00:00hi everyone welcome to the a six in Z
00:02podcast I'm sonal and today Michael and
00:04I are speaking to evangelist amou DS who
00:07is currently a VC but started his career
00:11at Cal Tech where he was one of the
00:13original researchers of machine learning
00:15and has some of the very early patents
00:16on machine learning then he became a
00:18software engineer and he went to Digital
00:19Equipment Corporation and then to IBM
00:21and basically the reason we were having
00:23him on the podcast today though is that
00:25he's recently co-authored a series of
00:27posts that focus on the topic of
00:29corporate innovation which we think is
00:32really interesting given that so much
00:34innovation especially lately is somewhat
00:36startup led yeah those two words don't
00:38necessarily always go together they
00:40don't and they they've tried this is not
00:42the first time that this is being tried
00:45but I want to say that there is a lot
00:48more seriousness and introspection this
00:50time around oh really
00:51seriousness an introspection this time
00:53around what do you mean so I think that
00:55corporates both based on the lessons
00:58they've learned in the late 90s which
01:01was their their last foray into the
01:03valley and other startup ecosystems as
01:06well as the disruption that they're
01:10starting to feel today from startups
01:13they are trying to they're thinking a
01:15lot more carefully or they want to think
01:17more carefully of what is the right way
01:20to approach ecosystems such as the
01:23valets and of course this time the world
01:27is a lot more complex we have more than
01:29one ecosystem that matters we have many
01:32more technologies that matter but again
01:35they're coming into it these the ones
01:37that are serious about it with with a
01:39lot more more serious intentions right
01:42well one of the trends that you and
01:43Steve Blank point out is that there's
01:45been a recent wave of companies setting
01:47up quote innovation outposts in Silicon
01:49Valley Amazon has a lab in Silicon
01:52Valley the DARPA now has a thing and I
01:55mean they've always partner in Silicon
01:56Valley and actually funded much of the
01:57early Silicon Valley for sure but
01:58there's now a formal organization it
02:01seems like every major company in the
02:04world has a innovation outpost in
02:05Silicon Valley and elsewhere one thing
02:07is interesting is you guys actually
02:08cautioned that some of this can become
02:10sort of innovation theater and I'm
02:13between how do you know it's just not
02:16innovation theater and it's gonna lead
02:18to something fruitful it's not just like
02:19another failed siloed function it's a
02:23very good question in what we are seeing
02:26is that a lot more companies want to
02:28establish their presence
02:30and most of them interpret this presence
02:33with the establishment of a venture
02:36effort but that does not make them to be
02:41truly embeddable into the into the
02:45valleys ecosystem and put them in a
02:48position to really take advantage of
02:50everything that this area has to offer
02:54and by extension other ecosystems around
02:58startup ecosystems around the world so
03:01what we try to caution is just because
03:04you're sending some executives here as a
03:08corporation it doesn't mean that you're
03:10automatically positioning yourself to to
03:13take advantage of the power of Silicon
03:17Valley and maybe then more importantly
03:19just because you're establishing a fund
03:22it doesn't mean that now you're going to
03:25start seeing a tremendous deal flow and
03:27the right deal flow and participate
03:30actively in obviously benefit the the
03:34the corporation that you represent
03:36you'll start seeing companies but they
03:38may not be the right companies and they
03:40may not be the companies that actually
03:41come around later and and take you on as
03:44a corporate entity right that's right I
03:46mean it is very interesting to me when I
03:50meet a corporate delegation that has
03:52recently established itself here and
03:56they will tell me how they participate
03:58in a couple of meetings and they
03:59interpret that is now being part of the
04:02part of the network and a little bit
04:04later they will ask exactly the question
04:07that you asked why aren't we getting
04:09into the right deals why can't we talk
04:12and why can't we syndicate with a the
04:16right VC is that what what's the answer
04:18to that why can't they well first of all
04:23there is the there is the aspect of
04:25trust you need to develop but more
04:28importantly I think you need to as a
04:30corporation you need to demonstrate that
04:33your what you're trying to accomplish
04:34just because you have a venture fund and
04:37just because you're willing to spend
04:39some some money doesn't mean that you
04:42become a value-add partner of a start-up
04:45or of a group of startups so what we
04:48have found is that the ones that have a
