hi I'm James Maguire editor of data mation and our topic today is what's wrong with big data why isn't it perfect yet why isn't it as simple and fast as we'd like it to be to talk about that uh we've got three great experts with us is Valentino tesro Tesco uh technical program manager at um for Big Data at Google hello to you Tino hello uh you are you are there at the googlex today correct oh I'm actually in Seattle right now you're you're in Seattle okay so you you you never say you see Sergey walking down the Halls ever uh I see a lot of SAR not so much Sergey oh okay interesting also with us is San sangani CEO of Eli hello to you San hi James it's great to be here so give us the uh the little 11c uh elevator pitch what does Elian do I know that the tagline is Yelp for Big Data but but what what does that mean yeah I mean I think it's relatively simp simpler similar to to yel or LinkedIn or Amazon or even to a certain extent Google I think all of those are systems that allow you to understand um a certain class of information we help people inside of companies understand the data and the systems of data inside of their companies and so like Google we crawl and scan those systems and we look for what people are doing with the data and use that information use that human activity Trail to build a very very rich catalog that people can update collaboratively socially and you know ask questions around um and you know why that's useful and relevant is because the world's only getting to be more complex place and so as analysts inside of companies try to M more and more information for profit um getting to that information faster and more easily with context is you know next critical problem sounds good also with us is uh Mr Mike matchet analyst with the tenasia group allar around smart smart person how you doing Mike oh good today good today thanks James I know that you and I have done done some deep Dives with Hadoop and big data so uh you're definitely one of the go-to people when it comes to Big Data uh so the the question is you know what is the what is the friction involved with big data when you think about you know it is sort of promised as this fairy dust we're GNA you know sprinkle this fairy D over our difficult questions that our big data solution was going to give us these gleaming insights of of information we're all going to go off and earn a zillion dollars with this great data it hasn't always worked out that way uh there's been a certain amount of confusion even some backlash uh so so what is what is the frictional about it in other words how can we get more people using more data faster uh sthan what what's your take on that yeah you know I mean I think the promise of bi is really you know very basic people get these insights and you know insights help you understand the true nature of a thing that's sort of the definition of the term and if you think about this notion of insight it's really relevant to you you got to have the right information with the right context you know for the right person and the right right moment one person's Insight is another person's learned fact 20 years ago and another person's Insight is something that somebody won't understand for another 20 years and so having the right bit of data at the right point in time is is really the critical thing I think the problem is that the translation effort around understanding mining visualizing uh and really just getting your hands around the data is itself a real problem you know data is built by computer systems computer systems are built by programmers using some very complicated algorithms and abstractions and fundamentally to get acquainted with the data you have to understand where it came from how it was produced who's worked with it what questions it can possibly answer and all this stuff takes a lot of time no different from somebody researching you know something in a library and trying to read up on it it takes a lot of time and so I think what we find is that people would love to get more insights and nobody says that they want to be less insightful but the reality is that there's only so much time in a day and most people can't spend the time it takes to gather the knowledge to start developing these insights arbitrarily and so for us to sort of build this kind of data driven world that's going to make us these zillions of dollars that everybody has promised I think there's a auxiliary reality that is that people have to figure out a way to lower the cost of learning about the data learn from the data you you saying it it's lowered cost and there's also a Time factor I think I heard in there as well so it's it's a question of just letting the time pass or or us evolving and finding a lower cost solution yeah absolutely I mean I think that you know if you think about it data like you know anything else has a well not like anything else but like any information asset has a half life you know um if you if you know what your customer is about to buy before they're going to buy it that's really great if you know about it two seconds after they make the purchase not super useful and so you know the cost of information is one piece of it and and lowering that cost then presumably makes it faster and easier for you to get to that information but you got to do both um you know if you don't have adjus in time in the right place in the right form not useful just opportunity goes Tina what's what's what's your view how does how do we make this easier and how do we eliminate some of the the innate confusion involved with big data well I I mostly agree with Satan's points um uh really big data is difficult uh for three reasons um the the current technology spectrum of big data is very very expensive and very complex uh to your leverage it requires a lot of specialized talent that you may not have in your uh in your company and in your Enterprise and of course um Big Data carries a lot of uncertainty when it comes to Value right folks think that they should have big data projects going on in their companies but not always do they know what exactly that will mean in terms of value um Google has uh long been a um you know has a tradition of hiring really