure out the question or questions they should be asking, but it’s a challenge. “We have a program called Data Discovery, but we ask the customer to commit, where we say this is what we found in the data, tell us which direction to go. Companies aren’t thinking data yet. Getting people not thinking about data to start thinking about data is our challenge,” he says.

In the meantime, Mnubo plans to keep growing by investing in R&D.

“The next five years is building a richer pool of algorithms, [insights] libraries and data science that keep us at the technical edge, but also to onboard more customers early in the IoT journey, to start deriving data insights earlier from the day you turn that product on,” says Pendyala.

The company’s competitive advantage comes from its experience, says Marcoux. “Mnubo solutions is unique as it provides out-of-the-box business value, faster time-to-insights, customizable IoT data science and a clear demonstrated ROI to product manufacturers. There aren’t many companies in the space that can claim such an impressive client list.”

They’re also one of the few IoT analytics companies that’s extremely specialized, says Pendyala. “When you google IoT platforms, you see a few hundred pop up. But when you narrow it down to who’s doing analytics around IoT data, there’s a subsection that’s very focused on it.”

As Pendyala takes a beer from the office mini fridge, a sensor sends a message to a digital chalkboard somewhere, adding one more tick to the Mnubo beer tally and turning novelty into added value. From a risky bet to a booming industry, from beer fridges to industrial agriculture, Mnubo’s success is proving that the future is smart – and Pendyala and his partners are four years ahead of their competition.

+ There are no comments

Add yours