04:51particular reason they're trying to
04:54solve a particular problem they have
04:56engagement from the highest levels of
04:59the corporation is not only anointment
05:02but there is actually engagement on a
05:04continuous basis they're the ones who
05:07are putting the right foundation to be
05:10able to truly take advantage of of the
05:13start of a startup ecosystem so that's
05:16some of the things that reveal what
05:17would make an innovation outpost work a
05:20broader question I have is just how big
05:22companies in general operate compared to
05:24startups and NV C's as well if we're
05:27talking about this innovation ecosystem
05:28and one thing that I found recently
05:30which I include in her in our newsletter
05:32is a finding that decision-making and
05:36companies has actually significantly
05:37slowed down I thought that was
05:39fascinating because one of the
05:40observations that you guys have shared
05:42is that sometimes the speed of response
05:44is even more important than the response
05:47itself and I won't even share some more
05:49insight about why that is well one of
05:52the observations we we made is that you
05:55need to understand what the corporation
05:57needs to understand why they want to
05:59have an outpost most of the times there
06:03is some sense of sensing being able to
06:07sense what is happening in a particular
06:09ecosystem so that they can they can
06:11respond it's kind of like an earthquake
06:13like very much so I mean I mean Caltech
06:19which is world class Institute on earth
06:21quake analysis geophysical analysis they
06:24they ran an entire network of sensors up
06:27and down the west coast right you should
06:30be thinking as your outpost as is
06:33providing you without sensing ability
06:36early warning system or something not
06:38the day before yeah again some
06:41corporations or most corporations feel
06:44that just because they establish a
06:46presence in Silicon Valley they're done
06:50right and one of the advice that we
06:52provide is that first determine which
06:56are the ecosystems that where relevant
07:00work for the problems that you're trying
07:03to address is being done so if relevant
07:07work is being done in Silicon Valley and
07:10in Israel and Berlin well you should
07:12consider having presence in each one of
07:15these ecosystems and maybe the presence
07:18is not of the same size but nonetheless
07:22it every one of these presences provides
07:27you with signal and then based on that
07:29sensing you can determine what would be
07:32is going to be the nature of your
07:34response so we have identified five
07:37different ways of responding to you can
07:39invest you can invent you can acquire
07:43you can partner and you can incubate and
07:47determining where which of these types
07:51of response is the right one to start
07:54with and the right one for the
07:56particular ecosystem that you've chosen
07:59to to be present in it's actually a hard
08:03problem most of the times when I when
08:06I'm asked to advise corporations I am I
08:09come in the middle of the movie as I
08:11call it because they have already taken
08:12an action 90% of the times is to
08:16establish a venture group and in
08:19retrospect we determined that maybe
08:21establishing a venture group wouldn't
08:23have been the right first step maybe
08:25they should have established a business
08:28development capability before they even
08:31think about establishing a venture group
08:33there were is this wave of corporate
08:36venture in the late 90s and it ended
08:39badly for the most part now we're back
08:43to 2016 or you know last year for that
08:45matter and they understand differently
08:47about what technology means for their
08:50sounds like if they're just coming back
08:51you know with the a similar approach of
08:53like well we'll just set up a venture
08:55arm it's not gonna work this time either
08:58so what do they understand differently
09:00how is it different and are they here to
09:03stay well so lately start with some
09:05numbers so back in in the late 90s
09:09roughly between 1998 and 2000 there were
09:15roughly 500 corporations that had
09:18established venture groups in our area
09:21and of those 500 by 2003 50% had closed
09:27down now by I call them unofficial
09:31statistics there are roughly a little
09:33over a thousand corporations which 10
09:37cents makes sense because many more
09:39geographies have opened up sin in the in
09:42the intervening you know 10 15 years and
09:44I wouldn't be surprised if we have the
09:48same drop-off rate when the next
09:50recession arrives and where the next
09:52crisis or whatever arrives now there is
09:56the one important difference backing
09:58back in 98 99 I feel that corporations
10:04here approach the internet broadly as a
10:09communication medium communication
10:11collaboration this time around they not
10:16only they are seeing the broader impact