smart people and giv them very very easy tooling to create numerous hypotheses and test them and reject them as quickly as possible um and so Google Cloud platform is really the externalization of Google's tooling and we feel like we have a really good story around giving folks um tooling that allows them to operate a really massive big data scale without the complexity and without the tinkering of that is associated with traditional B dat Technologies but if you would if you would advise some theoretical company out there that who who may or may not be using the Google solution who may be uh what what would you say is there is is there some way we can make this a little easier than it already is what is there is there a simple answer to that incredibly complex question of course um I I think the the goal should be to maximize the ability to iterate on Big Data to experiment with big data right data should be the goal and um a lot of companies don't have the core competency that surrounds a lot of Big Data projects so if you go it data Sciences is that when you say core competence you mean in our data Sciences or or not necessarily that's that's one example um when it comes to uh building Big Data projects on paramics for example um a lot of companies their core competency is not building data centers right and so that is one distraction that a lot of companies fall into when they start their big data projects they try to do it on premise and there is a lot of overhead a lot of time that requires before you're able iterate on Big Data if you start on premise so generally and Gardner agrees with us we we propose that maybe once the your big data project is mature you can put it on premise but to start to iterate on your big data you should start in the cloud uh Mike how do we cut through the friction I mean there there's there's sort of an obstacle between us and the answers how how do we limit that that obstacle for Big Data oh I get to solve the problem right this this is it I mean you know when it all comes down to it we're looking to you Mr mat for the answer you whatever you say uh uh well I mean I I I think you know just to just to step back a bit you know the way the way I kind of look at the the problem is it's hard to jump into big data and just mess around and get anywhere and I think what a lot of folks that you mentioned have gotten a little disappointed with is they say you know hey it seems like a good idea we've got a lot of data I think let's just go mine it and pull the value out of it and and you know build something in the back room and run with it and that's really not going to get you anywhere uh uh good um and and and part of the problem is as as we've just already mentioned it's a it's a skills problem you have to be able to approach these new you got you got business analysts who understand how to look at data they look at the data warehouse all the time they look at their business intelligence they want to make strategic decisions but then they jump over and try to take those same skills to a big data set where you're really looking try Say Hey I want to do machine learning I want to do predictive things I want to impact operations in real time and it's it's a little different way of looking at the world and different sets of data might be involved uh in different architect Ure certainly um that they're not uh it certainly is not used to dealing with I think that gives you some really big challenges right there uh and at the same time um there's these Enterprise it requirements that everyone forgets about so they go off and they build their big data project in a in a clean room and then they go to put it in production they realize hey wait a minute what about security what about data validation what about uh quality data quality what about Disaster Recovery what about what about all this other stuff I think that really throws a monkey wrench into getting imately getting the real value out of uh uh as referred to as a big data project um at the same time I see things as been mentioned there's there's definitely ways to approach this to take some of that burden off of the rank and file company or community and one of them is is to use a service provider like Google or or the number of I I'll give you guys a credit here there's a number of service providers that do that um you get a free Google search Mike for that I get pre Google I get pre ad placement maybe if I if I all right um and uh uh that uh uh there are tools now that can be brought to bear to track the data and to secure the data uh scrub the data keep credit cards out of it and so on so that you're not mashing up personally identifying information unknowingly uh and I think that I think the ecosystem is evolving really rapidly in that direction so um I I think we're going to see good things coming out of the next year uh at all the big uh uh uh big data shows I've been to this year it's finally been focused on security um which is to me a sign of maturity in the market so you generally have to talk about performance first which is what Hadoop was all about M produce to start with then then you talk about availability which is what all the dros were about like how do we make this system actually work and stay available uh and now we're talking about security finings like how do I secure that data and protect it and bring it into production environment so I really see the next year as H something where a lot of people start to really inul get past that uh disappointment into really recognizing value interesting on on the security uh security point I think it's you know one of the headaches the companies think about you know first of all who has access to all the data you know who who are the CIS admins who have access to all this data can it can it be hacked I mean Tino is that is that one of the things I mean if if a company is worried about security with big data why would you what would you say to them that's a really interesting point um when you when you think about security when you think about all the all the recent hacks that have occurred um I will say a lot of those hacks have occurred because a company was a company has made a decision to own the