[Music] from the cube studios in Palo Alto in Boston it's the cube covering the IBM thing brought to you by IBM welcome back everybody you're watching the cube and our continuous coverage of IBM think Digital 20/20 events it's we've been wall-to-wall for a couple days now and and we bring in you all the action a bass fall is here here he is the global league for quantum education and open science at IBM quantum gave great to see you thanks for coming on yeah thanks for having me here Dave you're very welcome love the discussion on quantum but I gotta say so I'm reading your bio in your bio I see quantum algorithms experimental quantum computation nanoscale device fabrication cryogenic measurements and quantum software development hardware programming etc so you're obviously qualified to talk about quantum but but how how can somebody learn about quantum do I have to be like a rocket scientist then understand this stuff so Dave this is one of the things that I'm very passionate about it's also my job to make sure that anyone can learn about quantum computing today so primarily what I'm focused on is making sure that you don't need a PhD to program a quantum computer when I was going through my graduate studies trying to learn quantum computing I needed access to a lab so I have to go to graduate school to do this but in 2016 IBM put a quantum computer on the cloud in that dramatically changes the field it allows access to anyone from the world with just an internet connection to program a quantum computer so the question I'm trying to answer on a daily basis now is the question that you asked how do I learn to program a quantum computer well I'm trying to make several resources available for you to do that okay well let's talk about those resources I mean you have quantum you have access to quantum computers I talked to Jamie Thomas the other day she said that you guys it's all available in the IBM cloud I can't even I can't even imagine what the infrastructure behind that looks like but as a user I don't have to see that so how do I get access to this stuff so there are several quantum computers available on the cloud now and every time I think about this it's fascinating to me because I needed access to a lab to access these things but now you don't you can go to quantum computing dot ibm.com and get free access to several quantum computers now the question becomes if I give you this access to the quantum computers how do you learn to program them the software that you use to program them is called kiss kit just like we've made access to the quantum computers open for everyone our software is also open source you can access it by going to Kiska torgue that's QIS ki t org and if you go in particular to Kiska org slash education we've put together a textbook to help you go through everything that you'd learn in a classroom about quantum algorithms and to start programming the real quantum systems yourself so everything's ready for you to program immediately what was the it can you give me the quantity IBM want them - computing URL again yeah that's quantum - computing IBM com once you create an account there you immediately get access to several quantum computers which is an impressive thing to think about the cryogenics that you mentioned earlier the hardware the software all of it is ready for you to take advantage of but I gotta ask you I know it's sort of off topic here but but if I had to look under the covers I'm gonna see some big cryogenic unit with a bunch of cables coming in is that right that's exactly it very cold inside that's right so the way to here's the way to think about it outer space is about 200 times colder than room temperature and the temperature where the chip the quantum chips it's is another 200 times lower than that so we're talking very cold here we're talking only 15 Mille kelvins above absolute zero that's zero point zero one five degrees above absolute zero so it's a very cold system and you'd have several wires that are going down into this coil system to try to communicate with the quantum ship well and what's exciting to me about this whole thing Abe is it is it brings me back to the sort of the early days of computing and the you know huge rooms and now look where we are today and so I would expect that over the next many decades you're going to see sort of similar advanced advances in quantum and being able to actually execute at somewhat higher temperatures and in miniaturization it's very exciting time and we're really obviously at the very very early innings but I want to ask you just in terms of if if I'm a programmer and I'm a Java programmer can I actually come in and start using quantum if you what do I need to know to get started so you need to know two things the first thing is you need to be familiar with any programming language the easiest programming language to pick up today by far is Python so kiss kit is built based on Python so if you're able to quickly catch up with a few things in Python and we have a chapter dedicated to this topic in our textbook that's the first thing the second thing is simply having the ability to learn something new simply being excited about this field once you have those two together you can learn quantum computing very quickly within a few months the question then becomes catching up with the research and reading research papers that can take some time but for us to be able to talk through a quantum program takes only a few a few days of reading let's talk about what some of the folks are doing with quantum we talked again to Jamie Thomas and she gave me some examples not surprisingly you know you saw for instance some some examples in pharmaceutical and to the other obvious industries but then banking came in it's a but what what is it what are people doing with quantum today maybe you could add some color to that primarily most of the working quantum today is focused on understanding how to take problems in industry whether it is to understand