10:19of digital technologies but more
10:23importantly there are several
10:26fundamental technologies beyond
10:29information technology that are hitting
10:32them at the same time so I was working
10:34earlier in 2015 with a foreign
10:38corporation very large corporation and
10:42their spaces materials but they were
10:45here to understand software and big data
10:48then by extension Internet of Things and
10:51by extension software as a service
10:54models is it because they wanted to
10:56become a software company I mean there
10:58are material science company or because
10:59they feel like software might really
11:03to their side of the business somehow no
11:06because actually they were seeing that
11:07software and data and 3d printing had
11:12already gotten into their business and
11:14so they were not starting being ahead of
11:17the ball they were they knew that they
11:19were starting from behind but but my
11:22point is that that they they had already
11:24realized that there are they're being
11:26hit or they've been threatened maybe
11:29from a variety of perspectives and this
11:32is not about just learning about one
11:35technology checking it and moving on
11:37with life in a business-as-usual way so
11:40now they're seeing that there there are
11:42many more things that are many more
11:44technologies many more business models
11:47right like software-as-a-service is
11:48actually a completely different mindset
11:50precisely that is that again and by the
11:52way they're customers now are asking
11:55them to start using this type of of
11:58models as opposed to them thinking oh a
12:01start-up is using it now let me think
12:03about whether it applies to my business
12:06as well I mean sometimes their customers
12:09are ahead of them and they have to react
12:11that's actually that place to be though
12:12and you're gonna do that that direction
12:14versus you know leading your customers
12:15what you want them to go I've been I've
12:17been in this I've been in the space and
12:19this industry long enough to have seen
12:22all of these ups and downs and what is
12:24fascinating to me today without trying
12:27to create any any hype is that what we
12:30are facing what we're going through
12:32right now is is truly a technological
12:36revolution not unlike what happened back
12:41in the steam age or the Indus River I
12:46mean that magnitude you know the the
12:49first internet wave was also very
12:52important was mostly a single single
12:55thing here we were being impacted by
12:58many more technologies many of which are
13:01acting as building blocks many of which
13:04weave together with business models so
13:07the the level of complexity that
13:09corporate executives
13:12have to content with right now is
13:15unprecedented and and their pace of work
13:18is does not prepare them very well for
13:22having to make this multi-dimensional
13:23decision right nor does the structure
13:26because first of all all these things
13:27are coming together as building blocks
13:28that might interact and completely
13:30unexpected ways so there's that
13:31interaction effect but there's also this
13:33thing where one of the PCs that we've
13:35put forth is that it's actually very
13:38software-as-a-service company to
13:40actually become a software as a surface
13:41company because it requires completely
13:44reaaargh attacking everything from the
13:46ground up from your sales organization
13:47to the way your departmentalized to just
13:50about everything customer service like
13:52sales the mindsets compensation right
13:54exactly and so incentives right and so
13:57the question that I have is when you
13:58describe this as a big structural shift
14:00what are the ways that over the past
14:02hundred years to now that companies have
14:05had to change their organizational
14:07structure and where I mean where might
14:09we go next but how is the nature of the
14:10firm being redefined I always love this
14:12topic so we have some hypotheses and and
14:15one of the major hypotheses that we have
14:18is that the innovation outpost if done
14:22correctly can become a major change
14:26agent for the corporate structure how
14:31this type of organizations can start
14:34impacting the way you do research the
14:37way that you are changing from being in
14:41or eyes on one corporation to being an
14:45arising to and arrives on three
14:47corporation how you need to to change
14:51your core without throwing away all your
14:55DNA I mean so these are very very
14:58important and very critical problems for
15:02corporations to start thinking about we
15:05do not believe that everybody would be
15:07able to make that leap and frankly some
15:11of the statistics show that the the life
15:13of the corporation of corporation is
15:16shrinking right you have companies that
15:18are that remain independent or for for
15:22shorter time yeah the span of the
15:24number of companies in the sp500 or that
15:26their longevity is shrinking quickly I