security story um I think it makes a lot of sense to partner with uh companies who can leverage economies of scale to provide a security story who are invested in providing security for hundreds if not thousands if not millions of customers um and which is why you see a lot of uh public Cloud providers um folks are starting to kind of realize that public cloud um is in is not as secure as private cloud is not necessarily true um and as as uh um as Mike was saying um that's kind of what we're facing right with big data and with public Cloud merging together uh the security story emerges something your take on on security sort of The Nightmare of security and Big Data like how do you there's there's some there's some it executive out there who's extremely anxious about the idea of you know when when the when the idea of security gets mixed with ad of big data that makes them really nervous I mean what what would you say to them yeah I mean I think it's really it's really tough I mean you know if you look at the Ashley Madison Scandal that just happened right that's the most recent one you know it's fair scary I mean you know on one hand I think that all of us have little sympathy for people you know potentially philanderers and you know looking for looking for people to you know cheat with on the web and on on the flip there's probably a whole bunch of people who are on that website who you know really you know are are you know maybe not have done nothing and possibly are getting accused based on intention and it's mining this intention data that's really interesting because now we're able to through all of this data get inside people's brains and get inside of their heads and I think that puts a huge responsibility on the people who are consuming this data about how to use it um who's using it and I don't think these are just technical problems right it's not just about securing the data itself it's about you know being able to make sure that the appropriate people inside the institution are able to see what the rules are around how to use it around what could be done with this data being able to warn people and Empower people with the appropriate knowledge a PRI privacy issu it's almost like a big brother issue with big data absolutely I mean I think I think there's a huge set of responsibilities that come up socially um around training people to learn how to use the data and when to use the data and where to use the data better and I don't know that that's just about sort of locking down and assigning roles and and rights and privileges I think there are Technical Solutions but I think the the the more difficult Solutions that if you're a CIO you have to face are you know what do I do about these insiders who have access to the data who may be a threat um what happens if somebody figures at an Innovative use as Target did around predicting you know whether or not somebody's going to have a baby and do we send them a mailer and how do we think about these problems internally and how do we make sure that those don't threaten our business and I think um you know those are really the ones that are much more profound and I think the ones that are much more difficult to contend with and I think we see that all the time with our customers because you know a lot of what we do is help customers broaden the use of data inste their Enterprises and you know the biggest question becomes well if we're going to broaden the use of all this information how do we make sure that we're doing it in a responsible way and how do we Empower people with knowledge that maybe they should only have contextually or only have in a certain particular way and so the security challenges I think don't go away um you know by any extent I think they just get you know worse and worse over time because as you give people more knowledge you give them more ability to do harm right I mean that the target issue you bring up where you know apparently based on shopping patterns the Target in theory knows where whether the family member may or may not be pregnant obviously it does create an enormously sensitive issue I mean does does Target or some retailer then begin sending mailers to the house I mean what what is sort of like there there's the Big Data solution there's some very very human issues about how that that solution gets used yeah um Mr matcher what what's your take on security and uh and big data is there any way to to to lessen the headache there yeah well so just just to follow up on that so I saw that there was you know the conversation descending at three levels and security at one level being about uh making sure you've got the data under lock and key and and that can include U adhering to laws like making sure you don't Geo repatriate some data that's supposed to stay in one country versus another country uh at a second level you've got uh regulation and compliance issues which is again still sort of laws enforced but to say you know I won't even though I'm allowed to have credit cards as a company uh I won't store them or use them in certain ways and I have to protect those from being accessed by uh certain members of my organization that might not use them in the right way or prevent them from leaking and then you've got what saan brought up which is intent which is uh you know Minority Report kind of stuff like uh uh I can mash up all this data and now maybe I'm getting to the point where I'm not just finding out um uh what someone's AR are done but what they might intend to do and should I leverage that or not and uh you know and I think the the rule has to be more of more than simply will I be embarrassed if I do it this way which seems to be what most people are talking about now someone's someone's pregnant and I Target them with a with a with with a I mean with Target did actually right so it was then that that uh very directed uh advertising which got everyone in the into an uproar and what they decided to do instead is now bury that inside other advertising and blend it in so it's not noticeable not solution to the problem but they're still using that data that way they never said they were going to stop doing that in fact they they they still exploit