how to simulate molecules whether it is to understand how to optimize a financial portfolio taking those problems and mapping them onto a quantum computer so that they can get solved so you'll see various various industries exploring how to take their problems and map onto a quantum computer so one one exciting one that I'm seeing a lot of progress in is chemistry learning how to simulate molecules using these quantum computers as someone with a physics background for me the exciting thing to see here is also how people are using these quantum computers which fundamentally are taking advantage of quantum mechanics to simulate other quantum systems so to understand nature better by using nature itself so this is another exciting progress that we're seeing in the field so exciting both from industry and from educational and science purpose so obviously it's a fascinating field and people would you say with curiosity it can get excited about it but but let's say I actually want you know some some kind of career in part of I mean what well how would people sort of get involved do you see you know on the horizon that this is gonna be something that is actually gonna be a vocation for you know young folks that want to get involved I could not tell you how challenging it is to find people who have the right combination of quantum computing knowledge and classical programming knowledge so in order to be able to take full advantage of the quantum systems today we need people who understand both the hardware and the software to some level and there is an extreme shortage of that kind of talent so the work that I'm focused on is exactly this problem of solving the workforce development problem so we're trying to make sure that people have access to anything that they need in order to be able to program a quantum computer and to learn how to then map their own problems into these quantum computers in the future the question becomes let's say we now understand how to use quantum computers to make financial portfolio optimization every bank in the world is going to want someone to implement this in their systems which immediately creates lots of jobs so this is going to become something that's in demand once it becomes possible on a on a large quantum computer so today is the right time to learn how to work with these quantum systems so that when the time comes that there are industries that are needing quantum skills you're ready to be hired for those positions okay so big skills gap you kind of gave an example in financial services where maybe some of the other things that you hope that that people are going to be able to do over time with these skills I cannot under I cannot over us overstate how important it is to learn how to simulate chemistry problems on these quantum computers that will have impacts anywhere ranging from whether it's drug design whether it's making better efficient solar panels more efficient batteries there are many applications where you'll see impact from these so the there are many industries that can benefit from understanding how to work with quantum computers that's something exciting I'm looking forward to see you know you read in the press that you know we're at least a decade away you know from from quantum being a reality but you're giving some examples where it's sort of here today I feel like it's going to come in layers you know not gonna be one big bang it's gonna come over time but but maybe you could you know frame that for us in terms of how you see this market developing I don't even want to call it a market but just this technology developing into a market what what has to take place and what kind of things can we expect along that journey sure so I think it's very important to keep in mind that quantum computers are fairly young technology so we're improving the technology as we go and there has been dramatic improvement in the technology itself but we're still learning as we go so one of the things that you'll find is that all of the applications work that's being done today is exploring how to take advantage of the quantum computer in some way if I immediately gave you a fully functional perfect quantum computer today you wouldn't even know what to do with it right you need to understand how to map problems on to that quantum computer so in preparation for that time several years away you'll see a lot of people trying to learn how to take advantage of quantum computers today and as they get better and better learning how to take advantage of whatever incremental progress is being made so as much as it seems like quantum computers are several years away many people are learning how to program them today just in preparation for that time when they're ready for use and my understanding is we're gonna get there with you know hybrid models today you're using you know traditional microprocessor technology to sort of read and write data from quantum that's likely going to continue for quite some time maybe maybe indefinitely but but but perhaps not right so Dave the important thing to remember is that a quantum computer works jointly with a classical computer if you ask me the question of how do i optimize my portfolio the numbers that I would need to compute with our classical there's nothing quantum about them these are numbers so there's classical information that you then have to take and map on to the quantum computer and then once the quantum computer is done you have to take the data out of that computer and then turn it back into classical information so you'll always have a quantum computer working jointly with a classical computer the question now is how do you make those two work together so that you can extract some benefit that you couldn't have attained with just the classic what do you see is the big sort of technical challenges that you're paying attention to you paying attention to I mean is it getting more you know qubits is a coherence working at higher temperatures what are the