 Bagaimana perasaan anda? Ada beberapa perkataan yang saya benar-benar terasa. So data-driven adalah satu. Big data. Digital transformation. Cloud. Semua orang bercakap tentang ini. Like it's a hell of a big deal. There are a lot of young people here. Any of you are... If you are 32 years old or younger, there's a specific reason for it. Because I've been in the data business for 32 years. I've been doing data-driven. So if you are in the company, you say this is really really cool. You are 32 years too late. This is not new. Seriously this is not new. The good news is because it's not new, the practice is very very established. All the problems that you can possibly face has been solved. It is published. It is proven. You don't have to spend 2 cents on research. You just have to do the damn thing. The trouble no one takes the time to do the stuff. So the practice and technology has changed. They are new tools. Data becomes a lot more volumous. But literally the practice is proven. If anybody takes the time to Google something, you will be the smartest guy on earth. Investors will come to you. So today we're just going to talk about how this thing is not new. And how people think it's this small thing. It is not this thing. It is this thing. And so the data scientists will not save your skin. So you need to put things into perspective and then hopefully you can say actually we've got all the moving parts. We just need to figure out how to do this thing right. So this is the actual title. Are you rich? Let's talk about expectations. If you are hiring, say about 15-20 years ago, hiring, say middle management, you see a resume and one of the skill sets there, someone says Microsoft Office Applications. It used to be very impressive. If you see a CV today and someone says Microsoft Office Application, you would really want to meet the guy because you will see what an idiot looks like. So expectations that people expect you to know certain things. It's like if I'm in Seoul and I want to have barbecue, it's already in Korea. So there are certain expectations that people expect you to have. So people expect companies. So if you are starting up a new startup, there are certain things that people expect you to know. One of them is things like this. People expect you to know how to consume data. People expect you to... So when I hire a director of marketing today, I expect the guy to understand what marketing metrics are. I don't expect him to know how to do it. But if he looks at me and say, can you tell me what KPIs are good for marketing? I'm going to slap his face. This is an expectation. So upfront, if you are going to this area, you need to make sure that you already have the capability because customers expect you to know. If you're going to give me a form to fill up and you're going to profile me 32 different ways, and then the day you still send me crap afterwards, I say, but I just told you 32 things about me. I can't you at least have the courtesy to figure out that I don't like ice cream or God knows what. So in our everyday life, we go to this, people ask you, so we are very obliging. We fill up the form, we tell them intimate details, and nothing has changed. So after a while, customers will loofah. Investors will look at you and say, these guys are idiots. Yes, they use the right words, but they can't do basic things. So they don't understand the point of impact. So a lot of, for example, so definitions also change. For example, for a long time, she mentioned digital a couple of times, everybody talks about, if you look at media, you look at training, everybody talks about digital marketing. Okay, today, is there any marketing without digital? It's just called marketing. The digital is assumed. Just like the Korean barbecue, and the french fries in Paris. So if you will go out there and say, I'm a digital marketing expert, what the hell are you talking about? Because the whole world is digital. Okay, the definition of digital itself has changed. For the longest time, when people say digital marketing means we move the digital channels. It used to be a paper, now it's electronic, it's digital. I send you an EDM. You do know, that's not the current definition of digital. Because if you don't, then today, digital means data. If there's no data, it is not digital. The fact is on paper, but somehow or other, data plays a role. So let's do this test. And if you hear the word, people talk about transformation, digital transformation, the definition is very simple. It's basically your company needs to learn how to really make use of data. In every sense of the word, it means also, so when people talk about data, data processing, you always think about big data analyses. People don't think about, if you don't collect the stuff, you can't do analysis. There are a lot of companies out there, big fat companies with hundreds of data scientists, millions of dollars worth of software, and nothing is happening because they just realize that I've got the factory, I've got manufacturing, I don't have raw materials. That's a huge problem. And then now with, you know, data protection, PDPA, there are rules, in terms of, there's nothing that says you cannot collect, but there are proper ways of collecting. So unless you understand the whole ecosystem, the last mile which the data scientist with the PhD, he's not going to help you because there's nothing for him to play. Or the data is not right. It doesn't represent the real world. So we all take bus. Here's the test. So we all know the clear channel panel at the bus stops. There's a whole bunch of them right now, but there's you have the the full electronic one. You have the the paper ones with the NFC thing at the side. Okay, you click it or you touch it. There's something that loads the website. I've got no sweat. You have the the paper one. Have you seen this? So there's the paper one with the camera. There's this sign which nobody notices. So if you're standing there looking at the stuff, please don't pick your nose or scratch yourself because you'll show up