00:00hi everyone welcome to the a 6nc podcast
00:03I am sonal and I'm here today with two
00:05general partners who are focused on all
00:07things infrastructure and enterprise and
00:09we have our newest general partner
00:10Martine casado welcome again thank you
00:13and we have Peter Levine who has
00:15probably pretty much every company in
00:17her portfolio that covers the space in
00:19his he's on the board of those come and
00:21hold general today we're gonna talk
00:25about one of the themes that's become
00:27really interesting to us is this focus
00:29on developers and how that changes
00:31everything from how you sell to
00:33developers to how you build a business
00:34that markets to developers to just every
00:36aspect of this I think one of the most
00:38notable transformations over the past
00:41five years has been the Popper to prints
00:43of the developer as a buying center
00:46within a company I used to be a
00:48developer and when I was a developer I
00:51had no budget and I couldn't buy a
00:53pencil like when pencils were popular
00:56you know what but like I couldn't buy
00:58anything basically whatever central IT
01:02had ordered that's what showed up on my
01:04desk and I wrote code and I compiled it
01:07now popped the program but I could not
01:09buy any I wouldn't didn't even know how
01:11to buy anything other departments in
01:13company sales and marketing an IT and
01:16G&A all had buying centers and budgets
01:20to go buy things developers didn't have
01:22anything until about five years ago
01:25that's when github got started and kind
01:27of this whole collaborative sort of you
01:30know let's bring developers together as
01:32a community and as software infiltrates
01:36every part of our economy it's not just
01:38tech companies with with programmers
01:40it's every company with programmers it's
01:43kind of the tail wagging the dog that
01:45these are the lead innovators and
01:49they're the lead buyers and companies
01:50and what we see are many of our startup
01:53companies now deliberately selling to
01:56developers as the first wedge point into
02:00an organization so 30 plus million
02:03developers and they all very much have
02:07opinions and buying potential and that
02:10has now become very well noticed among
02:14company's selling into enterprises
02:17convinced me a little harder that this
02:18isn't just a Silicon Valley thing that
02:20that that this is just the perspective
02:22that we who said in the heart of tech
02:24are biased towards this view because
02:26that happens to be the world view we
02:28have our big companies seeing this sort
02:31of transformation play out in their
02:32companies as well every company has
02:36software development inside whether
02:38you're a tech company or not that's fact
02:40whether you are GE or a financial
02:44institution or the government or a
02:47retail organization or an e-commerce
02:49organization so this is not a Silicon
02:52Valley phenomena this is a worldwide
02:54phenomena where there are developers in
02:57every organization building software
02:59they are the ones building the most
03:01innovative products in those
03:03organizations to move any industry into
03:07a next generation company so we are
03:10seeing across the board the
03:13proliferation and the opinion leadership
03:16of the development organization as again
03:20a key buying influence in every large
03:24so then how then do they get the budget
03:25to be able to buy those pencils or so
03:28the equivalent because I mean them being
03:30the center of power is clear but now
03:33does the company how did they get that
03:35budget the thing that's so powerful
03:36about the developer is they line to the
03:39line of business right so generally they
03:41line to the line of business which is a
03:42profit Center so if I'm a developer for
03:44you know a you know a piece of a car or
03:46whatever that you know the selling the
03:48car is what makes the company business
03:49where IT is often viewed as a cost
03:52center so it's just very natural
03:54business dynamics you see budget
03:56shifting towards the profit center and
03:57trying to contract on the cost center
03:59behavior always follows business so the
04:01business will very naturally align you
04:03know behind the cost center the other
04:05part of that is the ease by which
04:08developers can consume infrastructure
04:11programming tools open source is so much
04:16easier now than five years ago a while
04:19back the reason why we had IT
04:21departments is because we had on-premise
04:23software you had a first in order for
04:25anybody to even use this stuff you had
04:28her lab put the on-prem software on
04:30those servers do a beta trial internally
04:33now it's all try before you buy let me
04:37just give you an anecdote from the
04:38infrastructure side to kind of
