00:00this is Scott Cooper I'm here with
00:01Gaurav Dillon the founder and CEO of
00:03snap logic and thank you Cora for
00:05joining us pleasure to be here Scott
00:07we're gonna cover a bunch of topics and
00:09keep this as a very free ranging
00:10discussion but where I want to start is
00:12I want to start with this thing that we
00:14used to call EA I enterprise application
00:17integration and the reason I bring this
00:18up is Gaurav was the founder many years
00:21ago of a company called informatica and
00:23what would be great I think except the
00:24Ark for was if you think about where we
00:26were in you know 96 97 4 2000 around the
00:29time that you guys were coming to life
00:31what was kind of the enterprise
00:32computing challenge there what were the
00:34platform issues if you drilled back in
00:36that time you have to look at the macro
00:37trends what was going on was the year
00:392000 was approaching and people were
00:42trying to business process reengineering
00:44they just couldn't do it fast enough so
00:46this was a golden age of companies like
00:48sa P companies that no longer exist now
00:51like C bowl and PeopleSoft and there was
00:53a tremendous time for all of them but
00:55what needed to happen as that was going
00:57through his people needed to now bring
00:59those things together but it was a sort
01:00of very much that you were bringing in
01:03that product and you were helping to
01:07replace the old mainframe stuff with the
01:09new business process re-engineering our
01:12P stuff and when you say people need to
01:14bring that stuff in do you mean data so
01:16it was there information sitting in
01:18those applications that people needed to
01:20kind of be able to talk to one another
01:21as part of this range here and what was
01:22the nature of the IT problem people were
01:24trying to solve at that time it was more
01:26of a replacement of core applications
01:29like finance right it's open heart
01:31surgery for the enterprise to replace
01:32finance it's not simple it's in many
01:34cases you have to report earnings
01:35quarterly so you really have to get this
01:37just window just right so so the tasks
01:40that we built were how to get these big
01:43big applications to work with some of
01:46the other applications that existed in
01:47the enterprise but it's very much
01:49subservient to the big applications
01:52right if you bought something say from
01:54Oracle somebody in Redwood Shores
01:56California designed a way in which you
01:59balanced your books and that's what you
02:01did now you may need lots of other
02:02inputs and things to make it work you
02:04may need a data warehouse to report on
02:06it so you had certain amount of freedoms
02:08but the freedoms were very much
02:09subservient I would say to what that big
02:13was going to do so the business the
02:15business process itself was dictated by
02:17the key application been precisely and
02:18you know you were and people often used
02:20to use this term you know but it was a
02:22derogatory term or not but you were you
02:23were middleware to a certain extent
02:25right it was kind of you were connecting
02:27applications but with the kind of real
02:28business logic and business process
02:30being dictated by the SI peas and
02:31Oracle's and percent evil sauce and
02:33Siebels the world the other aspect of
02:34all this was that it was still a very
02:36limited audience of usage right
02:38remember this is all before the World
02:39Wide Web boy it was even before the web
02:45right this is the mid to late 90s so we
02:48things have changed quite a bit right to
02:50drill down then so for all the users of
02:51these applications they tended to be
02:53fairly specialized business users right
02:55there wasn't a desktop on every desk and
02:57you know people using email and all
02:59these other things these were kind of
03:00highly specialized and highly trained
03:01people so it was a lot of one-to-one
03:03connection of applications and a very
03:04risk framework and and so if you fast
03:06forward to kind of a little bit to today
03:08or at least we pick up you know maybe in
03:09the modern web era yeah what's changed
03:11from an architecture perspective what's
03:14changed from the types of applications
03:16the types of data kind of where these
03:17things reside like what's the big
03:18overarching architectural shift that's
03:20happened since late 90s to get us to
03:22today that even you know requires the
03:24need for a company like a snap logic or
03:27others that are doing things you know in
03:28a different fashion than what
03:29informatica was doing this may seem
03:30radical but besides the problem
03:33everything has changed okay it's the
03:35same problem and this is Robin being
03:37what applications need to talk to one
03:39indeed and how does somebody make a
03:42complex enterprise run because it is you
03:45know it's not just eight arms of an
03:47octopus there's 1,800 connections in
