Welcome to the 7th episode of our podcast!

You can download the podcast to your computer or listen to it here on the blog. Click here to subscribe in iTunes.

CP7: Podcast with Kumar Ramachandran from CloudGenix about Wide Area Networks (SD-WAN)


Martin: Hi today we are here with CloudGenix CEO Kumar. Hi, Kumar who are you and what do you do?

Kumar: Thank Martin, I’m Kumar Ramachandran, founder, and CEO of CloudGenix. What CloudGenix does is generally what we call the Software Defined-Wide Area Network or SD-WAN for short. What we do is transform the way remote offices are connected back to the data center and clouds by allowing customers to use low-cost connectivity by eliminating the need for the traditional complexity of networking and really basing a system that is based on application and policy definitions instead of lots of complexity that seems to rule todays networks from the legacy vendors.

Martin: Cool, when did you start CloudGenix and what did you do before?

Kumar: So we started CloudGenix in early to mid-2013. I spent several years of my career between applications and networking infrastructure. While I started my career out building applications for companies like Citi Bank, I migrated to networking. I had a long stint at Cisco where among various things I was responsible for remote office networking and branch networking products, which is the space that we are disrupting now.

Martin: Cool. Kumar, at what point and time did you think for yourself: “Hey, I want to start a company?”

Kumar: You know for an entrepreneur usually it is a combination of things. I usually like to say that there are at least two or three things that I look for myself. I think that space is big enough and interesting enough to form a company around it.

So, the first thing is that I like markets that are very large. The remote office networking, the WAN category is anywhere between $25 billion to $40 billion depending on which analyst and how you count the numbers in terms of total addressable market. That makes it a great opportunity for the team and the investors.

The second thing I like to look for usually is that you are delivering a significantly transformative customer value. Usually, I like to think of it as 10x better than any existing solution out there and that is exactly what we did with CloudGenix. My co-founders, in terms of head of engineering as less my technical architects they came up with a way to do networking that is 10x better while being 50% less expensive for our customers. So that’s why we decided to found this company.

Martin: Great and how did you start out? I mean how long did it take you develop the technology for the MPV version? How did you acquire and convince the first customers?

Kumar: That is a great question and a topic that I am passionate about. I like to believe that I am a student of the lean start-up philosophy. I have had the pleasure of taking Steve Blank courses during my MBA at UC Berkeley. Steve Blank is actually the godfather for the implementation model of lean start-up. It’s like Eric Ries the person that wrote the book on “lean start-up”, he was the student of Steve Blank too.

Really what Steve brings to the table and the practical implementation that Eric brings to the table is a model where you look at customer development not as something you do after building a product, but really up front in the process, where in you are building a product with validated information that indeed this product has a great market.

When we got started, even though we have deep building experience, we didn’t go about building all the bells and whistles engaging in a three-year project to develop an enterprise class solution. What we did is we went to customers to whom we initially showed the slidedeck where we subsequently built prototypes. Made the customer deploy those early prototypes in the labs and the production environments even in some cases. And what that gave us is valuable feedback that initial things we were building were adding immense value to these customer organizations. Right from day one we have been focused on doing three month spent while we have been delivering those products to customers in a very meager manner.

Martin: How long did this first or initial phase of the customer development take you? How many months?

Kumar: It took us about 6 to 9 months to purely refine the product we were building. Compare to a lot of start-ups everywhere, this is shorter time frame for us, it’s because we were very familiar with this space and the customer feedback was very clear and very strong. What that allowed us to do really is that space of 6 months we narrowed down exactly what we were building, how we were building, which market segments we were targeting, and after that we have just been making sure the product is hardening.


Martin: Kumar, let’s talk about the business model of CloudGenix. What type of customer segments are you serving what type of value proposition?

Kumar: I like to categorize the CloudGenix customers in two ways. One is because we are a remote office solution any vertical that has a lot of funnel in terms of their presence is a target customer for us. So some of the key verticals for us are:

  • retail, you have a lot of retail store locations,
  • healthcare that you have lots of hospitals,
  • banks and financial services,
  • manufacturing,
  • any place where you have funnel.

Now I’ll give a familiar example of a specific customer and do a little bit of color on the value to that. Columbia Sportsware, one of our regular customers, besides working with them, they really had three big problems. One, their WAN costs were way out of whack with that infrastructure cost. Two, their ability to deliver high definition video, a lot of rich media in-store video. Three, they have operational effort required to keep the network up especially through the holiday season when they need high availability was extremely taxing on the team.