on YouTube somehow. And then you have the new ones to just launch they call it a play. It's a full touchscreen, fully interactive. Okay, which one of these is the least digital? Anybody want to sorry? C, paper with camera. Remember my definition of digital? Data. Yes. A is a dumb panel. It is, you talk about dumb computers and dumb TVs. Now you've got smart TVs. This is a dumb panel. It is electronic. It has no awareness. It has got no freaking idea there's someone looking at it and scratching himself. Okay? This one has and he can do analytics and say typically old man with bald hair like to scratch himself. God knows what. That's inside for you. So you need to understand. So the definition of digital starts from point of awareness and sensing. So if your organization by design is not designed to be aware and sensing meaning collect data then everything down the road is a problem. This goes say if you're building a new organization say if you're a startup I'm building an organization from scratch. If you're the HR guy first thing you want to figure out is okay while we're building this organization how am I going to do performance management not just of people but process governance and everything else because at some point if you think we really need to do that then you start to you need to start to organize your company infrastructure workflow processes from the ground up to be sensing there has to be a way to collect data so you know if people are not following rules governance is broken performance is an issue my next door is dating god knows what because without it it's going to be there you go you end up with okay we've got to do self assessment because there's no other way and then someone yeah but it's the next best thing okay it's a cop out so definition of digital needs to be recalibrated because until you do it everybody is going to say yeah I'm a digital marketer because I do EDM it's electronic but there's no sensing I mean we all look at augmented reality you know the IKEA catalog from last year the piece of paper with augmented okay that's paper that's digital because every time you scan it's sensing it knows who's scanning it knows where you are it knows what phone you have and god knows what else you've profile the amount of data whether it what it does with it it's a different point the fact is it's a sensing ecosystem versus the digital catalog where you just stare didn't do nothing so let's talk about computer so if you're going to build a data driven organization so what does it take so there are a couple of things that needs to exist so we just talked about it it has to be a sensing organization so a data driven organization has to start from I am sensing all the time there's some awareness somewhere so the sensing creates data so I have to ingest the data into my systems whatever you whatever you call systems but think about this if your if your smartphone is your alarm clock for example then it's the first thing you do is touch your phone that's a data point if you haven't figured it out where it goes where it goes to apple google or idea god knows but it knows that some of us have to touch the phone 23x before we wake up that's a data point it has to be ingested now we're talking about smart nation we're talking about national sensor network talking about smart grid you do know thing about water knows that you people don't bathe in the morning which is very bad for you they know those people who stumble to the toilets in the dark without switching out of light because you are lazy say these are these are sensors so that's the whole idea here so without the sensors then nothing happens but with the sensors capability world-world develop that's why the first thing that's actually being put in place as part of smart nation is actually the national sensor network because I figured the big data things not going to work without the wrong materials at some point this is what everybody talks about so everybody's hyped up here okay we have all the analytics we process the data we do analysis then what so we have really really cool power points it will look as and then we all go home and then business as usual so nothing so nothing in the world changes so this whole thing with data driven is unless something in the world changes there's no ROI because up to this point we're all spending money this is not cheap stuff and there's no impact there has to be storytelling someone's going to look at this data and figure ha every time the MRT breaks down there's this little spike here this looks like a business opportunity let's pray it breaks down tomorrow and test see what happens and then you react so someone has to translate a data point into some sort of story which makes sense to the business and then someone has to actually transform your business something in the business changes and then magic happens up to this point without this literally nothing happens no one wants to talk about the G word governance because without governance this is out of control if you think about this because now especially if you are getting data from all over the place so I've got public channels I've got customers I'm doing 300 surveys a day so we have very creative so every survey is a different question so now I've got all these data points then how do we say so you describe big a certain way so we all have big businesses your big is bigger than my big so it doesn't make sense so governance now sets rules policies standards so at least when you start to I'm sure you've been in meetings if you are at some point in the corporate world someone shows a report we argue about the sanity of the data no one argues about the fact okay this business is not doing well so when you get data from you pull out on Thursday you should pull out Friday morning last night there was a batch no this spreadsheet is better that's what's going to happen if you don't have governance that means the business doesn't benefit it's all griping about their quality this is a multi million dollar problem a lot of companies get stuck