03:49these companies right so you've got a
03:51very sophisticated clockworks
03:52to make a large company function so the
03:55problem is still the same but everything
03:56has changed so let me take you to some
03:58of the changes one is you have of course
04:01the web what does this do is you're no
04:03longer building installing applications
04:05you are essentially using websites you
04:08don't just use the Internet to buy your
04:11books you use the Internet to balance
04:12your books if you were workday
04:13financials user you have a browser up
04:15boom if you're Salesforce user both so I
04:17think that's a profound change because
04:19the data types are now different it's
04:21not just the world of rows and columns
04:22that it was in the 90s it's the world of
04:25web data if you ever sort of
04:27accidentally open up your browser into
04:29its squiggly you know rows and column
04:31into squiggly brackets you've seen a
04:34JSON object you know and that's the
04:36stereotypes are different the network
04:37topology is different your data is not
04:39it's there so I think that's probably
04:42the most profound one now with that
04:44though is a millennial post web
04:48generation where you have self service
04:51as an expected requirement if you ever
04:54had the pleasure of installing Siebel
04:55sales force automation it came with
04:57about five people whose job it was to do
04:59reporting for the head of sales right
05:01you couldn't just do it yourself
05:02the fact that you can actually build
05:04sales reports on the fly in a meeting is
05:06astounding and this is where I think
05:08Salesforce has pioneered much of the no
05:11software wave in the enterprise and then
05:13there's some amazing work being done by
05:14folks like ServiceNow and work day now
05:16so the self-service model has
05:19fundamentally altered the landscape and
05:21then you layer in Millennials who are
05:23friendly to computers and iPads and so
05:26on and everything is different today so
05:30one other thing that I also at least
05:31I've always thought is different is
05:33proliferation of applications in the
05:34enterprise so yes I agree we've got new
05:36data types we have new data sources
05:38right they're now living and residing in
05:40other places one of the things that
05:41we've often talked about here is the
05:42idea that with the development of SAS
05:44and web-based applications what used to
05:46be the governor on the number of
05:47applications and enterprise can do was
05:49how many applications could the actual
05:51centralized CIO actually manage and
05:53implement and support over time right
05:55and we've essentially eliminated that
05:56governor now as you think about
05:58integration of data and other things
05:59across applications I assume the sheer
06:02proliferation of applications and
06:04business level users as opposed to
06:05purely kind of these centralized IT
06:07level users probably also you know kind
06:10of orders of magnitude changes the
06:11pretty big but that's not surprising
06:12right you expect with the web that
06:14people would use websites and so on
06:16what's crazy big is that people think
06:19they don't have a lot of applications if
06:21you bet companies a dollar that they're
06:22using X number of SAS apps they're off
06:25but not by half they're off by like 10x
06:28you know cuz somebody marketing is using
06:31something and they go it's not an
06:33application it's just a website and you
06:35go wait a minute but you're using it to
06:38do a critical task and you're doing this
06:41that is important to the companies a lot
06:43of data flowing from it back and forth
06:44it has all the implications in how it is
06:47part of your enterprise fabric as if it
06:50were an application is just a website so
06:52the sheer number of applications that or
06:56websites in the enterprise now is at
06:58least NX in my opinion but she was
07:00shocking is people think it's do X right
07:04they've layered on all these new
07:05applications many of which are now
07:07web-based applications at the same time
07:08probably we haven't retired many of
07:10those older applications right right and
07:12that is different right so Finance is
07:14let's say even the biggest companies you
07:17can think of the web monsters they run
07:19on good old financial systems that might
07:21serve so that is not changing as much
07:24all that would work their financials
07:26we've seen some some proliferation but
07:28everything else going going gone
07:30certainly Salesforce automation or CRM
07:33certainly helpdesk management certainly
07:35human capital management and and if you
07:38go by job title and vertical in every
07:40industry all those squares are lit up
07:43with SAS applications so given those
07:45changes and look obviously we're now
07:46many years into it but what's the real
07:48nature of the problem we understand
07:49those architectural changes what is it