that that that intention they just do it in a way that's not um straight line A to B so I think I think the moral issue goes deeper it's like should they be doing that should we be allowing them that and is Privacy Information that but you know we certainly can't solve that today um I I would say that there's a lot of value to be mined out of big data that doesn't go into those gray areas that can be about company performance or operational Improvement or uh predictive scoring on on which action should the company take uh themselves uh to be more optimal in their business processes and for a lot of companies they're probably safer in those areas than they are in um dealing with um uh client uh and external third party information I think that's a real sensitive area that has yet to be worked out what about the people who say if you ask them you know how do I make big data simple they they'll say they'll give you either one of two answers one would be well uh you you you you want to be on in in a cloud-based solution then that really solves all your problems or the other side says oh it's got to be on Prem because of course all your data is on Prem so does does does a cloud-based solution or an on Prem solution necessarily solve your headaches when it comes to Big Data is one clearly and obviously better something your take on that I mean I think it largely depends on who you are as a company and where you are in your maturity cycle I I think that as the cloud matures and you know as you've got amazing entrance like Google and you know of course there's Amazon um who who are you in incredibly sophisticated about managing infrastructure it makes less and less sense if you're a you know small company a medium-sized company and probably soon if you're a large multinational to start managing that infrastructure on your own I mean Google Now manages email for you know very very large Fortune 500 clients and there's no reason why they shouldn't manage the infrastructure for um you know very large data centers where you're processing and managing a lot of Big Data um so I do think that as As Time evolves the infrastructure and the cloud providers will grow up so that the cost of moving the data around is not as high and the infrastructure cost is relatively lower than one could build in house that that would seem to make sense um you know whether we're there or not today I don't know I think it really largely depends on the use cases and my guess is that Tino could probably you know have a much uh view on that um in particular um so I you know I think I think data is gonna but largely I think data is going to exist in a lot of different places I mean I think we live in a multi sort of you know lots of hybrid environment I mean there's there's compan that have stuff on COD there's companies that stuff will have stuff on premise and most companies have both and you're going to find data is going to move into and out of the each sense for a long time and so I don't see that largely changing anytime soon per perhaps over the long run it will move to the to you know sense you have an opinion about whether or not a cloud-based solution or or an on-prem solution is is better and I guess as as part of my question I would ask you like what about that salesperson who's on the way to a sales call and you know her her region is a southwest region she wants to ask a really fast query from her smartphone app and she needs an answer like in the next 11 seconds uh is is that I mean what what happens when you're when you're a CL when you're a cloud-based you know Big Data environment in that sort of situation so of course I'm a little bias because I work for Google Cloud um the one parallel I like to draw between the the private and public clouds is uh the banking industry right before Banks came around uh thousands of years ago you had to keep money in your own home right and if you had lots of money you had to build a fence around your home and you had to dig a trench and you have to provide security and so forth but then through uh economies of scale this central location that is vested in security of your money that has um economies of scale um allows you to do much more with it right but there's still room for a lot of folks to have that safe at home to keep their uh Family Jewels and gives a lot of cash in there so I think I agree with satian there's going to be private Cloud will never go away there's always going to be room for private Cloud um in the context of Big Data the public Cloud allows you to operate at scale without having to do lots of Investments without committing a lot of time and effort right so so that time to iteration and time to Value decreases greatly in public cloud and of course with the Next Generation uh Cloud native products like Bor product that I'm working on you're able to do some incredible things and one example of course is with borey the ability to scale from zero cores to 9,000 cores for 11 seconds to generate a terabyte sized uh SQL query and then scale down and only pay for those 9,000 cores for those 11 seconds this is this is absolutely impossible anywhere else outside of cloud right but what about this the speed issue for that theoretical you know salesperson on their way to a call and and there is actually just very a very bottom level Big Data query yeah is that is that is that going to work is is is there still a speed issue for that person or not necessarily in your view well I think you can replicate speeds in if performance is what you're after you can duplicate performance whether it's in your private cloud in the public Cloud but when you when you intersect performance with cost with um time to Market um it's a it's an equation with a lot of variables right and I think that equation overall is much more attractive in the public Cloud now of course there's companies like Etsy who have moved off of the public Cloud into back their own data centers because they have a very predictable analytics platform they know exactly what's going on they know exactly what uh what they're building they're very comfortable with it it's a very static workload and they're very comfortable running their own Technologies they consider