there and at some point spreadsheet you all know what to do just pivot it this way and then life goes on so let's talk about sensing if you take a bus you see this poster I couldn't find so the original version of this poster basically tells you that if you tap out too early you're breaking the law because you're cheating the bus company they moved on now it's basically they want you to tap out in time not because of money they want the data you can see there's trust so someone has figured this thing out actually we don't need the money we want the data because I can monetize this stuff if nothing else productivity is scheduling you see transformation so this is a sign of someone's figuring it out if you're a pokemon hunter the whole uber grab these are all sensors in disguise like play things by the way without this nothing has happened if you don't play the pokemon thing there's no data for them to I'm just making billions selling the fact that certain type of people will go out into the middle of the ocean to catch some stupid thing that doesn't exist if you are the marketing guy and say will it be cool if people will queue up for 3 nights for product launch you really want to know who are the idiots that will do it the guy who's out at the ocean catching the pokemon will do it I will pay for those people that's the business model here the same with the grab I mean what's interesting out of this there is so uber and visa has got a project going on I think it's in Thailand so what they do is they look at the data from bookings and they figured out that if you're looking for an uber after 6pm and you're not going home again through the app I know where you live if you're not heading home there's a 90 over chance that you're looking for food it's so trivial it's there so now if you're out there looking for uber it's after 6pm and my destination is not home they're going to serve you content about food destinations where you're headed and then you can figure out that's a huge chunk of change so we all know uber is bleeding money from the car hire thing but they will make this thing off when someone figures out that all this data is monetizable to a different way so this is but if you're not sensing and data cost money it's not free so they're willing to pay for this because now I'll make it somewhere else so what is sensing capability a lot of people think it's technology true and there is a technology play but at the end of the day you think about all the play things that we participate in if the customer experience design is sucky then you won't play that means the sensing mechanism doesn't work there's no data so if you're out there say yes back to the HR thing so if you're HR guy willing to build this thing if you're expecting your employees to every day go in and fill out a survey people are going to hate you at some point they were live so now your data points are off so it's not going to work so yes there's a technology play but customer experience design so this will lock up to get it wrong they have the technology part right but you don't understand customer experience and so people don't play or people get fatig or people give up and so now the data gets compromised or not at all and so everything down the road doesn't work so first part get the sensing right and these are some of the key capabilities so anybody here runs so this is not some financial thing this is what my running watch sees when I run a marathon it's like it knows that others will get a heart attack really soon so we talk about ingestion that little thing that you wear potentially with a heart rate monitor potentially with a transponder on your feet it ingests and generates impressive data but that's what it is we look at these things like ya really cool next slide so part of the I've sense so my watch is a sensor so I'm addicted I'm a sensing addict for me tragedy is all dressed up to go for a run and realize that oh hell my watch battery is low for some reason I cannot run without my running watch I have no idea why it's not that I look at the data but I need to collect the damn thing just for the hell of it and then save it and then delete but at least I got a sensing part right so let me just tell you what this data is so this is what my running watch is it looks very complicated but it's very trivial but there are implications this thing here so the watch actually has built into it a pressure barometer so it's tracking my route pressure changes so basically I just ran up a big hill oops this is my heart rate over the period this is my speed over the period this is cadence even tracks the number of times my feet hits the ground and the distance basically my stride length and there's a whole lot of stuff so from a sensing point of view it's quite amazing technology and this is not even new this is like a few years ago the new ones do a whole lot more what you do with it is the magic because now it's just it's just cool powerpoint so what does it take to ingest data so data management is a huge one of course because without data management it's all going to be rubbish just like no my big is bigger than your big then there's no benchmarking but there are specific skills so again if you just google data scientists there are a lot of these are hard skills there are a lot of mathematics a lot of statistics a lot of technical programming so these are guys who cut code the geeky guys with a PhD they build amazing product they spin the data so many ways and they visualise the stuff and you look at it and they tell you it's very high you say yes we can tell and it ends there so for it to move from today very high that's an interesting business proposition two things need to happen this is two points of epiphany Allah Singapore style someone has to ask why like that question if you think about it if you just look at and everybody just accept it and it goes all it needs is the first idiot in the room to say why like that because now it's going to force a reaction someone is going to explain storytelling I need to tell you why it's