04:39demonstrate how pervasive this is from
04:40an infrastructure perspective is now
04:42becoming something that programmers
04:43program - it's becoming a nice
04:45abstraction to programmers I think for
04:47example the containers phenomenon is
04:49infrastructure being consumed by
04:51developers the container is basically a
04:54piece of infrastructure abstraction that
04:56programmers can use so they have app
04:58portability containers can run on bare
05:00metal they can run on VMs but from the
05:02perspective of a developer it's a piece
05:04of infrastructure that if they built it
05:05to it and they can run it you know in
05:06Amazon or locally or anywhere else it
05:08just makes that portability really easy
05:09to do now it's something that developers
05:11rights to so now there's you know
05:13somehow a bleeding between developers
05:16and what's typically been in IT budgets
05:18there's no resistance anymore to really
05:21trying and embracing the best tools and
05:23once the developers in an organization
05:26have coalesced around the set of great
05:29it becomes a foregone conclusion that
05:31the company ought to step up and go
05:33purchase those particular tools and
05:36services so it really has got much
05:38easier for the individual developer and
05:41teams of developers to embrace and
05:43incorporate new technology into their
05:46into their workflow as opposed to the
05:49pass let was literally handed down to
05:51the developer so that also sounds like
05:53it would make it a lot harder to sell to
05:54developers because then you have
05:56developers all with their different pet
05:57I mean yes there's a period before they
05:59coalesce around a popular single one or
06:01two tools where they're just trying out
06:03a bunch of different stuff and at least
06:04before you had a direct arc to go to
06:06central IT department and say I'm gonna
06:08sell you this and that was that so how
06:10then do you know sell to developers you
06:12know any anytime users have more choice
06:15and more flexibility it benefits
06:18everyone and there was a time where IT
06:21would go make a decision and many of us
06:24have heard or have used the word shelf
06:26where many times where IT would buy
06:28something and no one would use this
06:30stuff because it was some slick sales
06:33guy who sold to some golfing partner who
06:35happened to be running the IT department
06:38and that's how the that's how the
06:40software got deployed and then
06:41got rolled out and no one would use it
06:43here we have the democratization of
06:45applications of developer tools through
06:48open source through SAS that I believe
06:51really moves the needle in terms of
06:54innovation and the best products are the
06:56ones that win now that doesn't mean that
06:59that's the end of the sales cycle that's
07:02really the whole indoctrination of
07:05open-source and SAS as a starting point
07:07is a marketing vehicle to build a
07:11pipeline for an organization and then in
07:14order for that to be sold into the it's
07:17sort of the upside down approach you
07:19start from the bottom and then you sell
07:21to the top in the old days it was sell
07:23to the top and then proliferate down
07:24here the real users start with it and
07:28then there are typically enterprise-wide
07:30licenses and all of that so you use the
07:33open-source SAS tools and services as a
07:36marketing and pipeline development and
07:38also a way to influence innovation in
07:41your company and then layer up an inside
07:44sales or a sales organization that then
07:46can sell value into that company to
07:50expand the footprint that's what we see
07:52happening you can't underscore this
07:54point enough which is the developers are
07:56influencers the developers have control
08:00and a large lateral voice on budget but
08:03in my experience they aren't explicitly
08:04the buying Center the buying Center is
08:06still it could be a dev manager it could
08:07be a VP of engineering it also is very
08:10often still IT or someone higher in the
08:12organization and so that means that you
08:14have to be very good at two motions
08:16right and you know actually appealing to
08:17the developer and still having an
08:21sales force which which again this is
08:22not common that you see both of these in
08:24the same in the same and and the most
08:26successful companies who are gonna crack
08:28the code on this will do both of these
08:30excellently yeah you have to do both
08:32really well and just to say that we're
08:35gonna have a free offering that's going
08:38to virally penetrate the enterprise I
08:41think it's a false hope it doesn't
08:42happen that well there is a point every
08:44company has to sell higher in to the
08:45company in solution why is that though I
08:48mean I still don't quite buy it because
