 Thank you very much. Thank you for the wonderful introduction. Exactly how I wrote it So my name is Owen Xia and I work for a local Digital marketing agency not a design house. So I don't really consider myself as a designer not anymore I used to consider myself as a UX design or web designer But I stopped calling me myself that mainly because of three reasons the first and the most important reason is that I cut my long hair I'm not saying that all designers should have long hair, but I certainly felt a lot more credible when I looked like that and the second reason is that Nowadays my job scope entails a lot of web analytics a lot of digital marketing in advertising a lot of Search engine optimization right whenever I go for client meetings if I feel that the client is kind of difficult to handle I just start throwing out all those acronyms DMP SSP DSP Which stands for accelerated multiple pages supply site management supply site platform and the data management platform and all that But when I was a designer, I find it relatively difficult to handle clients I usually run into very tough clients and I think all of them can be divided into two groups Here are the two groups The first is those I call micromanagers. They will tell you something like you know feel free to be creative and After you submit that your design they say oh, and I love your design, but can we just make a few minor changes? They won't take much time or yours at all and then you get something like that And the second group is kind of the opposite of the first group because the first group kind of know what they wanted And the second group really has very little idea what they want at the start of the project. They will tell you I'm not sure what I want, but I'll know it when I see it and when you submit your design. They will say You know what? I love your design, but can we make it pop? I never figure out what that meant because it's really difficult to make a page pop so based on my experience the more client is involved in the design the more expensive it's going to be in terms of Investment on your part, you know time and money and strength right if I design everything you approve That's relatively easy to handle and if you want to involve your entire staff and Assist me in the design and go through all that 80 iterations. That's going to double triple even quadruple the cost nice So I usually I used to sit in a client meeting and listening to their all their feedbacks and again Until I want to shoot myself But here's the thing when it comes to design Everybody seems to have a lot of opinions Which is why when I said if I go for search engine optimization meetings I'll just talk about the latest Google algorithm and change right and if they can keep up with me I really don't know why I would be I'm there But for design whenever you put out a design you always get ample amount of feedback everybody has opinions, right so But opinions can be dangerous Since we're talking about opinions, let me get your opinion on the following website This is a car leasing website. It's called Lin's car comm It's started by this Chinese lady called Lynn and It's a car leasing website. Can I have a show of hand? How many of you like the site? Right, so we have one person how many of you don't think it has a good user experience and you don't like the side Okay, the rest of you don't have an opinion or are you just won't raise your hand Okay, so for those who do nothing has a good user experience. What would you change? Give me some response. What do you say everything? so it's it's very 1999 right and For the slightly older crowd in the room if you recall that Yahoo used to have this geo cities section where Everybody can create their own site. Well, if you're nodding, you just revealed to me your age So they used to use these really bright color pure colors and fonts are all over the place, which is kind of this If we want to use a nicer word is a retro design. Okay now How about this website? It's a it's a dating website. Don't ask me how I know it and don't ask me whether I've used it But it's a dating website. It's called plenty of fish calm. Yes, very apt So what do you think of the design? How many people like it? How many people would actually use it again? Don't ask me whether I have used it or not. How many people would actually use it? Nobody Okay What would you change? Would you like would you per how many people do you think that they actually need a redesign? Okay, very good Now the third website I want to get your opinion on is this one How many people will redesign this website? I see some of you are hesitating because why? Because a lot of you already know about this website You know that craigslist is a wildly successful website and therefore you wouldn't change a thing But if you had never heard of this website Then you probably would tell me yes, I would like to change its website. It still looks very 1996 which it does Right so Craigslist is a pretty interesting website It was started by you guessed it Craig and Craig started this website in 1996 And over the past 20 years the site remains relatively unchanged and mind you 20 years in internet years. It's like a century. Even Coca-Cola has gone through multiple revisions of their can design right and if you think about it in 20 years They haven't really changed the style of their marketing. What does that tell you? Perhaps don't fix something if anything broken All right, but so redesign can actually be very costly very risky You might have also heard of this website dick.com in 2008 it was one of the most popular