 from the Julia Morgan ballroom in San Francisco extracting the signal from the noise it's the Q covering structure 2015 now your host George Gilbert and we're back this is George Gilbert we're live at the Julia Morgan ballroom in downtown San Francisco at the iconic structure 2015 conference we have a very special guest with us Jerry Chen formally I guess we could say leading a big chunk of VMware and now have a move to Greylock Jerry good to have you with us thanks for having me so give us some context when when you moved over to Greylock what did the industry look like to you what were the opportunities you were looking for them you know I joined Greylock back in 2013 so almost three years ago and you know I saw there are a bunch of different waves happening technology from so the big data wave you saw the cloud wave and one of the waves that I was focused on is something I call a DDI developer defined infrastructure and what we saw at my decade of VMware was a shift towards developers as making decisions around infrastructure Amazon is a cloud is a perfect example of developers making what used to be IT decisions and so that led to a bunch of investments or a bunch of investigation by me technology like Docker which I invest in two years ago other data plays that also have been spending time in our security place I've been spending time in really looking at how developers and data scientists really are rethinking what traditional IT was would that be would it be fair to say that this was the capture the developer as the design win and then you know you have to reach IT who pays for sort of uptime I think it is fairly correct me I think all these technologies now there they're two halves to a product there's going to be a developer or data scientist or a user aspect to it and that user is going to be a developer for technologies like Docker or a database a data scientist for either a database or analytics platform or an user for like next generation enterprise apps be like a mobile client and but on the back side there's going to be an IT buyer security manageability scalability reporting auditing and so when you go to market you had to think about what's the IT buyer want all those LEDs but also what do the users want and that user population used to be either strictly IT or strictly end user but now the number of populations you sell to in a company have changed right developers data scientists business analysts mobile sales forces construction workers you know the population of users for technology has just exploded okay so let's let's take a look at that that we have what seems to be a slow motion price collapse and infrastructure software because of open source and because of cloud meter priding pricing you have to reach different constituencies so that would suggest perhaps a higher cost sell unless it's more self-service how do you square that circle you know it's um it's not really higher costs but the where you spend the dollars are different so for example open source is a great example look at a Docker or Hadoop in clad era two examples of using open source as a way to reach potential customers that reduces the friction for selling reduce of the marketing cost there's a selling cost of trying to sell the enterprise product behind the developer or data science adoption but those dollars are shifted right so whereas you used to have on a large marketing campaign to try out my new database or new language the open source has become a great way for developers and you know vote with their feet so they like the technology like the framework they'll use it and that's more or less not free you have to still invest and spend money into a developer ecosystem but the type of dollars change do the do the economics work where in the old day to get that initial sale yeah you needed the license sure you know and that essentially was a lost leader now is that initial sort of land you know of land and expand is that self-service enough so that we don't need the upfront license you know it's a great question I think there's there's two things are trying to unpack one is a shift from perpetual license to subscription right as well as kind of the cash flow issues of getting paid up front versus not getting paid up front for open source right so I think the county we were what ten years into the status business model now I think we learned that look if you can get the customer acquire the entity the subscription models actually great because it gets predictable revenues predictable cash flows you don't have to worry about renewals or lumpy enterprise sales I think the question that becomes how can you cost effectively land that customer and that's a combination of using new techniques like you know open source app stores for the consumerization of IT like Apple App Store the Google Play Store through the web so there are things you can do to reduce our cost of land but it's not for every product there's some products that are like large systems like enterprise applications that still have a large land because you're selling to the entire company so it's hard to be absolutely those they those have not gone through sort of the price collapse the applications that's correct that's correct because in infrastructure space yours you like you said you're stuck between open source and a cloud space right OSS in one side and in Amazon Google Azure on the other side and so threading that needle of how to you know use open source as as momentum driver but then sell enterprise products around or behind it is kind of a new business model that a lot of companies around here pioneering I think a couple companies involved with like Docker and Cladera are kind of proving that it's possible to build a business in this this world so let me go back on this on this topic and drill down a little bit a lot of the infrastructure companies I talked to say we're we're gonna make money by helping the customer operate our software okay in other words they understand they're not gonna make it in the initial sort of selling to the developer the design win but if you look at a matrix of sort of application software middleware infrastructure and a couple other layers and then there's performance management change management and high availability yeah basically when they say we're helping the developer run the software they're talking about like giving them a console in that one box in performance management I mean I'm exaggerating a little bit I think I think it is hard to say that's for all products right so I think from installation optimizing running monitoring upgrading tuning that whole lifecycle upgrading is part of running it right and you can either do that as a software subscription you can run as a service in the cloud but I think that lifecycle are all things enterprise will pay for yes and if it's in the cloud for sure you can capture that because you also get the integration points you're the vendors responsible for the integration points but if it's on premise yeah let's say it's you know you're a messaging product or some part that fits into a bigger stack every one of those products has their own failure model their own high availability model their own you know fault tolerance security model how do you get those to play together even if each vendor gives you a lifecycle that's been um that's and that's where everything old is new again right so I think we the pencil swings with these integrated systems where people try to sell storage compute networking a big box all the sun and then broke apart into like storage as a separate category network a separate category search separate category then how do you make those things run together support the things together as well yeah and just like the physical world you had this pendulum swing from integrated to piecemeal and they were viable business models and best of breed in software you also have folks who want to get off of the whole suite versus best of breed software products and just like the physical world where you had best of breed infrastructure physical infrastructure best of breed software infrastructure can also play well together okay so then it's a question of perhaps the class of customer like the early adopter can sure and sort of swallow the multi-vendor components so all right I took you on a digression to you know sort of operating expenses perhaps tell us what you see as some of the more promising well let's just make it open ended categories where you know maybe not the next uber but they you know what do enterprises have to do to move to you know beyond just the ERP replacement sure that sort of thing well I think ERP is a great category where you have large horizontal apps like HR CRM financials but I think different verticals also lend themselves to disruption as well I just invested in a company called Rumbix and the construction software space so construction is a ten trillion dollar vertical that historically has spent a lot of money on technology but now they're using GPS to like track their tractors and cranes using drones to measure progress and Rumbix is using smartphones and mobile technology to make your workforce more productive and safer and so you think about there's a bunch of horizontal applications like ERP mentioned but I get really excited by different vertical applications like construction like health care like financials that have plenty of opportunity for this kind of end-to-end applications that are they're very disruptive it actually can make a big difference in both the top line and bottom line for a side business and I imagine on those again part of the part of the value since it's being run by the vendor yeah the the problem with on-premise was every customer had to write the integration points and here the vendor can do it just to the common the common applications that their customers I built on-premise off-premise but I also think in this cloud world and the way apps are being built today everything is API everything has kind of a underlying plumbing so the way to integrate applications is a lot easier today and I think you're seeing a bunch of companies now come together to either build a full-stack app but can also integrate with like classic databases ERP applications security applications so I really think we're so integratable by design by design it's it's kind of like 10 years ago 15 years ago if you had a business you had to make sure you had a website now if you write an application you better make sure I have an API right because almost every app now it's going to have a way to integrate and talk to other apps by design okay all right Jerry Chen we're gonna have to leave it there no worries always great to hear your insights and this George Gilbert we'll be back in a few moments this is structure 2015 downtown San Francisco thanks