07:51that makes the old way of doing things
07:53no longer applicable given those
07:54architectural changes so I would say it
07:56probably two things one is the data
07:59fundamentally being different require
08:01new kinds of plumbing you know this is
08:03digital plumbing we're talking about but
08:05you're no longer using rows and columns
08:07you're using a document model the way
08:10the worldwide web works the way browser
08:11talks to a website is the way these
08:13applications all function so I think
08:15that's a huge reason to rethink the core
08:18engineering that you might be doing
08:19because if you use stuff from the last
08:21century and look it's good stuff but
08:22it's designed for a certain time at a
08:24certain point in time and where we are
08:27today is the data types fundamentally
08:28being different and say let's think a a
08:30very very real catalyst of new kinds of
08:32technology but sometimes we find that
08:34the even more important one is self
08:36service C integration has always been in
08:40the basement in the dungeon in the
08:43enterprise and I glow-in-the-dark people
08:45who come in and do integrations and if
08:47they work that's beautiful if they don't
08:49you don't know why and if it start
08:51working again you don't know why so
08:52having the ability today
08:54that in a sense from the darkness into
08:56the light with technology that is
08:58self-service that works like SAS is the
09:01next requirement and it that is becoming
09:03an absolute requirement it's rather that
09:05people are shocked that that's not how
09:07things work right and I think there are
09:09then a way of thinking about the radical
09:11benefits of going from manual labor as a
09:14word to having business people get the
09:16stuff when they need it whenever they
09:18want it on their own terms and be able
09:20to sell serve themselves okay so those
09:22two things are I think is sort of what's
09:26what's causing this to be a huge problem
09:28that makes a lot of sense so we've been
09:29talking about aggregation of data and
09:31there seems to be no end to it I mean
09:33what what happens over time where does
09:35all this data go what happens in the
09:36enterprise there's a software life cycle
09:39in the enterprise people have thought
09:41about how they staged data how the stage
09:43applications and make them happen I
09:44think you look at a lifecycle for data
09:46as well you know there's all this
09:48production of data that is happening at
09:50the edges and the core at sensors around
09:53all the monitoring that we do around all
09:55the new requirements that we have in
09:57somewhat of a scary world and so there's
10:00a life cycle of more data production
10:01more data management and actually more
10:04consumption and it's it's a virtuous
10:05cycle it feeds on itself because when
10:07people see the benefits of analytics
10:08they want to go capture more so it's a
10:10healthy trend I want to shift gears a
10:12little bit we've been talking about you
10:14know movement of data right and
10:16integration of data across applications
10:17but I'd love to kind of shift the
10:19conversation a little bit to analytics
10:20what's the implication now what is the
10:22role of analytics what does it mean in
10:24terms of the utility to the business
10:25user so I think historically we had
10:28reporting and business intelligence we
10:30may have spoken about it as analytics
10:32towards the turn of the century but that
10:33was always reporting and if you think
10:35about reporting and business
10:36intelligence today that's what my iPhone
10:38does if I go for a hike I pull up my
10:40iPhone it tells me how many steps I've
10:42taken right so this rearview mirror
10:43historical perspective is it is no
10:46longer of incremental value to a
10:49technology company or do an investment
10:51bank you know Jamie Dimon once famously
10:53declared that I'm not a bank I'm a
10:55technology company a JPMorgan Chase so
10:57as people go through this soft rates the
10:59world phenomenon that there is a change
11:02in expecting new things so what are
11:04those things those are things that you
11:06find at Fang Facebook Apple
11:08Netflix Google etc and those are
11:10predictive forward-thinking analytics
11:12those are discovery engines
11:13recommendation engines machine learning
11:16artificial intelligence algorithms
11:17although I think AI is often frequently
11:20misused I think missile machine learning
11:22a police thinks and what those are are
11:24predictive types of things what is
11:26likely to happen so what we're seeing is
11:28a trend away from legacy data warehouses
11:31into data lakes which are then consumed
11:35both by people using modern
11:37visualization products like say a
11:39tableau and also by lots and lots of
11:42data scientists which is a new job title
11:44where you rub the data with the
11:47algorithms and you try to predict what
11:49will happen so if we stay on financial