08:50if it's something is taking off popular
08:52like my share popularity why do you need
08:55why do you need to sure that companies
08:57yet right now often why you have to sell
08:59higher in the organization is because
09:01the technology is being sold across a
09:03cut across silos for example when you
09:05look at cloud as a general thing it
09:07touches many traditional aspects compute
09:09networking storage and security so you
09:12either have to go and get every one of
09:13those decision-makers on board or you
09:15just go to someone that has all of them
09:17under them and sell to that point which
09:19is why you know existing companies with
09:22c-level CIO level relationships are
09:25successful at those higher level sales
09:27and like when you see companies like
09:29like startups for example move into
09:31infrastructure you want them to sell
09:32higher into the organization so they
09:33don't have to fight for different
09:35battles very similar to my answer but
09:38I'll start from the bottom so let's say
09:40we have a particular silo in an
09:43organization that has organically
09:47adopted a tool a developer toolset and
09:50that's in one organization here's why
09:53you need a sales organization let's say
09:55the other silo is haven't quite adopted
09:57that particular tool set and let's for a
10:00minute assume that those other silos
10:02will ascribe value to those services in
10:07slightly different ways than the
10:09original silo what a salesperson can do
10:13is bring highlight the value of a
10:17software product as different from wait
10:20maybe when one group uses it they get a
10:23value but they may not know even within
10:25that silo they may not know that there
10:27are ten other features did you know that
10:29if you did this you now have
10:32collaboration with a remote office and
10:35you might never know that and so within
10:37another silo it might be did you know
10:40that this product has these features and
10:41it could connect to this and what a
10:43sales organization is able to do is to
10:46promote maximum value for a product
10:49where is not self-evident through the
10:51use of that particular product in it of
10:53itself and that's why you need a sales
10:55organization on top of this what I'll
10:57call the freemium model I totally agree
10:59and from a predictable business driver
11:02standpoint organic growth is very
11:04difficult to control and if you're
11:06building a business organic growth is
11:08naturally fragmented
11:09you want to bootstrap drive-through kind
11:13of organic growth but then you need a
11:15real discipline enterprise engagement
11:17point to turn that into a predictable
11:19monetizable business and this includes
11:21things like for example using a
11:23beachhead within the organization and
11:24then driving it throughout through
11:26consistent messaging the enterprise
11:28sales force is is something that will
11:30drive any of these companies into being
11:32a large company yeah so you know one of
11:34certainly my personal qualities around
11:36this is it's very obvious to me that
11:38open source is becoming more important
11:39on the buying cycle because I mean a
11:41it's a way to get access to a lot of
11:43developers but also you get credibility
11:45with those developers but at the same
11:47time it raises all these questions in
11:50the company how do you market to
11:51operators like IT buyers you know like a
11:54lot of IT buyers they like
11:55certifications like the CCI is a classic
11:58one I mean a very classic way of
11:59marketing to operators or IT buyers is
12:03you create class rooms you put them in
12:05and you train them on specific types of
12:07products and that's not how developers
12:09like to to to buy it all developers
12:11don't like to take classes about other
12:13people's products they'd like to
12:14understand everything they like to be
12:16engaged and that's why I've always kind
12:18of maintained that open source is a
12:19phenomenal marketing tool to get
12:21developers engaged but then as a
12:23business now you have another question
12:25which is okay if you have open source
12:26out there to get the attention of
12:27developers what does that mean to your
12:28business model and your ability to
12:30monetize what's a viable business model
12:32you know there are a number of models we
12:33see out there one of them is you give
12:34everything away open source but then you
12:36end up becoming you know basically a
12:37service company or you could do open
12:39source and then come up with a closed
12:40source product behind it but this open
12:42core model hasn't shown to be a lot of
12:43success yet or maybe you know there's
12:46some open source and there's a service
12:47model on the back end so this there's
12:48now there's all these open questions on