15:29mean the company you worked at was as
15:30pod on the very few hundred-year-old
15:32companies left IBM and there might be
15:34very few of them left it's true and
15:36actually it begs a question because it's
15:38almost like this assumption that a
15:39company should have a long life span and
15:41it's actually like well do we really
15:42want that to be the case companies are
15:44no longer the containers of security and
15:47people can actually move around and you
15:48see it I mean first of all you you see
15:52it in the tenure of CEOs which is
15:55shrinking significantly but just what
15:58you what you mentioned I mean I you you
16:00see it in the attitudes of employees
16:02where the the aspiration of working for
16:08the same corporation for 30 years is no
16:12you I won't call it we have moved to the
16:15gig economy but we're moving in that
16:17direction sometimes by choice
16:19sometimes by necessity but but I think
16:22that these are all trends that that are
16:25going to to impact how the corporation
16:29looks and which corporation survives and
16:32and the like does the outpost ever have
16:36to become the center can it even become
16:38the center because I just don't think
16:40there's ever been a case where that's
16:41happened so I don't think it becomes the
16:44center but but it becomes the
16:46transformation agent the whole idea is
16:49that corporations do not experiment
16:52enough in our view what can a
16:54corporation take away from Silicon
16:58Valley or from any other innovation
17:00ecosystem is not the fact that you come
17:03into touch with two or three or ten
17:05startups is the fact that there is a set
17:09of now best practices and one of these
17:12that you can take away best practice of
17:14how to create startups how to finance
17:19startups how to manage a portfolio of
17:22startups and what's important about this
17:24is that the corporation's need to find
17:27ways to experiment rapidly and
17:32and and what we have learned here to do
17:35very well may be better than any other
17:38ecosystem around the world is how to
17:41quickly start companies to get them
17:44financed to allow them to pivot as
17:46necessary how to back our winners and
17:50let our not so strong players move on
17:55and I think the corporation needs needs
17:59that type of a playbook so the the
18:03innovation outpost becomes that change
18:06agent for launching this experiments I
18:09think a number of these experiments
18:11become very successful so essentially
18:15you need to find ways and what we are
18:18advocating is how do you change the core
18:20are there specific examples of that that
18:22you can point to that you've either
18:24observed or if you can worked with for
18:28me a very interesting example is what
18:30Google is doing with alphabet and the
18:33reason I find this interesting is
18:35because you're seeing their venture
18:38capital groups working in tandem with
18:42their the rest of the corporation
18:45you see them sometimes acquiring a
18:48company and letting it develop before
18:51bringing it into the core you're seeing
18:55them merge things like you saw for
18:57example what they did with their AI
18:59efforts you see how they deal with moon
19:01shots like the Google car I'm not saying
19:04that they always get it right but there
19:06are some interesting lessons there of
19:08how do you continue evolving the
19:12another example is BMWs I brand an
19:16organization that has been set up very
19:18differently from the major business
19:22units of BMW trying to to look at not
19:27only the manufacturing aspect the design
19:30and manufacturing aspect of the car
19:32which is what BMW and other automakers
19:34are known for but the the whole aspect
19:38of mobility services right how do i how
19:42do i service a consumer
19:45beyond just owning a car and how do I
19:48service that consumer throughout the
19:51lifecycle of the hoof a transportation
19:54experience not only when that person
19:56goes to a dealer so in there are other
19:59there are other examples but these to me
20:01are some more recent experiments that
20:05point in that direction you had an
20:08article out with another co-author about
20:10evaluating moonshots and Harvard
20:11Business Review can you tell us what
20:13that framework looks like the moonshot
20:15is a major effort by a corporation with
20:20the idea of addressing a big or a big
20:24problem so in the case of IBM which was
20:28the company that we evaluated through
20:30our framework they are developing Watson
20:34as a new type of software platform which
20:38uses artificial intelligence and can be
20:42used as there is the building block for
20:45building for creating the next
20:46generation of applications whether
20:49they're enterprise applications or
20:51consumer applications and they're
20:52investing significant sums of money in
20:54order to see them through they need to
20:57consider a variety of dimensions and one