things that you see is as the the scientists are working on to move things forward so one of the things that I can do immediately Dave if you and I agreed right now is we can go to the lab and take a quantum chip and put a thousand cubits on that quantum chip that's fine we can do that immediately the problem that you'll find is that it doesn't matter that you have a thousand cubits if the qubits are not good quality cuteness so the technology should focus on improving the fundamental qualities of the qubits themselves before scaling them up to larger numbers in addition to that as you're scaling to larger and larger numbers new problems come into the picture so making better qubits scaling up seeing how the technology is doing learning new things and then scaling farther up that seems to be the model that's working today so in addition to monitoring the quality of the qubits themselves I'm monitoring within the technology how people are implementing solutions to scaling problems in addition to that another important problem that deserves a lot of attention is the question of how do you make good software that can take problems and map them onto quantum computers in in quantum computing when I say I'm running upon a program really what I'm doing is building a quantum circuit and then running that quantum circuit on the real device well if that circuit has certain operations in it maybe you want to tailor the way you transfer that circuit onto the device in a way that takes full advantage of the device itself but then in order to do that you need to write good software so improvements in the software along with improvements in the quantum technology itself will be how we get to success and at IBM we're focused on finding a metric that wraps all of these things together and it's called quantum volume and we're seeing improvements in the quantum volume of our systems as we go yeah Jamie talked about that you're essentially taking the key metrics and putting them into a you know a single observable metric that obviously you can track over time so I want to ask you about security a lot of people are concerned that the quantum is just going to blow away everything that we know cryptography and all the you know the the passwords and security systems that we we've put in place is that a legitimate concern will quantum you both get us into that problem and take us out of that that problem I wonder if you could talk about that so there are two ways to think about this problem one is just fundamentally if you ask me what does it take to put the the cryptography that has our bank accounts safe over the internet connections that we use it takes roughly about a thousand good cubits okay if I tell you a thousand good cubits that doesn't seem like a lot of work but when you think about it what it really requires is an overhead of about a thousand cubits for each qubit that we have today so the numbers of qubits that you need are in the millions in order to put the the kind of cryptography that we're using today at stake so certainly there's a long way to go that's one aspect of the story the other aspect of the story is that we should never underestimate the progress of technology so even though the time when we can use Shor's algorithm which is the algorithm that can be used to break the cryptographic algorithms like RSA even though that's several years away you still want to be ready for that time and what that means is if you have sensitive information today you need to be making sure that that information itself is protected with quantum resistant cryptographic techniques so that when the time comes you can't use a quantum computer to get back the data from today and break so two perspectives one is we're quite a while away from this kind of danger but at the same time it doesn't mean we should be complacent today we should be taking preparations make sure that our critical information is protected yeah that's so that that makes a lot of sense but when you say we're a ways away or we are we decades away we years away we can you and you quantify that in any reasonable way it's hard to speculate on that number so I'll refrain from giving you a specific timeline just to give you an idea the quantum bits that were in development ten years ago had a coherence time so the amount of time that they can store the quantum information of roughly a hundred times smaller than they are today and ten years ago if you asked people how do we get to a hundred times better qubits nobody would have been able to give you a clear answer you could have guessed some ways but nobody would have been able to tell you we'll get there in ten years but we did so instead of coming up with estimates of timelines that depend on what we know today it's probably a better idea to monitor the technology as it goes and keep adapting we're probably talking this century where we're talking to the century hopefully it is my last mission to enable enough people to learn quantum such that it happens within my life very exciting field a I can't thank you enough for helping us educate the audience and and my and myself personally really I'm I'm so fascinated by this it's something that you know jumper and I and the team have been really focused on and I think it's really time to your point the start digging and start learning you've given us some resources there give us give them give us those two reasons one more time there's there's the IBM site and the the the the the queue kit site use that site what are those again just those to wrap so you can access the quantum computers at quantum - computing ibm.com and once you're there the way to learn how to program these quantum computers is by using kiss kit which you can learn about by going to kiss kit org slash education once here at that education page you can access our textbook which we make open-source it's a textbook that's co-written with professors in the field and is open source so it's continually getting updated you can access that textbook at tisket org slash textbook if you go to our youtube channel you'll find several videos that allow you to also learn very quickly so kiss gets YouTube channel is another great place to look so lots of resources and that's kiss kit with a Q which is why I wrote it that way so alright exact thanks so much it was great to see you stay safe and next time hopefully we'll see you face-to-face and you can draw some some cool pictures to help me understand this even better Dave it was nice talking with you I look forward to learning quantum programming with you yeah Cheers and thank you for watching everybody this is the cubes coverage of the IBM think 2020 digital event experience we'll be right back Brennan for this short break [Music] you

