Brilliant chat with Colin Gounden, CEO of VIA, a company that uses machine learning to enable encrypted analysis of confidential data at large energy companies. We discussed the comparative advantage of the VIA founding team, the way VIA solves for the data siloing issue which so many machine learning projects suffer from, how they solved for slow enterprise sales cycle with the “30 minute pilot”, how growing up in multiple countries affects your perspective, why energy efficiency of everyday appliances is a massive opportunity for cleantech founders, how they hire for their company value of “love in:love out” and more!
Here is the audio. The excerpts below are lightly edited:
MCWALTER: What would you say was the comparative advantage of the initial founding team?
GOUNDEN: I think familiarity plays a big part. Some of us have known each other for 10-20 years so there’s some advantages just in terms of direct communication. You can have very direct conversations, there’s no beating around the bush, there’s no kind of like mincing words, and you get a lot of stuff done a lot faster that way. Also, we tend to know each other’s strengths and weaknesses and have been able to say, Hey, this is my weakness, this is not my area of strength, and I could really use some help here. And sometimes, in other founders, some groups find it hard to admit weakness.
MCWALTER: Thinking back to those early days, what were the first 3-6 months like?
GOUNDEN: Lots has changed since that period. In the early days, we were really focused in the area of machine learning and AI to solve problems in the energy space. We actually started by working with regulators. The Federal Energy Regulatory Commission was one of our first customers. And I would say in total transparency, it was less about the software and more about solving the problem. It looked a little bit more like a consulting company with software, figuring out what software was the right software to solve these kinds of problems. And certainly, I think even today still, but in 2016 certainly, AI was very much in a custom mode and it needed a lot of hand holding. So the first three to six months in addition to working with some customers like regulators, there was also a lot of client conversations. We had a good CRM tracking system. Jackie Finn runs our marketing group and she was very insistent about us having a really good tracking system. So I think we did over 400 customer interviews in the first year. So a lot of it was just customer discovery, just understanding what the market needs and doing some paid work along the way.
MCWALTER: And what is an example of how VIA solves problems for utilities?
GOUNDEN: For example, we have a large European utility, who have petabytes of data. And they want good in-house data science people, but it’s actually really hard to attract certain skill sets to the utility industry and you see this a lot; they have six open positions and only one filled. And because they cannot often get the people in-house, they would like to make a choice of packaging the data in order to work with academics or consulting firms or maybe even general data scientists. In an ideal world, that would be great. But of course, there’s lots of confidentiality issues and so they cannot just send that data out. Instead of getting access to the data directly VIA enables third party data scientists to write up queries in Python and basically send that to wherever the data is, it runs there and results get returned. And we have a bunch of checks and balances to make sure the code you’re writing isn’t going to steal something,and the results that are getting sent back aren’t revealing too much. There’s also an audit element to make sure the utility knows who’s doing what when, and with what data. This solves for the privacy concern while giving utilities access to amazing data science talent.
MCWALTER: One of the big common problems with machine learning and those kinds of ventures is data siloing. What are your thoughts on that?
GOUNDEN: We have one customer who said they have two divisions, one in mainland Europe, and one in that In the United Kingdom, and they said it takes nine months to transfer the data from one unit to another. Mary, Queen of Scots did not take so long. It’s all these regulatory concerns that’s actually blocking that transfer of data. Our goal is that they could learn from each other without having to physically send their data. That is one of the problems VIA is solving.
MCWALTER: When a client shows an interest what is the typical kind of timeline to get a client up to the point where they’re getting real value? I imagine these jobs in general are just moving pretty slowly?
GOUNDEN: So we launched something in 2018, called the 30 minute pilot. And we thought this is kind of an industry first because so many pilots, prototypes and demos take too long. And so literally we said, well, what would it take? Let’s think about this differently. What would it take to get real data, real results in 30 minutes? So we created an environment to do that. And actually,it only takes three or four minutes to load and the other 27 minutes, is getting people comfortable with the idea that they have to go collect a little tiny piece of data on their own by digging around their files. So that’s why it totals 30. Also, we have a pilot we offer where for transformers for example, if the client is interested in benchmarking their transformer performance against the hundreds or thousands of transformer readings we have in our database without having to actually physically combine your data. So we can do that in a 30 minute process.
MCWALTER: With COVID a lot of companies moved from an office, or multi office kind of model to something fully remote, what did that transition look like for VIA?
GOUNDEN: Our VP of product the other day, he mentioned we seem to be more productive than, in a way and I said, that’s probably because, One, no one’s on vacation, Two, thankfully no one’s sick and Three, no one has any distractions or other things to do. So that trifecta means you will likely get a big bump in productivity over the few months while people have been working remote. I do think it takes work to get that to be efficient. So I’d say we learned a lot of tactical things for example, all meetings start on time. And you see people actually show up one minute early because that’s important and video is essential. And, in our Slack channel, everybody logs into the morning and says welcome and everybody says goodnight at the end of the day and that’s an important part. So there are little things that help the culture stays cohesive and I think that that takes a little bit of work. There’s also the little things like virtual pub night and trivia night.