We took on all three challenges head on. With CloudGenix they were able to use their Wide Area Networking cost by 50-70%. Second, they now have a network infrastructure all which they can deliver Omni-channel initiators in retail; things like video kiosk, things like rich media to the store, etc. Third, operationally because they are providing a highly available model for networking their operational teams have had an exceptionally high performing holiday season dilemma and now we are into 2015 where the marks are for network downtime in the storage has declined while the operational team itself doesn’t have to do any manual intervention. All this then goes back to the builders. Infrastructure and IT exist to drive and help the builders.

So now what we are now translating for them is:

  • one, better up time for the stores means better sales;
  • two, the Columbia Sportsware to expand like trying to roll out the new stores is reduced dramatically. Now when the business says, we need to make a change or we want to open a store IT is not a bottleneck.

Martin: Normally it’s with technology driven companies there is some kind of very interesting technology in the first place, which is very new, which then translates into business opportunities where some industries are disrupted and so on so forth. In your case, it seems to me that this technology is SD-WAN. Can you tell us a little bit about that? What it is and how it is used?

Kumar: Okay, absolutely. SD-WAN or Software Defined-Wide Area Network, there are some core principles to it. If you look at the old model of how to build the network you actually have to put physical hardware at every one of your remote locations. Oftentimes, these were custom pieces of proprietary hardware that are very expensive, very difficult to maintain from legacy vendors. Second, you have what is known in networking as really a fragmented system. Think of it in terms of a central big brain creating an operating network. The brain is fragmented across all parts of your system. A fragmented brain is very difficult to manage and very brittle. Third, because the brain is fragmented it relies on extremely expensive connectivity from the carriers.

What SD-WAN does with CloudGenix is:

  1. We have moved away from the need for using proprietary hardware for networking. You can obtain an update from where ever you want and you can run our solution as software.
  2. There is the central brain that builds and operates the network. Due to the need for a lot of complexity that comes from a fragmented model that incumbent vendors put out today.
  3. What we are able to do with our customers, we start going to cloud, we start looking for being able to take advantage of fully automated solutions in the CloudGenix model.

All that you do as a customer is you come and specify application intent. Now, what does that mean? So you are typically coming in and saying: “Okay for this particular application here is my security policy. Here is my performance policy. Here are any compliance requirements that I have. Once you specify that the CloudGenix system will be able to take it and build an operator network on behalf of the customer. What CloudGenix SD-WAN does is eliminate the need for hardware, provide a big central brain, and eliminate the need for a lot of traditional networking complexity.

Martin: At some point after you started the company you need to think about a sustainable competitive advantage. Something like an IP, patent, economies of scale, or some kind really interesting way of how to reach and serve the customers. What is your sustainable competitive advantage? At what point in time did you start thinking and executing on it?

Kumar: The core differentiators of the company really come down to the fact that we are providing an application based model to go build your network. Now there is absolutely nobody in the industry that has successfully done this before. So it’s truly path breaking innovation on that front that if you think about it in the past anyone had to build a network you literally had to go into the bits and bytes to go build that network. The analogy I like to give in application development is that if you were using an assembly language to write your code you probably won’t be able to build a model on that site like Google, Facebook, or anything like that. The same is true in networking. Today the language that the networking administrators are given by the vendors is an assembly language. What we are doing is providing a model language to that administrator that the administrator is able to talk the language of business to our system, capture that and our system builds a network on behalf of the administrator. Doing that sometimes requires a lot of technology under the hood that none of the other vendors out there have been able to build in the last few years.

Martin: Cool, now I would like to talk to you about some kind of parallel between the networking, natural or biological systems. Can you elaborate on that?

Kumar: The old model of building a network is if your brain was completely fragmented. The human body actually has an centralized brain. Our nervous system communicates information that we just get from our body, back to the brain. The brain makes decisions, sends those decisions back and that is a great model. We all assume to function exceptionally well. Like, my hand have certain independent capabilities: I can pick a pen and that brain can communicate just enough information. I am fascinated with how those networks are being built today rather than in the past.

The way networks are built then is really that it is how your brain would be fragmented all over the place and you are struggling to clear that coordination. So now if you have to even pick something with two hands somebody else would have to help coordinate these two hands together. That is why you have a lot of network administrators needing to program and manage the network, which is very brittle.