websites on the internet has received huge amount of traffic in 2008 It was actually worth a hundred and sixty million dollars and two years later in 2010. They decided it was a time for revamp So what did they do they redesigned the website within the first month of redesign the traffic dropped 26% and then it went down here from there in 2012 which is two months later It was sold for guess how much It was sold for half a million So it went from one hundred sixty million dollars to half a million now I'm not saying that redesigned was the sole culprit of this There were other factors like the rise of Facebook, but I'm sure redesigned played an important part It's a redesign can be very risky Now you as a smart audience in the group already know where I'm going with this When it comes to opinions, there is more to it than meets the eye. This is Lin's car.com Business week calls it one of the best websites on the internet You know why because the whole retro so I actually watched this BBC interview of this this lady her Rationale for creating such a website is that every website was looking the same, you know Now this was a slider carousel then several sections, right? So she really wanted to attract the attention of the public was more of a public public stunt and it worked Every website was reporting this most people the audience that they attracted actually already expected to see something like this right, so it was actually part of the traffic strategy and if you Actually get past the initial reaction of what the hell is going on here? You actually discover it has all the persuasive device necessary to convert audience for example Here he says I'm Lynn. I'm from dragons den dragons then how many people you know shark tank It's this American reality show where entrepreneurs go on to TV and you know try to get investment from vcs dragons Then is UK's equivalent of that Somebody was actually trying to buy by her out and she said no I said that social proof and he says in 2010. I rented 35 million of cars so and so forth and he has these assurance device That says what happened when I press order now, you know, don't worry. You won't be charged. You do not commit He has all the right call to action. It even has Life support and all these different elements are scattered around the page I'm saying if you do not leave if you actually go through the page a lot of them will convert the website Receives millions of visitors every year and it's one of the most successful car leasing website on the internet How about the next one the site I may or may not have used? plenty of fish two years ago in 2015 the site got acquired For five hundred seventy five million dollars again. This is very interesting story. So the creator Was somebody called Marcus Freind and he was a developer in 2003 there was this new programming language called ASP net I think it was from Microsoft right as some of you would have heard of it so Marcus Freind as a developer he thought it would really good if he could add one line that says Proficient in ASP net on his resume, but he didn't really want to read books and stuff So he thought why not create a website as practice I so he just slapped on a basic website and to his surprise people were just signing up. I mean and Then before he knew it Two years later. He was making ten thousand dollars every day from Google AdSense If you haven't heard of Google AdSense is that little revenue program where you make about two dollars for every thousand impressions He was making ten thousand dollars every day now he was it got acquired for five hundred seventy five millions and the interesting thing is that because He was he wasn't really in the whole startup game by the time he figured out what VC meant venture capitalist He was he had more cash than he would ever be able to spend so he owned a hundred percent of the company till the day he got acquired And if you go to plenty of fish I'll come I think right now is pof.com You will still see a very retro design a very basic design. I so what I'm trying to say here I'm trying to say that opinions can be very dangerous If I go to a client meeting and our discussion becomes your opinions versus mine It's not very productive because I can't really tell which one is correct Either one could be correct, right? We can joke about clients all day long because I enjoy doing that But the ironic and sometimes scary thing is your client may very well be correct And you may very well be wrong. I just don't know which one again You know where I'm going with this since the title of my my presentation is analytics, right? So instead of letting opinions drive Your design decisions you really should be using what data but here's the thing most designs are driven by opinions Although we don't call it opinions. We call it experience That's just another word for opinions What's the second thing that people usually base their design decisions on it's so-called best practices Which is why nowadays you see all these similar? Bootstrapping website some of you would know the or use the CSS framework bootstrap right they pretty much look all the same in fact if you Google every freaking boots bootstrap bootstrap site ever you will see the standard bootstrap layout and Also nowadays we you see a lot of websites with the first screen above the foot It's all the slider. We call it carousel With the left arrow right arrow then some dots at the bottom which nobody clicks So you're doing that only because everybody else is doing that. That's industry best practices The third one is competition competitors again often when I