it their core competency I I've always wondered about those companies that have the attitude how can they say to themselves we have a very stable environment I mean what's going to happen 36 months from now I mean really there there's such a thing as business stability especially for a business that's online but well okay that's a different question I think it's a fundamental question to analytics I mean you know it's funny I think that everybody um you know the old world of business intelligence was I wait two years you know I ask a set of questions that I want to have answered through a set of dashboards and I wait two years for those dashboards to get developed and I think to precisely to your point a stable analytics environment is sort of in some sense an oy moreon because I mean the entire purpose of analytics is to ask more questions and to learn so that you can ask even more questions and so by definition in some sense you're you're you're constantly going to change and I think one of the big complaints about data warehousing is well it doesn't answer my questions that it would have otherwise answered 18 months ago um and the reality is that should be true because your organization is learning and the same data warehouse ought to be evolving and so you know um whether you're you using Google or you know we've got Partners at altto scale who who you know handle a lot of the infrastructure oriented challenges I think the reality is that you've got to you know that managing the infrastructure is one problem managing the human process of asking questions is another problem and they are different problems at different levels to Mike's exact point um you know we we see that all the time and I think you know cios are really concerned with the infrastructure problem and we see business people very much concerned with the human process of asking questions how how does that relate just so is that relate to the the cloud or not premise thing or is almost a separate issue I think there I think there are two separate issues I think that there's a you know it's it's um a question of you know look I want to go I want to you know buy a product now do I buy it online do I buy it uh in the store uh there are you know different reasons by which for which you would want to do one thing and then there's the question of which product am I buying well am I going to buy this TV or that TV and how do I select which TV we're going to be buying and what what how do I figure out the analysis to determine which are the is the right TV and those are two different problems at two different levels you know you might want to buy the same thing online as you might want to do in the store uh or you might want to buy something totally different and and they're different decision really for a company to make Mike are you an on premise Advocate are you you a totally a cloud guy when it comes to uh Big Data well you can almost guess my answer is gonna be hybrid James so which is maybe maybe hybrid is really the only answer to that question really but I mean it's the Practical I mean it's the Practical thing and that's what we see happening outside of Big Data just in Cloud infrastructure in general or any kind of clustered architecture is you have a baseline cost for your stable workload as as pointed out and that's probably going to end up on premise uh in the long run if someone has the ability to do that just for cost and then you have the the the unknown part the variable part the the new part the ability to do research and development ask new questions um burst uh on some seasonality or Cy Cycles uh into a cloud so you know I think it's a hybrid it's going to be a hybrid world uh of some kind which is a half and half answer uh but enabling that has been difficult for some time because of Big Data uh it's big it's you know you have to make again make sure you can secure it in both places now you don't want to replicate it and pay for storage in both places um you do have to be on top of your game to do a hybrid uh solution of any kind um and uh you know I think I think uh you know we mentioned we talked about security before but I just want to throw out the idea of encryption uh is being a really key enabler of this and being able to uh secure your data wherever it is um but at the end of the day I think people who say that service providers aren't as secures on premise are fooling themselves it's it's one of the ironies really because you know those those remote service providers their entire business is based on being secure and if you look at all the headlines everyone's whove been breached it it's it's it's like actually public is safer than than inhouse I I I'm I'm I'm completely in that camp if you look at the fact that and that you know I'm sure Dino can can can throw in some stats here but uh the breaches are are almost always uh and even even this goes back to even uh uh you know who's who steals from the secrets from the government not just Dash Madison but it's always human-based it's a human factor problem most of the time and um in your data center those are employees that can get disgruntled in in the public Cloud they don't know you from you know A to Z you and as long as you can create the isolation if you can encrypt your data and keep it keep it from mingling uh keep it from leaking uh uh on there I I I mean you look at look at the data center I I've walked into Data Centers I think we talked about this one time where people have the door propped open with a chair and a fan sitting on it and and and a cord from the fan running over there and say don't trip on the fan because blow the circuit breaker which shut server like what what this is not secure so oh yeah the secretary also the password is written on a sticky note Above This the screen too and and so you know you don't see I mean you go to you go to any qualified service provider location and you're going to see if not tank traps you're going to see ditches it can't even drive a truck up next to the building and it goes from there right it's it's it's it's just a different environment entirely so yeah what about uh and I think about things that you know limit Big Data I think it's one of