What we have really done is moved to a model that you can have a central brain, a TCN Wireless Network  sees all the devices in the network and orchestrates them. Like the human body is a great example of how these systems work. Large scale networks really need central brains. Now the trick there is to make sure the access to the brain is super fast, the brain doesn’t become a bottleneck that you have a large capacity in the brain that you don’t have a headache or anything else. Those are all things that we brought to the table as the path breaking innovation. That is a great question by the way.


Martin: Let’s talk about your learning’s over the last years. So what are the Do’s and Don’ts that you can share with other people who are interested in starting their own company?

Kumar: You know by far the biggest thing I would say goes back to the principle of lean start-up. There are way too many companies that we see where engineering teams hunker down, build to an early product and then they see product market fit after that. In fact, this is similar to ours. I see companies and what it causes them to do is waste a lot of engineering resources and time. Money and time are backup to the startup you are finding.

In place of lean start-up get a certain validation. I like to say, don’t believe the nice between the ears, Steve Blank. Get out of the building, deliver your product to your customers, and get the validation. Getting verbal validation is not enough. Don’t be afraid to ask for responsible validation process, absolutely don’t be afraid to ask for things back from the customer. Ask the customer to test the product in the lab. Ask the customer for appeal. Ask the customer for something. There is a value to the customer or the customer organization because that is what provides you validated feedback.

What you will find is that it doesn’t matter how early the product, how early the market, you’ll find those early adaptive customers that referred to as beacon light, that will tell you if you are on the right track and on the right formation of the MVP.

That applies to anyone going down this path that in the end most start-up die of indigestion, rather than starvation. Be very judicious about what you build and how you build it.

Martin: What are the main or the key questions you would like to answer during the customer development process?

Kumar: During the customer development process, the high level single biggest question is that: “Am I solving a problem for the customer for which the customer is going to pay me money or compensate me otherwise?” This building process has to be about my assumption as a great engineer will figure out a solution to tough problems. So finding innovative solutions if you have a great team you are going to do that. The question really is that innovative team building innovative solutions for a rich problem set.

The single biggest question that will get answered is that if you think that you can potentially solve three or four problems for your customers, get validations against that. If those problems are really that important to your customer even when you don’t have a product, even when you are in the prototype phase, you will find that the customer is either investing in you by giving up their time, taking your early prototype into a lab or even cutting appeal for you. But you have to get something from the customer back. If you are just showing slides to a customer and asking the customer for feedback that is oftentimes where a customer is too nice. So it is critical to get to a place that you getting something of value from the customer. That is what we tell you that if the customer feels that you are really solving something worth paying for or are they just being nice by telling you what you are working on is great.

Martin: What have the customers offered to you? Was it like cash, payment, or some other type of value?

Kumar: You know in our case the thing we asked our customer was to take the product into their labs and spend a week or two testing it, then give us feedback. The only thing more precious for most customers in IT these days is their time. In fact, time is more valuable than even money because they guys are doing so many projects. Sometimes we didn’t even ask them to get approval from their managers to work on this project. It is not a side project, but it is really a formal evaluation. What that did in the very early stages is that it gives us very direct feedback. Some of the problems that we thought that we would solve, we had actually customers come back and say: “You know, I’m not sure that its interesting, but I don’t think I want to take the time.” And then you knew that that was real feedback. You didn’t know the people so the slides were great. The fact that they are not willing to take your product into their lab, you say: “Okay, this problem is not worth the time of this customer.” So you will find a different customer or find a different problem.

Martin: Before you raised your first round of financing what type of validation did you try to show to the investors?

Kumar: We were a little early compared to maybe software based companies these days. We did get validation from customers in two ways. One was we built a lot of mockups. We didn’t have a functional product because it takes a lot longer to build a functional product. But for some customers we had slideware and then we went a bit further with PMO that we actually built pick through mockups. So the customer could pick through the mockups and get a first-hand sense of how the product would work.

Martin: Cool, great, thank you so much for sharing you knowledge Kumar. It was a pleasure talking to you. I wish you all the best and luck with your new business.

Kumar: Thank you very much, Martin.

Martin: Great, thank you.


Thanks so much for joining our 7th podcast episode!

Have some feedback you’d like to share? Leave a note in the comment section below! If you enjoyed this episode, please share it using the social media buttons you see at the bottom of the post.

Also, please leave an honest review for The Cleverism Podcast on iTunes or on SoundCloud. Ratings and reviews are extremely helpful and greatly appreciated! They do matter in the rankings of the show, and we read each and every one of them.

Special thanks to Kumar for joining me this week. Until next time!

Comments are closed.