21:01of the dimension that that we talked
21:03about is experimenting with business
21:06models before they they identify the the
21:10right one through this experimentation
21:14analyze their failures as much as they
21:17analyze their successes because they can
21:20learn a lot from these failures and one
21:23of the issues that many corporations
21:25have and their employees is that they
21:28become a risk averse because they do not
21:31want to fail and when your our point is
21:34that when you are involved in moonshots
21:37not everything is going to go as planned
21:39and if it goes as planned it means that
21:41you didn't take enough risks so they
21:44really need to impart that type of
21:47approach to their employees is the
21:49moonshot is it a quantitative thing or a
21:52qualitative thing which is to say that
21:54you know what's a moonshot for Google
21:56may not be a moonshot for a century-old
21:59or something like that well I mean for
22:01us the the notion of the moon shot was
22:05more defined an effort to address
22:08something new by making a very
22:11significant investment so just to put
22:14into perspective if I'm a large
22:16corporation and through my venture arm I
22:18invest in a in a small start-up maybe
22:21I'll invest five ten fifteen million
22:23dollars publicly iBM has said that their
22:27the Watson moonshot is over a billion
22:29dollars I would imagine the Google car
22:33is similar if not larger investment we
22:38talked about the BMW I brand that's a
22:41multi-billion dollar yes so the the way
22:44you differentiate is not only by the
22:46fact that you're trying to do something
22:48new but also because you're investing a
22:52significant sum of money in India you're
22:54solving a very complex problem again
22:57what iBM is trying to do is trying to
23:01make artificial intelligence extremely
23:04easy by encapsulating it in this
23:08platform so that any application
23:11developer through the up through a
23:13series of api's now can get access to
23:17all of that intelligence I mean and and
23:20they're one of the major application
23:22areas they have is in the healthcare
23:25sector so you can imagine what they want
23:29to imagine is doctors you know naturally
23:32interacting with applications that are
23:35based on Watson in order to provide
23:38advice in order to make difficult
23:40decisions so again this is this is a
23:43very hard problem and that's why I said
23:46you need ste the proper experimentation
23:49the proper patience the proper personnel
23:51fact I mean hiring is a is another big
23:53dimension who do you hire
23:55so evaluating or defining the moonshot
23:57it's something new it's a hard problem
23:59it's got significant investment behind
24:01it the right kind of hiring Astro teller
24:04made an argument in an op-ed for me a
24:06few years ago that it has to be 10x
24:07better it can't be just you know 10 like
24:10has to be not just incrementally better
24:12of magnitude better he also made the
24:15argument interestingly that there has to
24:16be some sign today that you can actually
24:19get there so it's not just so pie in the
24:21sky like human beings need to fly
24:23without wings but more that there's some
24:25indicator whether it's Moore's laws on a
24:28trajectory or a curve that there is some
24:30indication that we can actually get
24:32there even if it's not available right
24:33now a question I have for you though
24:35Evangelos is Watson is an interesting
24:37example in multiple fronts but I can't
24:40help but wonder how can it compete with
24:42what a new class of startups can do and
24:46when you have other companies that have
24:48a whole different type of data that IBM
24:51does not have access to like Google and
24:53Facebook for example and what they're
24:54doing and that's just the example of AI
24:57it this plays out in countless domains
24:59where can a big company actually compete
25:03with startup driven innovation so I
25:05would say the short answer is yes
25:06because when we talk about start-up
25:09driven innovation we we think of it as
25:13the corporation adopting many of the
25:17practices that make startups successful
25:20one of these practices significant
25:22partnering and knowing who to partner
25:25with and how to develop these
25:27partnerships efficiently so to us I mean
25:31to brought power my co-author in bhpr
25:35article I mean this was one of the the
25:38most successful aspects that IBM has has
25:42done and we encourage them to continue
25:44to continue doing it because I think
25:47corporations are starting to understand
25:50that in this environment regardless of
25:53how much capital they want to throw with
25:55the problem and how big and how good
25:58they are they cannot solve things on
26:01their own and they need to go innovate
26:04Steve and I are finding that
26:06corporations that have adopted open
26:10innovation principles where they are Co
26:13innovating with other large corporations