okay welcome back live in Las Vegas is the cube ibm's information on demand conferences q exclusive coverage SiliconANGLE will keep on here live I'm John furry the founder of silicon Hank I'm Joe mykos Dave vellante co-founder Wikibon org our next guest is a Fred Balboni global leader business analytics optimization IBM GBS global business services you know obviously big data is powering the world I mean just can demand for information and solutions is off the charts afraid welcome to the cube anything there's a services angle here where you know services matters because one in the channel partner is this good gross profit for helping customers implement solutions that they have demand for so you've a combination of a market that's exploding with demand people know it's a game changer with big data analytics cloud is obviously right there in the horizon in terms of on prem of Prem then you've got now see mobile devices bring your own device to work which is thrown off more data okay and then people want to be in all the different channels the social business so you know CIO to CEO says hey this new wave is here if we don't think about it now and get a position and understand it the consequences of not doing anything might be higher than they are so we've heard that how do you look at that and what are you guys doing what's the strategy give us a quick update and from from GBS i think that the to make this successful first of all it services is important it's the last mile you know that means the point you may it's the last mile and without without that you cannot ever deliver the value the the really interesting challenge that every executive faces is you need to be able to we can easily get our head around big data technology and I shouldn't trivialize that but you can go and understand the technology what's possible in big data you can also get your head around analytics and the analytics algorithms and the kind of insights that can be drawn from that the real challenge is how do you articulate what's kind of possible to a client because many of the use cases are very niche and so clients often say yet that's right but it's big it's possibly bigger than that yeah that's right it's possibly bigger than that the other issue or the other challenge to get we've got a hurdle we've got a jump on me articulate this to the businesses clients businesses think in terms of process you don't think in terms of data you know you don't go talk to a CIO CEO and say you know tell us what's the key attributes of your customer and they don't think that way they can talk to you about servicing a customer or selling to a customer or managing customer complaints so that the processes but the data it's a tough thing so the first part the services is so crucial in this is being able to articulate the value of analytics and big data to a client in the businesses terms so it becomes a boardroom conversation kind of so that's that gets the program started and then quickly being able to fill in with use cases because clients don't want this to be they don't want to start from a blank sheet of paper and they don't like going to give me some quick wins here so it's kind of those timetable what kind of timetables mmmm back in the 80s 90s when client-server rolled out it was months and months yeah project management meetings roll out the Oracle systems roll out the big iron now I mean I'll see maybe shorter spurts little different hurdles what's the timetable only some of these horizons for these quick wins okay so project implementation I come on now let's let's know it's it's I think that that we're measuring project implementations in weeks I think cloud-based technology allows us to provision environments on the order of a couple of weeks and that used to be on the order of five to six months so I think that's going to that accelerates everything and that also allows you to do a lot of a lot more speed to value get applications or analytics use cases up there much more rapidly one two as you start to build these portfolio of use cases and if they're built on acceleration tools I mean acceleration so you've got those code sets that are already there that you can add you can jump on top of I mean you can get these use cases up there in 6-8 weeks we have one we have an example a really large major company i'd rather not i'd rather not because it's not externally referenceable but a really a significant client that had on the order of more than more than 5 million discreet customers and doing detailed customer analytics on their customer base against their products and we were able to get that baby up and running in three and a half months now that two to three years ago traditional logic would have told you that was a nine to twelve month project and by the way you know ten years ago that would have been a 18 to 24 month project yeah so I think that yeah we're moving much more rats the expectation now too I mean the customers realize that too right the absolute not but but there's one thing I want to talk about this it's still this is the one thing that if you'd asked me what's most important this speed thing allows you to go rapidly to places but you you better have a navigation roadmap on where you're