like that how dumb it is how good it is cause and effect and then someone should hopefully look at that and say then how because now the storytelling moves to recommendation and decisions without this you just have fascinating power point and unfortunately today that's what you have you have very very expensive fascinating power point because people don't take the effort so it means now storytelling the data scientist has done his job he's produced this unless he's part of the business and beta he has no idea what just happened in the business but he can show you the point he can spin the data and hopefully say there's correlation between this and this but so what so like in my case when I say I just ran up bloody big hill I look at my heart rate it looks fairly consistent you're saying this guy just ran up big hill he should be like dying of a heart attack at some point so why like that should be the first question so far I can tell you because I'm not very motivated or I'm probably following this hot chick in a short skirt and she was really walking so I figured there's no good reason for me to run any faster God knows what but that's storytelling for you and so if I had a coach she's going to say okay you really need to take this seriously then how happens she's going to say okay we need to put you on performance target you gotta be focused you gotta do this, you gotta do that you see the difference it used to be fascinating chart now there is a reason same thing here this is my speed and cadence you see it's very consistent every 2.5k my speed is very consistent every 2.5k I sort of stop if you run marathon drink points so you run with water it will spill so I'm very practical don't waste water I stop and drink and then I continue I figured I've got 5 freaking hours to finish this thing there's no hurry and then at some point it looks like I'm having a high attack so again why like that explanation never train lah you train 10k, run 42k race at some point it's not sustainable anymore so now that's storytelling so again coach say then how so okay structured training God knows whatever the mix so this is a skill set in itself this is very very missing this is not something that you can outsource to your data scientist so usually oh yeah before I get that anybody here of thick data no one talks about this stuff anybody talks about big data I have a million records I look at it as I say according to this thing people with brown hair do oh yeah very clinical thick data is just the opposite I have very little sample I have very deep insight this is where the emotional connection comes in so google this lady Trisha Wang she she does a lot of work she is an anthropologist technographics and basically she's the one that told Nokia that you guys are going to die if you don't be careful about her she spent like a year as a street vendor selling candies in Vietnam while doing her research shed 100 records only actually we did she told Nokia I know you got a million records sample size but this is how people buy phones and Nokia told her you got to be kidding now she's telling Nokia yeah you got to be kidding so thick data is not talked about a whole lot but this is where it brings a lot of softer emotional connections the context to what big data shows you so if you look at this stuff again, look at big data but also consider thick data and so it's a different skill set if you look at this a lot of soft skills a lot of communications you have to be able to explain very very complex things into relevant business ideas for that specific business so if you're selling this thing to a CFO it's very different if you're selling this thing to a CMO very different from selling it to the operations director in the same company so usually local expert if you have a willing capable local expertise this is where they come in because they bring to them business legacy so the data scientist prepares the data the local expert looks and they say ah, yeah, yeah this is what usually happens when it rains and this and this or it's a problem with supplier these are tell-tale science that's valuable a lot of this a lot of this don't show up in data points so this is the sales person essentially and then this is where something needs to happen transformation you like my charts transformation is when you've heard the story someone's going to say okay, we need to do this thing differently we need to stop doing that stuff we need to stop paying people this much we need to pay these guys god knows what so somewhere in there certain competencies breakthrough thinking is one of them people who think outside the box potentially people like you guys young people people who don't have walls around them and then ideas people who have I've been accused of being very irresponsible with my ideas we need people like that because once you say cannot take back that means no, it becomes somebody's problem but you need people like that because now it creates a very different organisational structure of course there is a lot of change management so I didn't really say what I do so I'm not startup like you guys I mean I work for myself I left a couple of worlds but it's a go I'm a management consultant I run a one person management consulting practice I do transformation consulting my clients are the biggest companies on earth so Apple is a client IBM, Microsoft and a lot of time when they invest in startups they try represent the investor so they throw me to the startups my job description is usually very simple adult supervision that's what my clients say you are adult supervision so I sort of meant we introduce enterprise capabilities to start up I would say if you behave like a startup people will treat you like a startup so start behaving like a big boy so we introduce all the rigor governance structure but you're still startup so you have agility but you have there are times when and you tell you tomorrow cannot wear jeans because we're meeting people who give us $20 million so everything has to match so I'm that guy so usually I sit on the