Todd McKinnon saw the first phase of the cloud computing revolution close-up. He joined Salesforce.com as its head of engineering in 2003 and saw firsthand how companies mistrusted the idea of using software they didn’t run on their own computers in their own buildings, but instead trusted to someone else.
Over time, CIOs everywhere got over their skepticism of the cloud, and Salesforce.com went on to become a billion-dollar company that is widely used in numerous industries.
But he knew there was more to it. Lots of other small companies where adapting the Salesforce software-as-a-service model for other job functions besides tracking sales relationships. There’s SuccessFactors, which tracks employee performance, and Workday, which manages basic company operations like payroll and benefits.
For companies large and small that are embracing the cloud, that’s a lot to manage. McKinnon’s plan with Okta, a start-up company that’s received investments from Andreessen Horowitz, Floodgate and SV Angel, is to make it easy for companies to take advantage of cloud applications and services, by giving them a single place from which to manage it all.
I caught up with McKinnon last week in Menlo Park, Calif., at the Sand Hill Road offices of Andreessen Horowtiz, which in July invested $10 million in Okta, its first cloud investment. The company is planning a big launch in January.
NewEnterprise: So your last job was at Salesforce.com, but it led directly to you starting Okta. Tell me about that.
Todd McKinnon: I went to Salesforce at the time their engineering team was really small. Their engineering team was only about 10 or 15 people. They brought me in to be the first non-founder VP of engineering. They wanted me to scale the company and scale the group.
NE: So you got into Salesforce early, and things look pretty good over there now. Why did you leave?
TM: We had a booth at Dreamforce, and I saw a lot of my old friends there, and they asked me the same thing. I was at the tip of the spear in terms of seeing this transition in the industry. When I first started at Salesforce, companies were first starting to adopt the cloud and the software-as-service approach, but they were really skeptical. They had to scrutinize everything about it before they would buy it. One time I spent an entire weekend–I remember because it was my birthday–reviewing the code line by line with a big bank, for security. Over the years, that attitude changed. Rather than view it as something risky, they viewed it as beneficial, and they got more comfortable with the risk.
That to me was an important moment. When the industry shifts, that creates an opportunity, the kind of opportunity that small companies can take advantage of. I couldn’t sit there knowing there was so much opportunity and so much disruption going on and watch someone else go out and do it.
NE: Clearly we’re seeing Salesforce broadening out from its original offering. We’ve seen it launch Database.com and acquire Heroku in recent days. It didn’t occur to you to try and build this within Salesforce?
TM: It did. But there’s challenges there in terms of what the company is going to focus on and how long to focus on the primary mission before it branches out. I wanted to build a company. I wanted to have an impact, but I also wanted to build a company.
NE: So explain how you got from there to starting Okta.
TM: I had been thinking about building a monitoring system for big companies rolling out their mission-critical cloud services, like call centers. The idea was to monitor performance and gather compliance data in a way that was similar to what companies were already doing with the systems they had in-house.
I called a bunch of people who said that monitoring was a good idea. But one thing I heard was that it would only make sense to large companies. The other thing was that there were a lot of companies, some big, some small, that were deploying cloud applications. I started to see companies who were running most of their IT infrastructure outside the firewall. Once you think about a world where the center of gravity is outside the firewall you have to solve a lot of problems in the cloud that have already been solved inside the firewall.
I kept running into a simple example: If you have a Windows network, you have file servers and print servers and email. Microsoft has made that work well. But if you’re trying to build your corporate IT in the cloud, there’s a bunch of unsolved problems. You have some file utility, and email from Google or a hosted Microsoft Exchange, and maybe a print driver on your copy machine. What I learned is, before anyone needs monitoring, there were all these basic problems that needed solving.
NE: What kind of basic problems?
TM: There’s the the most basic problem of identity. How do you get users authenticated in a consistent way across all these cloud services? How do I make sure that when someone joins the company they get access to what they need to do their job? And more importantly when someone leaves the company, how do I make sure that they don’t have access to all the things they did when they were an employee?
When you run your corporate IT in the cloud, all your files and services are out there on the Internet and so you have to make sure you de-provision that user’s access. There’s a great bonus with the cloud that files are accessible from everywhere, on your PC or your phone. The downside is that they’re available everywhere and so you have to manage that. If you’re truly going to reap all the benefits of the cloud, like lower costs, in the next five to 10 years, companies are going to have to rethink how they build their networks.
NE: And that’s where Okta comes in?
TM: Right. We set out to build a domain controller for this new type of network–we call it a Cloud Area Network.
NE: At most companies there’s a mixture of infrastructure that’s in the building or in a managed data center, mixed with some cloud services. What you’re saying is that if I want to manage my infrastructure on things like Amazon Cloud Services or Microsoft Azure or Google Apps, some combination like that, then you’ve got to have a way to control who can access what.
TM: Exactly. Okta sits in the middle of your services, and knows what you use. Right now it’s focused on applications: Workday, Taleo, Success Factors, GoToMeeting, Salesforce. We have hundreds of prepackaged combinations. You tell it which ones you have, and you get three very concrete benefits right out of the box. The first is that your users get a single dashboard to access them all, with a single sign-on. Administration staff gets a single point to create accounts across all of your services. And then, most importantly, when someone leaves you can automatically de-provision them, so you can cut them off from the services all at once.
NE: So what are your launch plans?
TM: We’ve been selling in the marketplace for a while now, and the main thing we’re going to talk about is customers.
NE: That was my next question. What kind of companies are you going to be talking about?
TM: These are known companies. They’re not huge. Our biggest installation is about 1,500 seats. But when companies deploy it they want to give it to every employee because it manages so many applications. One customer has 21 applications in there, and one had 15. It crosses all the job functions, and touches all employees.
NE: So if you counted up all the active seats in use right now, how many would it be?
TM: We’re going to be announcing that in January too! (Laughs.)
NE: So what’s the business model?
TM: It’s a subscription model just like all the other software-as-service companies out there. It’s based on per user, per month. The default is that they choose to license for the whole company so we get a lot of broad deployments.
NE: So where do you want to be a year from now?
TM: The big thing a year from now, we need to start to position ourselves toward our bigger vision of becoming a platform. The initial product is easy to understand. Over time we need to turn it into something bigger. We want to get beyond the applications and turn it into a platform-as-service. We’re starting with applications, because that’s where the adoption is right now. Ultimately we want to be the controlling layer for a lot more: Programming tools like Heroku and Force.com and Google App Engine, and then there’s the infrastructure-as-service layer like Amazon. Our ambition is to be a domain controller for all of it.
CastTV was founded in 2006 by a husband and wife team, and had raised $3.1 million from investors including Draper Fisher Jurvetson, Ron Conway and Marc Andreessen.
Leanback users can give feedback on whether or not they like recommendations, and the channel will refine future suggestions automatically. It seems very similar to streaming music on Pandora. The idea is to offer users a never-ending stream of videos they like, so they’ll watch YouTube longer. 
