Data and analytics key to ensuring water quality

Safe water depends on analytics.

Without strong water quality data, the drinking water people consume every day, the water used to fuel the agriculture industry and the water industrial organizations use in their everyday operations could all be unsafe.

To ensure the safety of all that water, real-time data is necessary.

Water quality needs to be constantly monitored for changes, and data collected in the field must be transmitted for cleansing and preparation before the data is visualized so users can make decisions based on the data.

But real-time analysis isn't enough with water analytics.

The ability to predict problems before they happen so they can be prevented is also critical.

Six years ago, Meena Sankaran saw a lack of quality water analytics. She saw utility companies in the power and transportation sectors taking advantage of new technologies to increase efficiency, maintain safety and prevent problems, but the water sector was lacking.

In response, in 2015 she founded Ketos, a water intelligence vendor based in Milpitas, Calif.

Sankaran recently took time to discuss all aspects of water analytics, including why analytics is so critical to the water industry, what drew her to water analytics when she decided to start her own company, and what technology and services Ketos provides its customers.

Meena Sankaran

What is water analytics?

Meena Sankaran: Digital water analytics, by definition, is having a digital footprint of your water data that's accessible online for providing actionable insights to proactively run your operations and make smarter decisions.

The analysis of what your water data tells you can be instrumental to building sustainable operations if that information is derived with a focus around early warning.

What is the problem Ketos' technology aims to solve?

Sankaran: When you think about water in terms of technology, it's trailing [other utilities] like power and transportation -- you've heard the term the last frontier when it comes to technology adoption, and in the water sector, it even goes as far as still having to digitize assets like PDFs and spreadsheets. There are layers and layers of a lack of [technology], like a lack of automation. There is fundamentally a huge gap in terms of how quickly the adoption cycle is happening. In the last five to seven years, however, it's finally started moving, and the trend of digital water and smart water is finally coming in.

Fundamentally, [water quality] is one of the forgotten problems, but when it becomes a problem, it does so in a very mean or bad manner, like in Flint, Michigan. The reason is that the technology is very complex when it comes to measuring water quality versus the quantitative measurement of water. Understanding leaks and understanding usage and understanding pressure are easier to do, which is why almost 70% of the market has gravitated toward solving that problem. Less than 30% of the market is looking at water quality, and there's so much to solve in water quality. Ketos is fundamentally diving into that challenge of how to blend automation and data science, while pulling digitization into the segment to provide real-time water intelligence.

Fundamentally, [water quality] is one of the forgotten problems, but when it becomes a problem, it does so in a very mean or bad manner, like in Flint.

Are water quality and water quantity mutually exclusive, or is quantitative data also needed for understanding water quality and using predictive analytics to stave off potential problems?

Sankaran: If you want to truly understand the water available to you, you have to understand both water quality and water availability from a quantifiable standpoint. In the U.S., there are 14 trillion gallons of water that are lost annually, and a good portion of that is treated water, which is really unfortunate. Looking at both water quality and quantity through the lens of complete automation, however, allows you to layer in augmented intelligence and machine learning so the data can be taken to the next tier of predictions, and you actually forecast where there might be a problem to prevent water loss. We started trying to improve water quality, but the problem we want to solve got bigger because you have to look at water holistically in terms of quality and quantity.

What motivated you to get into water analytics?

Sankaran: I'm a firm believer in being self-aware and knowing where your competency lies in solving a problem. You might have a hundred ideas, but which of those ideas can you actually solve and make an impact? I went through a rigorous process of research and process of elimination and mapped it to what problems there were in the market. I'm an electrical engineer, so I understand chip fabrication and sensor design. I'm also an electronics communications engineer, so how to build around IoT edge devices came very naturally to me.

But the potential for starting something that was impactful was also very critical. If I was going to do something in the world of IoT, it had to make a meaningful impact. I had three choices -- either air quality, water quality or food safety -- and water quality is something that affects the other two, so that's why I gravitated toward water.

I also looked at where I could potentially prevent a disease outbreak, and whether we had a substantial amount of water quality intelligence to do that, and there was a gaping hole. There is no central database that is democratized where we can go and just understand why public health or safety is affected by the water we're drinking, and in order to build that, you have to start by gathering the data before you can make something of it. That all weaved together for me to me apply what I know to a sector.

Who is the target audience for Ketos' water analytics platform?

Sankaran: We focus on business-to-business. About 60-70% of water is used for agriculture, about 30% is for industrial use and about 10% is drinking water. Because we address all types of water except sewage -- irrigated water, drinking water, treated water for water recycle, frack water that's coming out the rig -- our customer base is a mix of agricultural, industrial and municipal organizations.