go for client meetings I would hear things like, you know our competitors are ABC a is doing this. I really like it B is doing this great See can we combine all of them? I? Don't know if we can combine all of them This would work if your competitors actually know what they're doing Which is not always a case right if you blindly follow them You're gonna it's a classic case of the blind leading the blind And the fourth one is a little bit tricky. It's called hippo a hippo stands for highest paid person's opinion Which happens more often than you think? Especially if you're dealing with the boss on the other side Right, although we tend to think this we will speak our mind But in the actual in the in the actual boardroom whenever the boss opens his mouth the boss says I like blue Like I don't like yellow the website will probably become blue like the highest person's paid a person's opinion It's also very dangerous Right These are all opinions, but you should really be using data to guide your design decisions Why because as they say numbers don't lie But the problem with numbers is that numbers can also be very ambiguous Right if I tell you right now that I don't have any girlfriend in other words my girlfriend count is zero That data point you may draw the conclusion that I am a very boring and unattractive guy Which you might be very correct, but I think you should look at other factors For example, I haven't told you whether I'm married or not In fact, I am married and I have a loving wife That's all of a sudden that data point isn't so negative at all right it means that I am Law you and faithful like all the husbands in the room. I hope so When it comes to data, we'd like to talk about the framework of data inside and actions Data themselves are not worth much data. Do not have intrinsic value Or we can't really do anything about data the insights derived from the big data is what we want But then again if you derive a piece of insight you can't do anything about it. Let's say if I tell you that 56% of your users use sing-tel or use a predict startup as ISP, right? You can't really do anything about it That's a piece of insight which is not actionable So in the context analytics, that's not very useful what we want are Actionable insights derived from data and that's what analytics means to me So I think there are four main areas that analytics can help you with the first area Very simple is decision-making right if you say which version is better a version or B version here's an example Here's a website with the photo of a cute baby Now if I do not show you the heat map then I will conduct the poll I say how many of you like a version a and version B probably 50 50 Then it's opinion your opinion versus somebody else's opinion But if I show you this heat map by the way, if you do not know what heat map means It actually means a visual representation of how much attention your users are paying to every particular Element of the page so if I'm constantly looking at say top right top left corner and Everybody else does so then that particular area will be represented by a hot color such as red or yellow And if I'm not totally paying any attention to say the bottom right then that Screen part of the screen is usually marked as a cold color such as blue or white right so this is a heat map And it tracks the movement of the pupil That's the most advanced form of heat map tracking the next best thing is the track mouse click, right? But anyway, so this is a heat map. That's data inside action. So what is the data here? Can somebody tell me what you can? What do you see in this visual representation? What are the part of the screen that visitors pay most attention to the face? Actually the eyes now as humans thousands of years of evolution or you know basic social etiquette Dictates that we should look at the other person when he's talking to us So we usually instinctively look at the other person in the eye Which is also true when it comes to web design if I see somebody looking out of the screen at me I instinctively look at his eyes But we also have another instinct which is if I look at that then a lot of you will also look at that So instead of having the baby look at you where you will stare at the right at the baby and not the text You should let the baby look at the actual part of the screen where you want people to pay attention to which is the headline and the first paragraph I saw That is data Data tells us people pay attention to eyes. What's the inside the inside is? Line of sight should guide people to whichever part that it's most important on the screen Action well choose be and also in our future designs. We should also pay attention to line of sight Right well Obama does that Hillary Clinton does that so there's no reason not to do it Right so that is data inside action so for the following slide I always want to ask you to Discuss everything in this framework. Let's let's look at another example So Here's a website. How many people like it? This is no longer a trick question. They actually did redesign this website and got better response Okay, so if I look at this website, my first impression is that it's a little bit busy people things are all of the place a Lot of arrows a lot of pictures a lot of words so what we can try is Say a text-based website where you explain to me in detail what the benefits of this particular software is Or you could just use imagery you can slap on some really attractive image and Ask me to sign up, but I really don't know which might work. So Well, you let data to tell us right so they came up with