going because if you're going to do all kinds of little code drops that's great but you want to make sure you're getting leverage so you're going somewhere so therefore there's a scale but this is where roadmapping becomes really really important for every the technology side of the business you have to have a technology roadmap the other thing that's really important out of this is if you don't let's use the client-server example you used because this kind of has a you know we've all been here right here we've all lived seen this movie before yeah if you if you don't in the build this roadmap another thing that happens do you remember when CIOs finally said okay I'm taking control this client servicing sure what do they end up with they ended up with all these departments of computing in the costs work going astronomical so if you've got a road map you can also address the issues of managed services because you don't the least thing you want to be is having all these data Mart's that are scattered everywhere because you get no economies you get no economies of it but a cloud would bring you you get Noah kind you get no economies and being able to do that and you end up having to have all these maintenance teams you know that maintenance and by the way analytics by its nature has constant maintenance little adjustments and changes you're getting new economies of that because they're all managed is discrete units so therefore there's a lot to be as you build this roadmap you've got to think about the managed services environment as well so Fred you talked about earlier clients don't think in terms of data they think in terms of their business process is that a blind spot for clients because there are some companies Google for example that does think in terms of data in your view should clients increasingly be thinking in data terms or does our industry have to evolve to make the data map to business process I actually I kind of just take it as a thick I don't I don't I don't choose to question why I just accept it um i but i would say i which i would say customer's always right I just I just think the industry i thought that definitely but i think just the industries at a stage where you know we've always you know back in the old days of you know i'm going to show my age here but you know the procedure division in the data division oh my god looked at all and and and we you know the procedure division is where you actually did all the really and i think if the reason is we got understand the paradigm under which modern computing was created I don't to be like we go into history lesson but the paradigm under which modern computing was created was that we use computers to automate tasks so we've always taken this procedural approach which went then we went to process reengineering and that became a boardroom conversation so just I think we've conditioned over the last 40 years businesses to think about using technology to gain business efficiency they've always thought in terms of process so that's why this data element yeah companies like Google founded on analytics clearly have got a whole different headset in a different way to approach these which gives them a built-in bias when they address the problems they've got in their businesses sure but you don't come a decline saying hey you got to rethink the way in which you look at data you come in and say let's figure out how we can exploit data in your biz erect what we do it two ways we do it two ways first of all let me not dress let me not dress monton up as lamb at the end of the day it's its data its data okay now the question is how you articulate that and it's twofold we tend to I like to use a metaphor to describe the data so if its customer that the metaphor we've been using recently is DNA DNA strands to be able so you use a metaphor that there's a language that the business can relate to and you can create a common language very easy one in that way you can have an account because you're never going to drag a CEO into your fourth normal form data model so so therefore you've got to you've got to talk a language one number two you talk about as a collection of use cases so you use use cases as a vehicle to have the process conversation and because with the use case you also can talk business outcomes benefits and you can tell kind of a story you don't have to drag them through the details of the process but you can tell them a story whether it's you know I if you can understand called detailed called detailed data records and the affinities you can understand the social networks and therefore you can reduce churn within your telco customer base as an example quick but if you follow I do so you talked about its little use cases and they begin to understand wow what's possible and then you talk about their data as a DNA chain and they get I got it I actually need to get the DNA chain if I'm going to actually think about think about my customer base or my product base or whatever the lingua franca the business is still the businesses language it doesn't result of data but data can enrich the conversation in a way that can lead to new outcomes the data in rich's the conversation when you talk about the business outcomes that are created as the part of the use case well it's like a three third order differential equation but i go back i watch this yeah i just go say your tweet your epic soundbite machine just can't type fast enough on the crowd chat it's good for good for Twitter viewing yeah I've just opened a Twitter account please look me up I'm looking for friends I promise to start posting you got people watching all right all right so so in terms of customers right give us a little bit peak of some of the customer responses when you when you open the kimono show them the road map you know the messaging around on IBM right now is pretty tight here at IOD last year was good this year is better you look really unified face to the customer when you show them the road