How is your platform different from a general-audience business intelligence platform and designed specifically for water analytics?

Sankaran: The way we have designed it, we are collecting all of the data from our edge devices. We don't process the data in the field, so if anyone ever breaks into the devices, no one can steal any of the data. And then all of the data that gets transmitted from the edge device to the platform is completely encrypted. Then in the platform, you're able to have a fully processed data view that shows you, by the second, how many parts per billion of arsenic, for example, there are in your water, or how much lead is in your water, or that, over time, manganese is decreasing and nitrates are picking up so you need to manage your chemical feed. Users are then able to interact with the platform -- it's a very bidirectional platform -- and set some thresholds to deliver alerts. Think of it as a live network operating system that helps a manufacturing plant be proactive, sustainable and safe.

What is the setup of the water analytics platform?

Sankaran: In the enterprise world, things are about a decade ahead of the water sector. In the water sector, typically, people purchase a piece of hardware and the water operator is then responsible for cleaning it, calibrating it every other day and assuming all the data precision liability. Then they're finding a person in the organization who is an analyst who takes the data and processes it and figures out whether [water is safe]. There are multiple labor-intensive steps in that process, and then the data is stored in some sort of a PDF. If they send samples to a lab, the lab sends results in a spreadsheet or PDF in a different data format. You also might have an old [testing] system that's connecting all your pumps that's in yet another different system. So, as a water operator, you might have [lots of] different data formats.

My philosophy is to let the water operator be the expert in water and not make them the technologist. Our system is designed so that we go in, install the piece of hardware and maintain and service it, so the liability is on us. We set up the communication for it, we have our end-to-end cloud platform and we have automated the entire solution. We're able to give them an entire integrated stack as a service.

The problem in the U.S. is that the water is clear -- it's not necessarily clean, but it's clear
-- so people just have a tendency to think it's safe.

What is your customer makeup at this point?

Sankaran: We're about 60% industrial, 25% agricultural and about 15% municipalities.

How has Ketos' water analytics platform grown since the company was founded?

Sankaran: My philosophy has been to apply the philosophy of software vendors, so upgrades are free and we make new parameters available the way software upgrades make new features available. Today, we have 26 parameters. When we started in 2017, we had just three. By the end of the year, we'll add four more, so we'll have 30. What that means to our existing customers is that if they like any of the new parameters, they just become part of their service upgrade. Someone could say, 'Hey, we saw that you're adding fluoride in the summer and we're curious to try out fluoride for six months -- can you add it in your next service visit?'

Beyond adding parameters, what have you added or enhanced since 2017?

Sankaran: One thing is heavy role-based access control. Cybersecurity is something the water industry is very mindful of. We have been pretty picky about security, and we layer it at every tier of our solution on an end-to-end basis.

We're in a constant feedback loop with customers, so we're working very closely with them.

Does Ketos work strictly with customers in the U.S. or do you have international users as well?

Sankaran: In the early days, for our quantitative product, we deployed in India, Mexico and the U.S. We've deployed our quality-based product in the U.S. and Canada, and we just recently made our first shipments to Israel and Brazil. We are looking to close on installations in Peru, Chile, Mexico, the U.K. and Australia later this year. There's a massive demand internationally because the water is very different, and certain countries invest [in water quality] a lot more than the U.S.

Regarding the U.S., the perception is that something like what's still happening in Flint is the exception, but in reality, how big a problem is water quality in the U.S.?

Sankaran: It's massive. The difference is that when you're in a developing country or developing region, the color of the water looks bad so you look at it and decide not to drink it. The problem in the U.S. is that the water is clear -- it's not necessarily clean, but it's clear -- so people just have a tendency to think it's safe. Some cities are really trying hard [to improve water quality], but water quality is a massive problem and there are probably 20 other cities that are on the verge of becoming another Flint.

Which cities have done a good job regulating water quality, and which ones have not?

Sankaran: I would say New York has really good water, and they've taken a lot of steps toward getting there. And in some parts of the country, drought is negatively affecting water quality. But you have to look at water quality from several angles, like what mining sites might be nearby that have been decommissioned. And then, what about all the agricultural runoff, the fertilizer runoff? The question isn't so much the water quality itself, but how aware the local population is of the water quality.

I think the only way to raise awareness is more and more people from technology helping out, and that's beginning to happen. Oracle now has a massive utility sector. Amazon has opened up something related to water and IBM acquired a water analytics company. Microsoft has a new initiative called AI for Earth and Cisco joined the SWAN [Smart Water Networks] Forum. I think the tech companies are realizing that there's a massive opportunity from a digitization perspective and an opportunity to help. It's just a matter of time, but it's not happening fast enough.

Editor's note: This interview has been edited for clarity and conciseness.