board of couple of startups and literally I work with them and then sometimes I convey their message back to the enterprise sometimes I convey the enterprise message back to them sometimes I'm the guy who says I think we should not cut our losses and kill these bugger sir unfortunately that's part of the thing so I do a lot of this change management and then I push people down this path culture experiment so yes you behave like big boys but you have agility to do this so this is a culture if you have this then the data thing works because if you don't experiment because the whole idea with a lot of data driven sometimes the recommendations that come out of the data is counter intuitive that's why we don't do this thing according to the data we should do this but it sounds like a bad idea are you willing to do it because if you trust the data if you trust your methodology then it should work if you think I don't trust the thing then why do you do it in the first place or should we hedge our bets so that's a a competency in itself and then finally governance I found this by mistake it's really really cool Governance in a simple sense is rules simple illustration say I'm not saying this is going to happen say if Facebook one Friday decides that we will start to rank Facebook users by how funny their friends think they are so Facebook is not going to evaluate you Facebook is going to read your comments the comments of friends so when people say ha ha ha ha LOL so 4 ha means something 3 ha so at some point Facebook needs to tell all of us there are now rules you cannot anyhow ha ha ha because it will cock up our measurements of course we will say what a stupid idea then if someone gets really too smart and say now we will have a ha button I miss you if you want to say ha ha ha you click 3 times so you just see what happen Governance well intended customer experience design really sucky which means sensing is going to fail no data the business analytics is just not going to work now someone is going to say so he got nothing to do with the technology think about this someone just didn't think through the customer experience design over investment in technology and every blind thing you forget about the human being because I have no time to follow your rules I say ha ha ha as much as I want so this is a simple example of governance so unless you have robust governance a lot of your data driven is going to be very subjective credibility the moment people lose credibility with your practice no one is going to trust any more power point that comes out of you or the Excel chart or anything else so this may be a bit of an overkill but it doesn't so you understand the whole concept of governance risk management and compliance GRC if you have internal auditor compliance person you don't need one of those guys but it helps to have someone who is capable in some of those things so I am a certified GRC professional I don't practice this stuff as a consulting thing but I bring it into the data practice so when we deliver information strategy is 100% GRC compliant so the governance structure that we put in will pass through any internal audit compliance officer any regulator it makes a big difference so you want to think about that if you think your data practice is going to support critical businesses then you want to make sure that the governance is robust so people don't question the background so now we've figured out we've got skills culture is huge as a consultant so I love looking at maturity models because it's not about how much you like it because it's your company mature enough to deal with this new thing this is huge because if the company as a culture is not mature and you really want to do this thing then you got to stick your neck out really really long because someone is going to get killed so some sort of there's so many of this stuff I like this one because it's granular enough it's simple so if you are in an existing company if today you guys people struggle agreeing about which spreadsheet should we use then you know recommendation model is not going to work because no one trust this thing this whole thing will be optimization automated AI not going to work because you can't even agree on the spreadsheet so again just having the maturity model it's an idea of how far you are and if you're the guy driving and saying how difficult your life is going to be it's a damn good sanity check some again not rocket science but fairly obvious attributes as an organization you want to be curious so you're always looking for evidence you're always proving for example if you're a marketer someone runs a campaign campaigning go very well zero it's a valid answer it's not a good answer but it's a valid answer I don't know it's not a valid answer simple things like that so if you cannot measure we have a problem because you'll do it again but if you know it doesn't work okay just stop doing the stuff you got to be active so looking around looking for answers a lot of discovery understand uncertainties because there will always be risk this is what I just talked about the recommendations from the data could be counterintuitive so how brave are you or how risk averse or how do we mitigate in case we get it wrong agile bit transparent this last one there's a lot of maturity here you want to say if you got too much IQ and too little EQ then you're an asshole this is dead just because you have the data sometimes you want to be careful with it so you know great power comes with great responsibility who say that Spiderman like so but this attribute hopefully gives you a nice balance you know when to act when to mitigate and when to I think we just go with the flow and then save it for another battle sometime if not you'll die too fast I'm finishing and then finally hopefully you appreciate this is not a one thing it is an ecosystem it's a huge ecosystem let's just build this out and we talk about this so if you are the end user company and you want to do this then you need to have access you don't have to own all these