12:50what's the delivery model how much is
12:52what role does it play how do you do
12:54pricing and packaging and so forth well
12:56let's answer those questions well I
12:58believe that the most successful open
13:01source companies not only will innovate
13:03on technology but they're going to
13:05innovate on their business model the
13:07problem has been about how you really
13:09make money as compared to the
13:10proprietary counterpart has been a bit
13:13of an enigma right wait sometimes in my
13:16work here and not there and every end
13:18and the original sort of thinking was
13:20well we'll just go copy the Red Hat mom
13:23which was you give away all the software
13:26for free and you charge for support and
13:27maintenance and that model has worked in
13:31my opinion exactly one time that's
13:34worked for Red Hat and not so great for
13:36everyone else and that's because Red Hat
13:38started at a time where open source open
13:41source wasn't even known they had a lot
13:43of time to go build out their scale and
13:45capacity to do the two things that they
13:48do incredibly well which is support and
13:50service and updating products online and
13:53for other for startups to come in and
13:57reach that scale is very difficult to go
13:59and do and then of course it's open
14:01source and Red Hat can grab the bits
14:02just like anyone else and it becomes a
14:04bit of a free-for-all I am very
14:07skeptical that we are going to see
14:11franchise businesses as a large as
14:15proprietary companies being built on the
14:18Red Hat model and I think moving to
14:20models like software open-source as a
14:24service so if I run open-source as a
14:26service offering in the cloud you can
14:30monetize it any way you want there is no
14:32differentiation between free and paid
14:35you just the customer pays for the
14:38service and so I am very bullish on
14:41companies that use open source and stand
14:43up that open source whether it's part
14:45proprietary or part open but standing
14:48those products up as a service
14:50completely eliminates this
14:53disaggregation of the pricing model that
14:55occurs with free and and and support I
14:59think one more level of nuance to add
15:01which I find often gets lost is that if
15:04you look at these examples of quote
15:07unquote successful open source companies
15:09in almost every case they were going
15:15after and cannibalizing an existing
15:16market which is way more difficult than
15:19actually doing new innovation creating a
15:22new market and people miss this like
15:24they look at open source like oh this is
15:26like Red Hat was massively innovative
15:27and like the reality is it's much easier
15:29to say listen you've got Unix you pay a
15:31lot of money for it I've got Linux which
15:33isn't as good but it's a lot cheaper and
15:36develop it and so I mean when I look at
15:38my sequel I think okay it's going after
15:40an existing very mature database market
15:41if I look at Android even though that
15:43wasn't a standalone independent company
15:44you know palm and and Blackberry and
15:48iPhone created this massive market and
15:50so open-source not always but often is
15:53something that enters these existing
15:54markets is a commoditizing for us and so
15:56that almost already limits what you can
15:59do from a company perspective and so
16:01like when you're viewing open-source and
16:03what it means I think have to be very
16:04careful about and very clear about how
16:06are you gonna monetize what the impact
16:08isn't in business and what type of
16:09market you're entering so I think
16:10another way to monetize is let's assume
16:13that open-source proliferate sand is out
16:16there there's a very to me there's an
16:19interesting trend where companies ought
16:22to be moving up the stack to where you
16:25build on open source and you build
16:27whether it's proprietary or open build a
16:29great product that leverages the open
16:31source base so you're not necessarily in
16:34the open source business but you're
16:36leveraging the pearl of proliferation of
16:38open source using standardized api's and
16:41building data and analytics products or
16:44applications that now sit higher up in
16:47the stack and I believe that that's an
16:49excellent new business model based on
16:52the what I'll call the commoditization
16:54of some of the infrastructure components
16:56with new products and services and then
16:59as a start-up you can put all your would
17:02behind a single arrow it's very
17:03complicated as a start-up to do all
17:07ie manage an open source project and
17:09build all these applications on top so I
17:12think that the the the sort of
17:15separation point on these things will
17:18very likely unlock some really
17:20interesting franchise opportunities are
17:23you guys seeing any interesting examples