map what's the feeling they get it they feel like okay I got some trust IBM's got some track record history do they is the is the emotion more of okay where do I jump in how do I jump in there doing it and this little shadow IT going on all over the place we know with Amazon out the area so so when you're in there you've got to have these are conversations what do they like and what's that what's the level of response you get from CIOs and then also the folks in the trenches so there's always a question which there's a couple of questions first of all is how can I get how can I get value from this and that in that and that's you know a I'm tightly coupled to my existing transaction processing which is kind of like if you will call that turbocharged bi and and which is which is where so many people have come from is this turbocharged bi environment and listen that's an important part of your reporting business you need to do that to keep the wheels on the question is as you move to this notion of analytics giving you great insight then then you've got to say okay I need to go from turbocharged bi to really augmented components so clients I'd say there's a large there's a large group of people that are right now moving from turbocharged bi to the notion advanced use cases so there's this some disco a large discussion right now how do I show me do use cases by which i can I can rapidly that would be advanced how to linux up the calling advance limit well no we have well 60 60 use cases industry-based use cases that we as a services business put together on top of that we have about seven or eight key code fragments that we uses accelerators I mean we call them wink we call them assets and we just them up as accelerators but their code fragments that we bring to a client as the basis that we put on top of the the blue stack of technology to actually get them a speed to value because we really want to be able to get clients up and running within this notion of non idealities it's like literally being best practices in the form of technology to the customers well you're on an IBM thing I mean dare I called an application no I wouldn't dare call it an application we're not in that business but the point is is that it is it's starting to feel like an application because it's really moving down these unreal integrated solution is really where we going it's an accelerant this code correct so it's leverage the economies of scale is every success breeds that's exactly it more and then on top of that we would have that just don't throw a few other things that we do to accelerate these things we actually have five what we call signature solutions which is services software together with a piece of services code coming together to solve a problem we've got that round risk and fraud around customers I mean some specific very narrow things if somebody wants to you know because often IT departments they want to buy something they want to buy something they don't want to go down the parts they want to buy something and so fine here's a package solution let's go buy something um and then last but not least one thing we haven't talked much about but I always like to throw this out there because I think this is one of the things they and we didn't talk about it much in the main 10 or any better sessions but let's not forget about IBM research I'm really proud to report to you now since we started this category we've done 61st of a kinds with IBM Research so this is about client says I've got this problem i think it's unachievable i cannot solve this problem you know help me map in my oil exploration like things that are considered big problems big problems let's let's apply this group that does patent factory you know that IBM is but 15 years in a row let's apply those people to my our problems and we have 60 we have 16 so we do about 15 to 20 a year so it's not like we like we're not cranking these out like I'm hundreds of thousands of licenses but it's where basically our services business our software business and IBM Research go work on solving a client specific problem you heard Tim Buckman this morning when he was asked to know why IBM that was said IBM Research was the first answer that's right he gave we talked to him about that on the cube you know in his is insane me as a customer and we you know we always love to hear from customers I mean you know the splunk conference just had was just last week as an emerging startup because probably well aware of those guys they have customers that just say just glowing reports you get to the same same set of customers you know he is someone of high-caliber at the command and control in his healthcare mission and he's automating himself he it's and essentially creating this new data model that allows it to be pushed down to be listen you've got to do this and I'll tell you why you remember the the governance discussion is it was well I'm most excited about is the governance discussion five to eight years ago was an arcane discussion available of data modelers and like what do we do the governance discussion is quickly moving into the language of our business people and the reason is because they're beginning to do you remember the days of accounting systems when they say we want our accounting department to focus on analyzing the numbers and not collecting and forming the numbers well we're here again and if you've got good data governance you can focus on creating the insights and determining what actions you want from the insights as opposed to questioning the numbers and questioning the validity and the heritage of the number the validity and the heritage of the numbers and in this place everywhere yep financial services companies are the most stressed about it because the validity and heritage is required when you want to prove a compliance to a federal statute yes but it means everywhere if you're a consumer packaged goods company and you don't believe that sales are down in a certain market