capabilities but you need to have access to the various capabilities whether it's a vendor best case you know you got one guy who can do all this stuff but usually difficult so it could be a vendor it could be academia but unless you have the full ecosystem it is not going to work it comes from information architecture upfront design to connect we talk about governance data management so there's a lot of hygiene factor that needs to be put in there because if not again your data becomes very questionable then we all know this all the data sciences and analytics the storytelling talked about this and then the transformation if you are startup tech startup and you try to build some of these tools you should know by now this is bloody crowded the whole world is there you actually don't want to be there because most people can't tell the difference this one not bad there's a lot of old school people assessment institute stuff like that everywhere else it's actually very vacant nobody does this stuff if azar can do this for the biggest company on earth with one person and I mix it up while I'm doing the AV thing you can so this is from an opportunity point of view this is what the data driven economy needs is automation with information architecture governance automation it's a huge one because people who do this work hate their jobs so if you can make it more efficient they will love you the internal auditors the compliance people will love you storytelling getting very popular there's a lot of training available for business storytelling it's huge this is one if you are a good technical guy if you can do this well this it's a good transition to get you out of technology and in front of the business people because a lot of technical people have a hard time telling stories then the transformation again if you have if you've been in the corporate world for the longest time you've been to hell and bad on a lot of stuff this is a good time to get paid for this pain that you put to and then finally last words very simple 3 things awareness sensing people forget the starting point for being there you have to be a sensing organization without this nothing happens then of course obviously you have to be learning so what you sense you got data then you learn stuff but this is a money slide it is about applied analytics you can do all the discovery that you want if it's not applied you're not solving a problem somewhere there's no money in it and so some of these things are very very simple if you look around look at the masses SMEs look at the problems that SMEs need to be solved the very very simple problems that can potentially be solved with a spreadsheet with 7 columns without a pivot table but because it's so simple the guys with a PhD say this is a pedestrian problem I don't want to solve it that's where the money is so you are out there building the software to launch a space shuttle the 3 companies that need to launch the space shuttle already have the solution the 3 million SMEs that can't even figure out the expense report they need the solution so you want to take a look at that and again so applied analytics is where the money is if you can apply even simple things to simple problems but you can solve it then people will buy and then from there you got food in the door you can make it as complicated as you want and basically I'm done thanks so you have questions for us any questions? my question we want to take one or two and we've got one more speaker to go so please go ahead okay my question is lots of people now are talking about big data but we all don't hear about small data no no no on the newspapers or on the mass media we're all talking about big data yes but why is it that nobody talks about big data which means that all the data can be can be researched doubt did really really meant for the person person person very personalised and then translate it into application for business or let the people take action because if you look at it again this is my point of view if you look at the ecosystem as the word big says this is a lot of scale I can make a lot of money amassing that and the complexity but it's thick data literally people are talking about small numbers a sample of a hundred and with 3,000 data points but a sample of a hundred it's like I survey a hundred people and I collect like a few thousand data points and that's thick the statistics will tell you about budget of error this is like ten standard deviations of or some ridiculous number whereas if I have a sample of 3 million records so unfortunately that's where the scientific or mathematical world goes so the only people who actually preach thick data in a big way are literally so anybody who got anthropology training you'll be surprised how in demand you guys are for the longest time you couldn't get a job if you're a PhD I got to go back to either join the museum pretend I'm Indiana Jones I teach at university today big data companies also all the analytics companies like the SAS Institutes of the World will hire anthropologists because you are good at this and I mean anthropology is about the study of civilizations things like social media analytics when we say you suck if you went in social media and you type my vacuum sucks that's an interesting context some software will say positive because the vacuum are supposed to suck I'm not kidding so it takes an anthropology to understand context civilization and language to say this is negative in this context this is positive because now we have AI someone needs to tell the robot under these circumstances it's not a good word under another circumstances it's performance measurement it's actually good so this is where a lot of thick data practices come from the anthropologists the ethnographic show up and then the smart companies bring them in and marry them or pair them off with the data scientist so now they present balance so again not necessarily the politically correct answer but that's how we look at it and that's how we tend to apply them you can do one more okay cool thanks