MCWALTER: What are your thoughts on the opportunities for founders in a world that targets 100% electrification?
GOUNDEN: Because I come from the world of data analytics I have a bias which is in a world of complete electrification it means a world where we can optimize power consumption. We can learn from each other about how power consumption varies in the world. In Japan everything consumes gigantic amounts of power, the televisions, the rice cookers, the refrigerators. And if you go somewhere else, like Kenya or parts of India it turns out because power is expensive the exact same 27 inch television does not use the same amount of power in Kenya or India as it does in Japan. And so what are the opportunities for an entrepreneur? It’s saying, “Well, how do I take everyday items and make them low power? Because that helps everybody. So pick your favorite object that you know well and that can be air conditioning or rice cookers or refrigerators or dishwashers or cell phone chargers or whatever it is and ask how do I make that charge faster, smarter and with low power consumption? There’s hundreds of opportunities like that.
MCWALTER: What are your thoughts on like the increased risk from cyberattacks in a more and more connected kind of power grid?
GOUNDEN: Unfortunately, that’s increasing as an issue. You once had a house with a door and a lock, and now you’ve got thousands of windows, and it’s easier to break in. One infamous example of a retailer who had a big breach of credit cards happen by someone hacking in through an air conditioner unit. So I think that’s a real concern.The thing to do is not just depend on cybersecurity to keep up with hackers, but I think edge processing will be a great way to deal with it, in other words don’t move the data. Don’t let the data have access to a window, and do as much processing on the edge as possible, and you minimize some of those risks. Keeping it fragmented and distributed is one way of kind of making sure we are dealing with some of those threats.
MCWALTER: What are your views on things like blockchain applied to cleantech?
GOUNDEN: We actually have a small blockchain component we use to basically audit logs with an immutable ledger; it’s basically a smart contracting function that controls who can access to what data when and for what purpose and then tracking, the usage and and what data was used at what point in time. We never saw really the big crypto coin opportunity. A lot of others did, but the idea of this kind of decentralized database infrastructure and auditable ledger that makes a lot of sense, especially in a world where you need technologies that are beyond trust, or you don’t know all the parties. So, in a power grid, where I have a manufacturer of a electric vehicle, and I have a charging station from somebody else, and I have electricity powered by some distributed generation site by a third party and then a grid that connects them; all of those things may not know each other so they need trust to be able to kind of know what’s happening when and be able to have some auditable trail makes sense. So we see a good use for blockchain in those instances.
MCWALTER: Where did you grow up?
GOUNDEN: I grew up a little bit everywhere. I was born in Canada, but grew up in Scotland and I went to 12 different schools around the world. Some time in California and in Australia and I think that influenced me in a lot of different ways. You become more open to lots of things and you realize the world is filled with very different people and you have to adapt quickly to stay alive and be respectful of others? Because there’s huge diversity and what you think is normal may not be normal to anybody else. I remember going to school in Scotland and on the first day of school meeting somebody and saying “hi” to them. And “hi” was just not said in Scotland in those days. It was a very North American thing .And I remember thinking, well, that’s what you learn, you learn to adapt. So being respectful and open is probably the number one lesson that comes from growing up in so many places.
MCWALTER: I actually had not a dissimilar background, growing up in different countries. And one of things I found was you definitely started to notice, and ask the question, “Why are things like this?” And often people do not notice certain things because it is the water they are swimming in.
GOUNDEN: Right. I think this is an age old question. I think Voltaire wrote about the foreigner who comes to the court and asks all questions like, “why do you have all of these odd, behaviors and rules and standards?” The outsider often has an interesting perspective about why we do things.
MCWALTER: When building company culture is the primary lever used at the hiring stage or rather the building of certain values within the company over time?
GOUNDEN: The screening process is super, super important. And one thing we do is give little case studies,whether you’re an office admin, a developer or DevOps person, you have different kinds of case study assignments that everybody gets. It’s easy to be charming on a 45 minute call, right or in a meeting, but you need to actually see, are they interested in the work? Can they pay attention to detail? Are they committed to it? We have a value, which is “love in:love out”. If you love what you do, it’ll show in the output that you have. We have a 3-4 month probationary period and some people are like, “yeah, I take that bet. I bet I’m going to be great.” You know, that’s not much of a leap. Generally the best people are the ones who are willing to take that bet on themselves. There are probably 50 other little things that we do along the way to keep people motivated and which attracts the best people.