11:51sector just for a moment what's an
11:53example is if we want to take a hunch
11:55where you know we manage some money and
11:59we want to take a hunch on the price of
12:00gasoline in Houston Texas for example on
12:02Labor Day weekend what would we do well
12:05we would reach out for historical prices
12:06of gasoline we would reach for other
12:08commodity data and we would try to have
12:09predictive algorithms go in a sense
12:12model fit it's like girl fitting but
12:15that they're far greater degree but do a
12:16broader audience it's on a more massive
12:18scale and much more profound set of
12:20results that emerge and for that you
12:22continuously need to dip into the data
12:24lake bring forth data and you need to
12:26iterate on the algorithms so so this is
12:29what is happening and this is causing a
12:31profound change in how people are
12:33thinking about analytics today
12:36now you'd look at the work of predictive
12:38analytics and you divide that into two
12:40halves there's a data scientist this is
12:42a person she's an expert in algorithms
12:45they might have emerged with a computer
12:47science degree or a mathematics degree
12:49some some cases economics degrees and
12:51you might have even called it statistics
13:05the word was statistics once upon a time
13:09your fitting is this predictive models
13:11and that's where the science comes from
13:13but there are some modern things to it
13:15that didn't exist because we didn't have
13:16the compute right right so all these
13:18decades of Moore's law I have gone
13:20have kind of made us think about
13:21algorithms that we couldn't even imagine
13:23that would that are sort of more
13:24heuristic in nature more bear matching
13:27in a sense that just regression alone
13:29doesn't do all those data scientists and
13:31people who are doing predictive
13:32analytics need to iterate the data with
13:34the algorithms so you divide up that
13:36task into a on the one hand a data
13:38scientist who is talking to the business
13:40person trying to figure out what might
13:42happen the price of gasoline in Houston
13:44Texas but that person needs to go either
13:46themselves or through help get data so
13:49there's data engineering task that has
13:51to go on capital one who is I would say
13:54the original data science company
13:56figured out tens of billions dollars of
13:57business giving credit cards to people
13:59who others had denied and making it
14:01profitable using data science you know
14:04and for them this is absolutely quarter
14:07their business and of course we know
14:08that the big web monsters use it but you
14:11also see it used on and it traditionally
14:13has been a big deal in Wall Street but
14:14now it's expected and and this is a
14:16really mushrooming because of the
14:20increased computational and the
14:22plummeting cost of hardware storage so
14:25when you put those two together you
14:27really get fired you've got the
14:28ingredients of combustion there so you
14:30use two words data warehouse the old you
14:34use this exciting new term data Lake
14:36that's right so I just let's kind of
14:38unpack those a little bit you know what
14:39was the problem again five ten years ago
14:41when people were doing data warehouses
14:43what were they trying to solve what were
14:45the limitations that made things like
14:46predictive analytics hard and I how did
14:48we get to this transition to this
14:49concept of data Lake so I think first of
14:51all I would put my hat to do visionaries
14:54and thought leaders in the industry
14:56bill Inman and Ralph Gimple they came up
14:59for the concept in different
15:02perspectives and they're there is
15:04actually a religious debate on forums on
15:06the Internet to this day about which one
15:08is better we stay out of it we're like
15:09look it's data how do you want it so so
15:13but they really pioneered in different
15:14styles a way of being able to deal with
15:17time series data along multiple
15:20dimensions right so for example if
15:22you're in the retail business you have a
15:24product how has it done in regions over
15:28time the most common three dimensions of
15:30everything in the in in revenue how are
15:34my products doing etc etc so
15:36historically data warehouse went through
15:38this real sonic boom in the 90s in the
15:41early part of 2000 because it was a
15:43resonance with the change from the
15:46mainframe to ERP so when he brought in
15:48ERP you're like oops we forgot about
15:50reporting what do we do now so then you
15:52have this second sonic boom in the world
15:54of business intelligence and ETL and
15:56data integration with informatica do
15:58bring that forward and that was then so
16:00that was what that term is and look
16:02they're still running and lots and lots
16:03of places and you get a good historical
16:06perspective but that is no longer enough
16:08you know that is they particularly in
16:10this millennial generation and