26:15are becoming more successful at adopting
26:20startup driven innovation I think that's
26:22actually a key point because look the
26:24common problem with absorbing and
26:26outside innovation whether it's from an
26:28outside R&D lab an internal captive R&D
26:31lab that's just headquartered somewhere
26:32else a start-up or partnering with
26:34another company the main problem is NIH
26:38they're not invented here syndrome and
26:39how do you then get over that so there
26:41is certainly an advantage in shifting
26:43the mindset of the company and its
26:45practices so it can actually absorb
26:47what's new and the time comes it just
26:50still seems though that there's no way a
26:53company can overcome its core incentive
26:56to make money from its core business it
27:00sounds like you might be arguing for you
27:02know the sort of the natural atomic unit
27:04of innovation is the startup and once it
27:07reaches a certain age or a certain size
27:10right so I mean honestly Google
27:11struggles with this which is why we see
27:13alphabet now but how do you I mean in
27:16some ways it makes me think of like
27:17discharge but we're desiring use like
27:19you know a company just want the youth
27:21that these startups have but is there
27:23something more sort of engineered and
27:26cultural that they can wrap their arms
27:27around that they can bring in does
27:29Facebook fail at some point for lack of
27:31being a start-up anymore Google same
27:35thing Facebook's also experimenting the
27:36way Google is in the sense that when
27:38they acquired whatsapp they made sure
27:40that Jan koum had a seat on the board do
27:42you think it's impossible for companies
27:44who didn't come up like Google and
27:46Facebook and others did the short answer
27:48is I do not know what we are saying is
27:51that in this environment that is being
27:54set up with these forces that are at
27:57play right now the more traditional ways
28:00that corporations have had to innovate
28:04need to be rethought right okay we
28:07believe that the startup playbooks or
28:11best practices that we've learned from
28:13startups and their entire environment
28:16have an important role to play in this
28:20rethinking will everybody get it the
28:23same way right No will everybody do it
28:26no but there's still something too
28:29significant to consider especially for
28:33for corporations that are thinking how
28:37there you know horizon 2 in your eyes on
28:40three I think corporations because of
28:43their of their scale issue because of
28:46the sometimes because of the short
28:49termism that they they operate in
28:52whether it is because of activist
28:54investors whether it is because of their
28:58own their own doing they have become
29:02very good and have optimized how to deal
29:05in their with their operations in the
29:08short term and because of that more so
29:12than in the past they are facing risks
29:16on how to deal with a longer term in
29:19this risk as I said it's not only
29:21because of what they're doing today but
29:23also because of all of this
29:25technological revolution that is
29:27happening so they're being hit from
29:28asset from variety of directions to
29:32address their problems they cannot just
29:34do one thing okay they need to be doing
29:37several things we talk about what
29:39corporations can learn from Silicon
29:40Valley but we run the risk of the sort
29:42of the outpost becoming an outpost right
29:45so what can Silicon Valley learned from
29:46Silicon Valley and how do we ensure that
29:49what's here and what's so valuable
29:53continues many times whether it is us as
29:57investors or entrepreneurs or or startup
30:02executives we do not give the proper
30:06credit to the corporation we are not
30:10seeking to have a value exchange is this
30:14give-and-take in one of the the recent
30:17examples that brought this to bear for
30:20me is the automotive industry I've been
30:23writing a bank have been working with
30:24automakers and have been writing about
30:26the disruption and the challenges that
30:30the incumbent automotive industry has
30:34there is also a lot of enthusiasm in the
30:37valley from our own companies that we
30:41can completely overtake the automotive
30:44industry and I think this is a fallacy
30:46completely obliterating it and making
30:51you know the next Detroit it's hard for
30:54me to to imagine that maybe I'm too
30:56short term thinker myself but I think
30:59this is this provides a great
31:00opportunity where the local automotive
31:04ecosystem can really start collaborating
31:07with the incumbent automotive ecosystem
31:11and build something great right so
31:13that's how I believe Silicon Valley and
31:17any other ecosystems that Apeco system
31:19avoids being disrupted itself that was
31:23all we have time for
31:24Evangelos thank you so much for joining
31:26the a 6nz podcast thank you very much
31:28we'll see you at the outpost