the ironies is that it's very hard to get the human element out of big data and that you know we have these Solutions but we ourselves we humans are are you know contextualizing the hypothesis we're we're getting the insight back we're getting a series of answers back we've got to decide which of those answers is real or not isn't it kind of one of the ironies of big data is that the human the human element to that is is as important as as as ever do you know a take on that I think I I definitely agree with you at least at the current state um of of Big Data of course there are some really interesting things going on in the Deep Mind Labs that are uh self-educating processes but overall um human element is as important as ever because and here's the prevailance of SQL right you you have to have people who Define the hypotheses like you said you have to have people who question who POS the questions and then ask those questions and then Define what the questions are and uh this is why big data is so hard to deal with because it's very hard to ask those questions with a lot of current Technologies it's very hard to iterate all those questions and so um uh I think any big data expert should focus on offer its human element the best tools to iterate as quickly as possible right whether that's SQL whether that's programming languages and so forth um I think that's really important maybe it's also up to us as human beings I mean sometimes I think part of the problem is accepting the answers we get I mean maybe it's a very counterintuitive answer sometimes the data knows a lot that we don't know we almost don't want to think that way um sthan what what are you into this this sort of human Big Data conflict here is there is there a side you take on that one yeah I mean I think data is just at the end of the day a means right I mean it's it's just another source of information it's a source of information that's produced by computers but you know I think its promise is that it lacks bias right that unlike other sources of information an article on the web this webcast you know theoretically everything has some element of bias and data is this thing that's just pure observation purely seeing what other things have record and that's that's wonderful and great I I think that what we need is to make that source of information much more accessible so that you can validate the hypotheses that tinos are referring to so that you can actually um you know get in front of an idea and say look I can go gather the data to be able to get it and I don't think these ideas are as complicated as one PE most people you know consider them to be but I think the steps that you have to take in order to get the right data are themselves comp complicated so there's a fatigue Factor then most most people don't have the the intellectual sort of marathon capability to go through and stand up a server and get data put into that server and then you know build a schema and then build some Aggregates and then build an algorithm and all that stuff takes a lot of work and takes a lot of time and too many experts have to participate in that process and so I think the real human problem is how do you coordinate people so that the knowledge is available just in time um we spend our time thinking a lot about that that there's a lot of embedded knowledge inside of companies and really the big challenge is how do you ins surface that big embedded knowledge whether it's in a glal cloud system and a terata install you know in somebody's head how do you get that in front of somebody you know then that sort of where we where we exist Mike what do you think you know human versus Big Data is there a winner there um in the hell hell new Hell 9000 I think that the astronauts were just in out of the space that famous line sorry Dave I can't do that Dave so far ahead of its time I'm I'm thinking back to the Industrial Revolution in the Lites and so on and the you know the John Henry with his hammer versus the steam drill right um I think the steam drill wanted that one if I remember yeah certainly certainly remove more uh more coke and iron uh but you know the whole point of Big Data the whole point as sain was getting at and and I think is that you're minding it to get some value to make an informed decision you want to create a a more aligned strategy um you want to work in a context uh today with some pre some knowledge of future events or some prediction of future scenarios or events so uh it's it's not really one or the other um you still have to take that data and use it in order for it to have value and I'm thinking maybe we need a new metric you know it's not time to time to value or time to that side did we lose you mark the hell 9,000 oh there you are okay so heard me okay you're talking about a new a new metric I'm fascinated to know what the new metric is uh time time to uh time to uh intelligence rather than time to uh value or time to install right it's it's time it's the time it takes to get to the point where you've actually Incorporated the result into some actionable thing and and uh and realize some some return on your investment right um so uh you know that's that's I think what what it's not really an either or question right you're just you're you're trying to harness all this technology and data science and everything to help some human at some point do something better what about if you the three of you put on your sort of your your you know cry you look at your crystal balls and think you know what what's what's going to be happening with this three to five years from now will will we look back and and think gosh the idea that there was any confusion or complication with big data what a what a quaint idea we've come so far from that or what what are we going to think be thinking about in terms of Big Data say three to five years from now Tina what's your take on that I think the holistic intelligence is in gathering everything that's possibly can be gathered whether it's a website every single click whether it's iot every single thing that happens around every single device I think that really go is going to go mainstream because uh you've seen companies like you know