16:12particularly in areas like marketing and
16:14other places that is no longer enough
16:16and which is why you're seeing this
16:18fever about AI breakup so when we go
16:21from data warehouse to do this concept
16:22of data Lake what's the difference
16:24either architectural II or what what are
16:26the use cases one of the things that
16:28this new concept today like enabled so
16:30Inman and Kimball came up with an
16:32organizing principle you know data
16:34warehousing was never a product right
16:36you couldn't go and buy one it didn't
16:38exist but it was an organizing principle
16:40to end to corral Marshall and get
16:43benefits from the data that you had in
16:45your enterprise a data Lake is at an
16:48earlier stage of development you know in
16:50the packaged furniture industry for
16:52example initially you go and get planks
16:55of wood you saw them and you make you
16:57know furniture and then somebody says oh
16:58we'll have furniture pre-made and you
17:00can paint it and then someone comes
17:02along and they prefab it you know
17:04Restoration Hardware Bed Bath & Beyond
17:06and IKEA and you buy the stuff and like
17:09my eight-year-old can assemble it with
17:10the screwdriver right so we're going
17:12through an evolution rapidly towards
17:14people thinking about a data Lake as a
17:16place where you put all the data you can
17:18afford to keep modulo ie without
17:21duplication all right so get the data if
17:23you can afford it keep right because it
17:25may not have signal today it may be
17:27noise but as the algorithms improve and
17:30as computers get faster which we know
17:32happens every 18 months they double
17:33you're going to be able to get signal
17:35out of the noise right so so did the
17:37data leak fundamentally is about more
17:40data of more types of data it is
17:42expected that you will have all sorts of
17:46Oh Lord JSON web exhaust machine data
17:48right you know the industrial internet
17:50where you have exhaust remember the data
17:52warehouse was born in an age where most
17:54of the data was the barcode scan went to
17:57the grocery store you bought shampoo or
17:59beer and you scanned it and initially it
18:02was done just to speed up people the
18:04checkout line and sell more but the
18:06actually the data warehouse industry
18:08dwarfed the supply chain industry on
18:10restocking because people said wait a
18:12that could be useful to me to know
18:13what's going on in what store how people
18:15are buying this and correlations and so
18:17on and so on so and now we have that on
18:19you know on like a 10x or 100x larger
18:23volume because you've got exhaust from
18:25your website your security sensor your
18:28web traffic to your web browser so data
18:37Lakers fundamentally have different data
18:38types in reality you want to keep all of
18:40it how do you wrangle that data then how
18:42do you move that data forward into a
18:44consumption by data scientists
18:46ultimately so because there's a lot of
18:48it that is a non-trivial task so we
18:49think of it as a model where you have
18:51three stages you've got the raw sort of
18:53the the the water sitting in this lake
18:55right it's just the water that came in
18:56and then there's some element of
18:58purification that takes place in some
18:59element of bottling that takes place and
19:01we're seeing a lot of companies adopt
19:03that because they need a organizing
19:04principle they need a place to hang
19:06their hat they need to think about how
19:07to deploy these technologies
19:09industrially how to go through dev test
19:10prod and how to essentially mechanize
19:14what is today a very hand-built
19:16lots and lots and lots of open-source
19:18engineers running around in the
19:19enterprise is very hard to corral that
19:21and make it reproducible make it secure
19:23and so on right so if you think about
19:24where we are today and then is it safe
19:26to say the architecture for today for
19:28predictive analytics is we've got the
19:30people side but we've got data
19:31scientists obviously who are incredibly
19:32important here but we architecture looks
19:34like multiple applications some which
19:36are SAS based applications some which
19:38might be kind of behind the firewall
19:39applications right and then you have
19:41aggregation of and creation of the data
19:43lake itself and then on top of that you
19:46would put what reporting tools like a
19:47tableau or something else like that
19:48would be the mechanism to actually
19:49surface some of that or no or you think
19:51there's a you think there's another
19:52stack now that comes on top that really
19:55rink visualization will certainly
19:58data gaining and importance because we
20:00have the geo signal we didn't have the
20:01Jewess level but if you have somebody
20:03coming to your website on their
20:04smartphone you capture some geo signal
20:05and there's there's value in that but I