or a certain chain store first thing they do is they start challenging the numbers if you have good governance you can now start that you can now start to trust these systems of record but let's talk about data quality data quality but it's also the governess in the death of mindset is much broader iteration right how we said the first you know that folks from the nonprofit said you want to go on the record but he's basically saying I'll say basically when you put stuff out when you package and then bring it out it still might have some flaws in the data quality but it's the iteration is transformational but once that's in market saying that's changing he things prepare pre-packaging data and then bringing it in is not the better approach but I want to ask you about the your what you just said about this governance conversation that is date the core of this debate around the data economy what is the data economy in your mind given what you do the history that you've lived through we've seen those movies now the cutting edge new wave that will create new well for new ways change from transform business all that stuff's great but what is the data conn what does that mean to business executives that they're focusing on outcomes is is it changing data governance is it changing the value chains is it changing what's your thoughts on that the data economy is about discovering those points of leverage that that the data tells you that your instincts don't the data tells you that your instincts don't one of my favorite stories three years ago four years ago we were called in and clients said this is my problem the going and problem was I got to take 200 million dollars out of my advertising spend budget two hundred million dollars out of my advertising spend was he's a retailer end and the problem is is out of my 600 million dollar advertising budget the problem I have is also have all kinds of interesting theories and models that my agencies have told me I'm not quite sure do I just take 200 off the board across the board do I take 200 off to minimize my risk just spread it around how do i how do I manage the process and what we actually did was we built a super super set of sophisticated analytics which tied to their transaction systems but also tied to their social media system so we also understood and what we did was we were able to understand which customer cohorts responded to which media types then we added one more parts of the model which is we understood the trending in the cost of free-to-air cable radio internet all the different media types and as we looked at the cost models of them and we understood which customer cohorts responded to which media types we suddenly realized that they were super saturated in certain media types they could like doubled their spin and they wouldn't got want any lift in the advertised in their in their sales what we did was we got 200 million out of their budget and increase they got 300 million incremental sales that Christmas season because we help them get really smart about the play let me tell you I tell us privately i maked media buyers look at me like like I'm like a pariah yeah but but it is actually really you know really started to rethink now there's just a really great example because I think we've all can relate to that but that's the data economy where you find these veins of gold in these simple correlations and from that simple correlation you can instantly go and your business you can get the lift listen I can get five percent I IBM get five percent ten percent lift in some small segment business I've got the volume that's going to make a significant difference to my share one small piece of data could open up a window kind of had with Jodie Foster we would contact words like one piece of data opens up a ton of new data I mean that totally is leverage and it changes the game for that customer and and that to me is that is the guts of the data economy identifying those correlations and and what we're finding is our most recent study we just released it here the thing the IB the IBM Institute for business value big data and analytics study w IBM com it's the Institute for bit I bv study on big data just released and said 75 percent of all companies that are outperforming their peers have said big data analytics is one of the key reasons and the human component not to put are all on machines it's really about it's an ardent science its a mix of both the math and the human piece well you know there's this notion of not only do you create the insight but you've got to take action on the insight you know it's not enough to know if I could predict for you who's going to win tonight's basketball game you still got to place the bet you still have to take action on the inside and so therefore this notion of action to insight is all about trust trust in the insight trust in the data and trust in the technology that the business trust the technology and it's until you take that leap of faith remember when the Indiana Jones movie when he liked the leap of faith and you've got to like to step out and take that leap of faith once you take that leap of faith in you suddenly have trust in the data so that's that trust to mention and that's a human thing that's not a that's that's not a that's an organizational thing that is not a lot of technology in that one okay Fred we gotta wrap up i'll give you the final word for the folks out there quickly put a bumper sticker on iod this year's and put on my car when I Drive home what's that bumper sticker say for this year it's not all about the technology but it starts with the technology ok we're here live in Las Vegas we're going to take about that bet that was going to win the games and I will be the sports book later this is the cube live in Las Vegas for information on demand hashtag IBM iod this tequila right back with our next guest if the short break exclusive coverage from information on demand ibm's premier conference we write back the q