the Google acquisition Nest that uh essentially transformed an industry that hasn't changed in almost 100 years right which is the ability to holistically understand what's going on um and so more and more companies are starting to figuring out that um holistic data Gathering just being greedy with data collection and being able to experiment with that data and to Loop that data analytics back into product is really going to differentiate these companies um and as more and more companies are getting into that space the more and more companies Embrace that level of thinking um the companies that are not doing that they're going to be in a little bit of trouble yeah I think about what you said reminds me of the way app that you know incorporates it's like social media and traffic so it's all a loop the information is is a real-time Loop that is helping drivers who would then turn you know put more information in so it's actually it's the the information almost feeds on itself in a real-time Loop uh SE your sense of uh five years out Big Data where where are we going with all this yeah I mean we we talk a lot about this theme of sort of everybody being an analyst I mean if we think about kind of a lot of the Solutions with software you know whether it's um you know the spreadsheet that made everybody sort of capable of being a financial analyst or if you think about things like Adobe Photoshop and everybody being you know sort of a photographic editor um you know I think I think we think that there's sort of this this notion that you know to do your job you will have to effectively become a data analyst and in a world where you know data is a fundamental part of decision-making and a fundamental part of how you think about what you what you do you know that's going to tax um the systems to be significantly easier significantly more transparent um and there's going to be a large element of collaboration that's involved in this because being able to validate things and test your hypotheses in part have to do with you know sort of dealing with data and Mining the data and figuring out what the you know distribution looks like but it's also about talking to your colleagues and figing out that you're coming up with the right hypotheses and so you know we see a lot around sort of broadening the skill set around who can become an analyst I think decrypting the data scientist is something that's going to happen by and large because ideally anybody who has a hypothesis should be able to as you might at Google you know sort of test it and validate it um and then beyond that I think we see sort of a world where the tooling is going to get just simpler and simpler so I don't think we're you know the technologies that exist today are only going to get stronger and I think we're going to see more and more technology you know really to the point kind of the Moors law I mean it's just going to become you know more scalably easy to test the hypothesis to assemble this site yeah it reminds me I saw this uh video recently of Gartner and the Gartner research a firm and they were talking about they had taken big data off uh off of their hype cycle because big data is is is being broken up and being is being incorporated into so many different elements that it is Standalone but it but it's also Incorporated so many elements it's it's not totally Standalone and and sort of on the F flip side of that is what you're saying is we all need to be data scientists there there isn't just a small tribe of people known as data scientists in the future you know many more of us will need to be data data scientists yeah pleas Mike's point I mean there's this notion of sort of he was thinking I think time to you know I'd say time to Insight is the appropriate metric and I think in a world where if there is only a small amount of data scientist that's a pretty sad world that we live in I think I think that you know um I think there are visions of the world where are controlled by computers there are visions of the world that are controlled by humans and you know I think we just got to make more people smarter in order to get to the world that I think we all live in Mike are you ready to make more people smarter where where are we going at five years from now it's again torn a little bit I mean I I I see great strides in machine learning coming down to the masses you know uh things where you can push a button and it'll tell you what the best algorithm it is to use on your data set and and what uh patterns can be found in there before you even think to ask the question uh which which is some very interesting stuff in in the days getting bigger uh at the same time I see um The Internet of Things coming along as you mentioned nest and and some other stuff when everything's producing data there's going to be still a lot more data than what we think of as big today and so there's going to be this race I think in scale and it's possible that the service providers like Google are are only are going to Prevail here simply because of that explosion in scale there's no way people can do it on premise and it's and and so that's you know a good bet there for Google um but I you know I I'd like to say that big data is really only to be successful uh not just when Gartner drops it but when we don't think of Big Data as big we just think of that as data analysis of course it's large uh and and we we have that in under our tool belt uh as a matter of course um and I go into Excel and my Excel spreadsheet connects to some cloud data bank and and and whatever I'm doing there is working on Big Data uh with the same simplicity so I I I want to have had doop on my smartphone is that is that coming soon um or or some head doop front-facing app at any rate I I appreciate I'm sure you can get Google with your smartphone I'm sure that'll work you think it's possible okay I'll try that do with friends uh thank you very much for your expertise the three of you was great you know I I certainly learned something and I will send you the link and we can all tweet it uh thank you very much thanks thank you guys thanks
This transcript was generated automatically from the
video's captions and may contain errors.