two versions one text-based and one image-based in this case Which one one the image-based and I think now you can always rationalize this and derive insights. What is the inside here? data is text-based version lost to the image jury version and the interesting thing is the image jury version really doesn't have a lot of Arguments as a headline, but it's not explaining all the features and so on It's it's just very short the the text-based actually goes on and on all the information you need to make a decision Normally you would assume that the more information it gives user the more likely the user would convert not in this case What is the inside here? One of the possible insights is that Purchase is very different from trial. This is a software trial in order for me to take up this no risk trial I really only want to know the major benefit the major bullet points And a happy person I will click Trial right, but if I'm selling the software, maybe it's a different story I would hazard a guess that if I'm selling the software for 99 dollars without any trial People would respond better would respond better to the first version. Does it make sense? But there's no way of telling we always have to have to test Now the second purpose of analytics in my mind is to help you conduct user research Now I know a lot of you are UX designers and you are obsessed with UX documents like persona task analysis All these different user research are trying to help you achieve the task of finding more about the user while you can conduct primary research Survey interviews and all that but it takes time and also in focus group or in interviews People's response might be actually different from how they would behave in in real life unsupervised So the more reliable data is actually what they do on the site again. This is a heat map. What's the data here? What's the hottest part of the screen? Like these menu items so there are one two three four five six seven Or how do we usually as UX designers decide the information architecture or the manual structure? Sometimes we Conduct this exercise called card sorting. How many know have used card sorting? So basically you give a deck of cards to users with all the different Pieces and you ask them to reorganize right based on how they perceive your information architecture You would decide on the menu structure But this tells us what people pay attention to the third and fourth section of the website because They probably find it very important Right, so they should actually be the most prominent part of your navigation structure and Sometimes I also tend to see a lot of people pay attention to logo, which is not the case here but a typical failing design is that All the all the attention is given to a huge logo that occupies like two thirds of one third of the screen And people are not actually paying to attention to what you want them to pay attention to But let's say this part of the screen. It's a video Let's say you spend a lot of money producing a video and you want people to actually watch the video Then in this case, have you achieved that objective? Probably not because people are not playing the video So that's insight what action should you do probably move the video around So that's user research. Let me give you another example When you ask yourself, who are my users? What are you actually asking? What's their gender with their age? What are their interests? What kind of devices do they use like ISP some are really irrelevant some are really important What do you think are the most important traits important information when it comes to who are my users? What would you like to know if you can only know say five things? Oh tough crowd income Very good education. Okay Why income why education that actually depends on the on the website? Why would you say income is the most is pretty critical? Yes But it doesn't actually mean the richer a person is the more likely is going to buy that particular piece of software or Particular piece of particular service, but I agree. It is one of the factors that we would like to find out Which is kind of difficult to find out from analytics what we are usually what we can find out very easily is Demographic data age distribution Gender by the way, this is a screenshot taken from Google Analytics It is the most popular analyst software out there because it's free and I'm sure a lot of you have actually used it, right? so Google Analytics is divided into four sections the audience section and the content of the conversion section the behavior section and acquisition section they tell different stories one is Who are your users one is where are they coming from which is kind of a behavior? one very important is what they do on your site behavior and Technically speaking conversion is also part of the behavior, but because it's so critical They have separated out as a fourth section. All right, so these four sections all tell their own stories in the audience section We will get data on age Gender interest geo so geography. Why do we want to know these they serve many different many purposes one is Every design decision you make every content decision you make must align to their Demographic data and their interests right. Let's say most of you are in this case. We see that the The biggest group of visitors are 24 to 34 and male So probably developers Right, so if these are the demographic then maybe we can get away with making the website a little more technical Then it is but if we're looking at Pinterest Then we'll probably see a data like this and it's all women, right? So