20:07think just going back to it first you're
20:09going away from a batch architecture of
20:11the 90s do a real-time streaming
20:13architecture like we want our stuff now
20:14this is 2016 and you know we want to see
20:17a movie now so so you're going to go
20:20through a real shift toward streams of
20:23data coming in a very very large scale
20:24on a very big way in the enterprise and
20:26that's I think the one thing that's
20:28taking place that we may not have talked
20:29about and we see that all the time but
20:31how are people consuming it is going to
20:33be through certain Web Apps and and
20:36sometimes you're just putting that
20:37signal into an app to make a decision
20:40we're seeing precursors of this for
20:42example in AD technology that market is
20:44so efficient in how they serve up ads
20:47and real time exchanges where you bid
20:49high low and you put web ads that sort
20:51of thing is gonna happen for lots and
20:53lots of applications and you're starting
20:54to see companies like Salesforce for
20:56example make investments in machine
20:57learning to improve the efficacy of your
21:00sales people calling out companies like
21:02inside sales but what that is is pulling
21:04forward the benefits of these real-time
21:06trading exchanges that are informed by
21:08signals in those data about geolocation
21:10by a preference etc etc and all those
21:13changes are gonna trickle down through
21:15every process in on an industrial scale
21:17so let's talk about where we go from
21:19here right so you've been obviously an
21:20active participant and you know studying
21:22this industry for probably more years
21:23than I care to confess but so where do
21:28we go from here it does the data lake
21:29become a data ocean and or you know what
21:31what what changes can we foresee over
21:33time that actually might also cause
21:34interesting architectural shifts so I
21:37think the the data Lake as a organizing
21:39principle is yet not fully formed you
21:42know it's like an object slightly out of
21:44focus you can see the outlines of it you
21:45know what it is but you don't exactly
21:47have the recognition but the thing that
21:50is changing this is where it'll be
21:52fascinating to see the worlds of cloud
21:54and big data collide and our hunch is
21:57that the all the benefits of SAS app and
22:00cloud apps right you don't have to go
22:02rack and stack you don't need to worry
22:03about personnel and air-conditioning and
22:06you basically get into provisioning data
22:08are going to intersect with the data
22:12do where the data lakes might become
22:14cloud formations you actually might have
22:17data lakes exist in Microsoft Azure or
22:20in AWS as s3 buckets or we'll see what
22:24Google does with bigquery BigTable that
22:26that is going to give you in in a sense
22:28a turbo mode of execution around
22:31analytics predictive analytics where the
22:33data platforms actually go to the cloud
22:35and so so the world we see is a world
22:37which is definitely streams of data you
22:40have streams of data going there's no
22:42longer concept of you know batch or
22:44monthly god forbid or even overnight
22:46it's all now so you have streams of data
22:48coming through you have a world that is
22:50hybrid for quite a long time you'll
22:52continue to have legacy applications
22:54look the mainframe still oh I was gonna
22:55say right yeah no it's older than both
22:56of us right it's like the old Roman
22:58cities right we build on top of this we
22:59don't really the enterprise's retrofit
23:02job it always has been you know it's a
23:04retrofit job like ATM transactions no
23:06matter how bad it COBOL might be it
23:08makes no sense to redo all the
23:10agreements between bank transfers into
23:12new language you just throw more
23:14mainframes at it more water cooling you
23:15know and so on so so what you see is
23:17this layering of a new stack of web apps
23:20cloud apps that are streaming data
23:22they're connecting to multiple data
23:24Lakes which are probably going to go to
23:25the cloud for some machine data factory
23:28floor data power plant data it may not
23:30go just because you know bandwidth you
23:32can buy but latency you get from God
23:34right so that may not go but but a lot
23:37of other stuff it makes so much sense to
23:39put your marketing data into a data Lake
23:40in the cloud with Azure because you are
23:44going to have your website be the
23:46primary producer of data your
23:47on-premises days is actually very small
23:48and then you have a bunch of marketing
23:49applications in the cloud so why don't
23:52we just have the data Lake in the cloud
23:53when you think about the future is there
23:55a rate limiting factor in your mind as
23:56to what prevents us from getting to the
23:58Nirvana that everybody as a business
23:59analyst or any business user would love
24:01to which is how do I actually get real
24:03predictive value out of my data is it is