17:24out there of people doing really
17:26interesting examples of this or other
17:27monetization with open source you know
17:30one of the companies that were invested
17:32in a Remo formerly a de Tao builds
17:34machine learning and big predictive big
17:37data applications that sit on top of
17:40spark and and Hadoop installations they
17:44don't they don't support the spark and
17:47Hadoop they let that they leverage that
17:50once it's out there and so all of their
17:52development efforts are for these higher
17:54level applications in the past open
17:57source hadn't been so prolific and now
18:00that it's more widespread it's easier to
18:03deleverage that base and build things on
18:06top of it because there's actually a
18:07base there so then clearly there's areas
18:10where startups can focus in this space
18:12but what are these are the pitfalls or
18:14obstacles that they have to they should
18:15think about as they try to build these
18:17businesses and how can they overcome
18:18them so the first one is if you're doing
18:20open source really understand the impact
18:23to the value of your ultimate product if
18:24basically that becomes your product know
18:26that you're turning into a services
18:27company and if you're doing an open core
18:29model understand that you know pretty
18:31much nobody's pulled that off or not
18:33with the type of success you'd want to
18:35be clear it's not so much that it's bad
18:36to be a services company but there's not
18:38that much profitability and its services
18:40company it's a it's a different it's a
18:41difference traction of company so just
18:42understand the impact to your pricing
18:45especially like your repeatable revenue
18:47like license revenue on open source the
18:49second one and I Peter out the sums you
18:51know as you move up the stack which is
18:52important realize that it's a more
18:54fragmented market and so you have to
18:55come up with some way of dealing with
18:57that fragmentation I mean that the
18:58classic example is the languages like
19:00like the problem with languages is
19:01there's so many of them and it's so
19:03fragmented and bifurcated so as you move
19:04up the stack you tend to have this
19:06problem and and you know what we said is
19:08one thing that you can do to rein that
19:09in is actually have a direct sales force
19:10that actually shows the value across and
19:12then the third one is open source is
19:15hard enough to sell on its own sometimes
19:16because you're talking about you know
19:18less features sometimes or whatever so
19:20know if you're entering a mature
19:22existing market and your value is
19:23basically monetization or understand if
19:25it's really innovation I mean market
19:27category creation is alone is hard
19:29enough by itself so if you have to do
19:30you know create a market category and
19:32explain kind of some new model something
19:34else it may be too many things so I
19:36think like understanding whether you're
19:37trying to do market category creation or
19:39whether you're trying to enter an
19:40existing market will impact the way that
19:42you bring these technologies out and
19:43then maybe a couple of others just that
19:45we you know highlighted developers want
19:48to use not buy so you know use before
19:51selling to and it will naturally take
19:59Organic virality works to a certain
20:01point but in order to be a successful
20:03organization in this new ecosystem the
20:07notion of viral spread plus an
20:10enterprise sales organization I think is
20:12the winning formula so then to
20:14operationalize a little bit more what
20:15does that mean for hiring and the order
20:17have build up your workforce for a
20:19particular startup does this mean now as
20:21a rule that you need to hire a community
20:22manager that's really a great point I
20:25think Community Managers to developers
20:27is actually a very important early
20:29business function almost before a sales
20:32organization right so you want to
20:33coalesce that community and then layer
20:36up a sales organization because that's
20:38the community that's going to try your
20:40product and it's where you're going to
20:41get your pipeline of users that you can
20:44then feed into an inside sales
20:47organization following the adoption of
20:49your product actually think along with
20:51that and it's it's kind of not what
20:53companies often think about you know in
20:55order to get something virally adopted
20:58should be it should be obvious and a
21:00given but you're right so you end up
21:01putting basically your best development
21:03resources in my case I put my best
21:04developer on the open-source side of the
21:06house which from a company is kind of
21:07difficult to do because you basically
21:09want that best developer do we work on
21:11the core product on the inside but you