maybe you have to make The language limit a little more Emotional a flowery if you are selling a card you have to describe Ah the the seat is the the leather is taken from Italy 10 people are Supervising it if you if you touch it you will feel you will feel like in heaven or whatever If you're explaining to male audience you say Zero to 100 kilometers per hour 4.3 seconds, and it was like oh great car and Also it ties to advertising because I deal with advertising a lot right this section of the website interest Let's us target people by interest Right normally without analytics. You're just blind blindly guessing Here's another one which devices do my users use Why do I want to find out about that? Some people are still not convinced that we should always go mobile first the mobile first is the philosophy where you start The design with on a smaller screen and gradually expand on a larger screen There's still people aren't convinced because they think I'm not sure two years ago The number of mobile users had surpassed that of the desktop users, but just by how much I'm not sure But let's look at this the words are a little bit smaller, but Essentially the top 10 group almost all of them has a width of 360 370 400 320 what does that tell us it tell us the vast majority of users are browsing through a mobile device Therefore you have to rearrange the layout you have to make the words slightly bigger You have to make the call to action button slightly bigger so they can navigate and you have to really Make sure that the information texture architecture the manufacturer really elevates the most important part of the Section of the website which once again depends on the analytics right as I shown you in the previous screenshot Next question. How many people find my site useful in other words? How many people think? Yeah, we I have a good user experience What do you think is a typical trade or behavior of someone who? Enjoys my site What does he or she do spend more time so more time on site come back? I think that's very important now as a marketer. I deal with Marketing funnel every day so marketing funnel essentially says that people do not convert on the first go They go through a typical process of awareness Consideration conversion and loyalty or when they first come to your website very few people a small percentage would actually convert They will remember your site. They will probably compare you to your competitors That's a consideration stage and they then they will come back and convert So I want to know how many people are actually returning right? So this is returning users versus first time users So if I see this chunk to be really really small then I would be worried that means the vast majority of my users My visitors are not coming back Chances aren't they're not gonna convert now as the next step as inside I would actually add the so-called segments. So most analytics software let you Define your segment segment means a group of people with a much homogenous trade So I would add a segment that only takes data from first time users I want to find out exactly how many first time Users who do not return what actually convert and how many of those returning users would convert and After how many visits right Google analytics and all the other analytics software would tell you that Now the opposite side of that question is how many users do not like my site? How many users? I should think I have a terrible user experience What do they do? they never return and chances are They don't even proceed to the next page right, they come to your page and leave and we call that a bounce right as Opposed to a sticky site where they stay so if somebody comes to your web page and immediately leave It's considered a bounce so we want to find out what is the bounce rate if you have a 50% bounce rate that means right? That means you're wasting half of your traffic Because people do not proceed Then I usually get the question that Basically says how our website is the latest trend. We only have one page. It's a single page website Technically all your visits are bounces right they just land on the side and leave what should we do? Now there's something you can do instead of tracking page to page you can track different elements Which will I would talk about in the next few slides So if you really want to find out how many people are Not finding your site useful or you want to know that whether you have a terrible experience look at the bounce rate That's the first data. I would look at and as well as the time on page right one more question How do users behave on my site? That's again a broad question now as a UX designer especially for the UI designers in the room They would usually come up with all these mock-up and art board You will have screen flow right screen a flow so screen B B to C C You know diverge D1 D2 and so on. That's the ideal situation. You want people to go from say Using us an example. We want people to know who we are as a marketing agency So about us right the next step is probably services or maybe case studies because people want to know whether we're trustworthy All right, so social proof is big elements about us services Case studies and then eventually we want all of them to go to the contact page and submit a query form or you know Quote contact us. That's the idea flow, but are your visitors really following that flow? All right, so Google analytics again similar to other analytics software will show you that Where people are going from one place to another perhaps they land on the homepage first and Then most of them depending