24:05it the state of machine learning today
24:06is there a compute limit this problem is
24:09very real and it's shockingly large even
24:11in 2016 good chunk of what can make this
24:14is modern technology coming in and
24:16finally we have all these data are
24:19lassoed up and JSON objects and modern
24:21api's do make out of many one I had
24:26dinner with the CIO of a large retailer
24:27and they might decide that the
24:29point-of-sale terminals should be iOS
24:31devices somebody else might decide to do
24:33it in Android because they might be
24:35doing it in Asia and they might have
24:36more fluency with Android somebody doing
24:38in San Francisco may have less you know
24:39so as you do this you really want the
24:41ability to create your own enterprise
24:43with intelligent choices of building by
24:46and using something to glue it together
24:48and that something gives you option
24:50values on the choices that you make and
24:52gives you self-service to be able to do
24:54that without having to have a huge
24:55programming team and skill sets matrices
24:58and so on and so on yeah so if you think
24:59about if I if I kind of sum up a little
25:01bit for for folks we've kind of almost
25:03gone full circle and hopefully we're
25:05going even farther down the circle which
25:06is you know as you described we've gone
25:08from kind of a world of the application
25:10vendors the ERP vendors in particular
25:12effectively dictating the business
25:14process and therefore kind of data flows
25:16and how you think about analytics and
25:17everything else to take back your
25:19enterprise right which is the idea that
25:21essentially self-service you know as
25:23we've seen in so much so much of compute
25:24where you know consumer self-service now
25:26drives into Enterprise self-service that
25:28really allows people to get to the end
25:30they're looking for the the analogy I
25:31will use is you know it's like you don't
25:33have to listen to music the way that
25:35backers did on the CD you can have a
25:36playlist right you can compose what you
25:40want to hear you take back your
25:42enterprise and you you put together the
25:45things that you need whether it's
25:46point-of-sale things or internet sensors
25:48or cloud applications to work with what
25:50you have on premises or what might be
25:53very mission critical for you and you
25:55can have it on enterprise so you know
25:57we've been talking about architectural
25:58shifts but there seems to be a broader
26:00organizational shift going on in the
26:02within the IT organization itself you
26:04know where do you see things going
26:05there's a bunch of things going on
26:07technology is pervasive it is as much a
26:10part of the enterprise as telephones
26:12used to be in the 70s or 60s it's
26:15so the the roles are changing quite a
26:17bit that there was this tension between
26:19the chief marketing officer in the CIO I
26:22think they've largely kissed and made up
26:23because a the CIOs are getting to be
26:26more business people and the marketing
26:28people are getting more technical so
26:29that tension is disappeared in the sense
26:31now with the CTO and the CIO is
26:33interesting most organizations are have
26:38for the CIO because what they do is the
26:40demarcate many of the more sophisticated
26:42technical problems security for example
26:45you have a chief information security
26:46officer you have a CTO who's looking at
26:48where should these applications run
26:50should they be on-premise should be in
26:51the cloud where should these data go
26:53what loud should we pick should we have
26:55a multi public cloud strategy so at that
26:57scale if you are becoming a better
27:00business partner as a CIO and you no
27:02longer have tension with the chief
27:03marketing officer or the head of people
27:05or they had a finance now you have in a
27:08sense to delegate what would
27:10traditionally be a CIO job which is
27:12being a technologist into someone who is
27:15a technologist working for you imagine
27:17your multi-billion dollar business you
27:19have to worry about scalability you have
27:21to worry about the optimal mix of
27:23on-premise and cloud computing for
27:25yourselves storage optimal mix you have
27:27to think about the mix for you sort of
27:29that all the time in addition to that
27:30you have to worry about security you
27:32have to worry about all the innovations
27:34that are happening on machine learning
27:36and predictive analytics and keep an eye
27:38on the competition if you're in retail
27:40you have to compete with Amazon which
27:42has all this stuff baked in and is able
27:44to come out with with new things like
27:46drones and so on so it's a it's a time
27:49of great challenge it's time of great
27:51change but it's also a time of great
27:53opportunity well thank you Gore for
27:55spending time with us today
27:56you're welcome Scott real pleasure to
27:58catch up with you on this thank you