21:13know a community managers are great a
21:14community manager is not going to drive
21:16adoption like the best engineer on the
21:18planet building something good so I mean
21:20if you're gonna be doing this you mate
21:21you have to make sure and there's a
21:23tension internally that your best
21:24developers are one that are driving the
21:26open-source project because that's how
21:27you'll get real credibility and and then
21:29you have to realize that that that
21:31developer is not contributing to the
21:32core I also think along these lines you
21:35have to navigate the roadmap of what
21:38goes into open source on how you think
21:40about your business model exactly
21:42because you could you know now you put
21:45your best developers on open source and
21:47there is a very natural tendency to go
21:49make that product arguably the best
21:51product it is go add all the features
21:54and all the functionality and if your
21:56business model is the assumption that
21:58you're going to upsell something into
22:00that but your best developers have
22:02already created that an open source
22:04you're kind of you're kind of down a
22:06path where there's nothing to upsell
22:09you have to understand the dividing line
22:12very clear in an organization about what
22:14goes into open source and what and and
22:17how you monetize all of this as opposed
22:21to just say okay build the best product
22:24you can and then we'll figure out this
22:26segmentation we've met with a number of
22:28companies that are trying to do this and
22:30I'm always very interested in how you
22:31structure the organization like who does
22:33the open source team report to because
22:35if you have them reports like the core
22:36product team I mean there's a lot of
22:39internal dynamics that have to me this
22:40is fundamental tension that was actually
22:42my next question which is it would kind
22:44of close this out then how does this
22:45change all the knowledge we have about
22:47how how to do product management that
22:49we've all learned through these
22:50traditional companies it heightens the
22:54importance of really putting a go to
22:56market plan in place early on because
23:00the natural tendency we have been
23:02conditioned to listen to our customer
23:04and I can tell you what a customer says
23:08when you release open sources make it
23:10easier and more functional and I will
23:12adopt it so let's say we now take that
23:14back to our developers and they say
23:17great we're gonna add let's take all of
23:19these features that we thought we'd go
23:20charge for and throw it into open source
23:23because that's gonna increase adoption
23:25but the defeating point there is that
23:27well there's no revenue anymore because
23:29the customer has taken all this stuff
23:31and it works great and they're using it
23:33everywhere but we don't have anything
23:35else to sell and so having a real
23:38awareness on that and I'm not saying
23:40don't listen to customers but I think we
23:42have to be very very careful in terms of
23:46that fluidity between open source and
23:49the revenue opportunities for the
23:52company and to make sure we always
23:54understand the trade-off between adding
23:57another feature into open source versus
23:59the monetization and go to market plan
24:03that we established when we started out
24:05and so I think it becomes important very
24:08early on we're in the past it didn't
24:10really matter that much you kind of
24:11figure out features as you go and
24:13customers would tell us and all that
24:15stuff and now you have to be a lot
24:17clearer and a lot more intentional about
24:19your plans upfront so that when things
24:22change you actually have a milestone by
24:24to measure the change against and I and
24:27I think organizationally you almost have
24:29to reflect that which is I give the open
24:30source is subservient to the closed
24:32source it is very unstable because
24:33basically the the closed source product
24:35is going to keep it anemic in a way that
24:37feels you know like like it like a
24:39negative dynamic and if you happen to
24:41the other side of the open source
24:42actually controls the closed source you
24:44know it would be very difficult to
24:45enforce you know that separation so I
24:48you know I've tried both models and I
24:50ended up doing the one where they're
24:51actually separate teams so that someone
24:53above them that actually is very clear
24:55about the go-to-market about the
24:56dividing lines is able to arbitrate now
24:58that's just my personal experience but I
25:00do think it's something that
25:01entrepreneurs should think through is
25:02how to structure the teams given that
25:04you've got this very natural tension
25:05internally that's great you guys well
25:07thank you thank you awesome