on the thickness of the bar They go to particular page and that particular page go to the next page Now if they are going all over the place, then you know that they are confused They're not following the user the screen flow that you had planned for them This will give you a rough overview of Users behavior, of course, there are other sections of the analytics that we can make use of But this is the most visual one Now that was inter-page behavior. How about on a particular page? So back to the previous question I asked if I have one page website Or if I have e-commerce website where I want to know what people are doing before they order my product I want to know how many people are adding it to the card how many people are banning in the card I can track different Elements on the screen I can track how many people have click on the quick view So that's e-commerce website where you examine the detail of the product How many people are adding it to the card how many people are removing it to the card that statistics tell me Whether my prices are setting too high whether my product description is detailed enough so on and so forth Now the third area so again the first area is to help you make design decisions second is to help you conduct user research The third one is to help you Define UX success now. We like to talk about good UX experience user experience and bad user experience all the time We say good UX bad UX, but you can't really quantify them if both sides have good user experience Which one is better? There's got to be a way that you can measure So all of us have KPI UX should also have KPI to us The ultimate goal of good UX is not to make users say wow, this is really creative. I like the color Right to us the ultimate UX goal is to generate as many money generating actions as possible So I repeat the number one goal for us is marketing agency Is to generate as many money generating actions as possible That money generating actions could be an add to cart could be Find out more Right, they all lead to the final conversion, which is a sale or trial or sign up So you got to quantify your user experience in that framework, right? You can't just say it's pretty or it's not pretty So the next question is how do we? Define our objectives. What are the possible goals that we can define as a UX designer now? if you're actually selling something if you are you e-commerce platform or if you are a Software seller then you would have sales or trials If you're an agency you would track people who are calling how many people are submitting queries Even how many people are signing up for the newsletter because newsletter's purpose is to lead nurture, right? Nurture the lead to the point that they are willing to convert So all these different goals can be tracked within Google Analytics. Now. This is the coding part How it works is that you place a piece of JavaScript code in the source code of your website And once you do that Google Analytics This will track all of your page every time somebody clicks something on your page That action will be sending a message to the Google Analytics server and Google can track all of that But then I another frequent question I get from Users is that Well, we don't sell anything. We don't even want people to call our website only serves as an information portal We have lots of pharmaceutical clients now in Singapore. They are forbidden to sell anything online So the only reason for their website is to provide information What kind of go can I set? What do you think? They don't have any sales. They don't have any phone calls. They don't have any inquiries What do you think? What can you track as the goal? If you have two designs of such website, how would you decide one version is better than the other? Engagement very good So that means the time spent on site the number pages browse right so you can set up that as well Google Analytics and other websites analytics platform allow you to do that you can define A goal to be greater than X pages or duration greater than X minutes Right. That is a simple way of quantify your UX What I'm trying to say is that every single page should have a goal. It's either a conversion or a Money-generating action that leads that contributes to another page's conversion What if you know that ten out of a hundred people who are arriving on the side are Filling out the form meaning you have a 10% conversion rate right that could be bad or that could be good depending on your historical data But that doesn't tell me how I can improve Because the form is a form right it's either removing that form or let it stay What can I do to improve the conversion rate of the form on that page? This can be achieved by Event tracking an event is Google's lingo. It basically means anything that people would interact with on the page So if I click a button if I download a PDF if I click play on a video, that's an event. That's an action Right. So if I am marketing through videos I want to show the features of my product and I want to know whether people are finishing watching the video I can track the Progress of the play right whether it's 20% 50% or 75% because that would tell me how we engage the user is or how I am doing explaining using video similarly, I can design I can assign an event to each single field of this form Why would I want to do that? because you will tell me The point where most people are dropping off now we had this client is a banking client and they had this Big-ass form right and we're getting a really low conversion rate So we what we did is to assign each form as an event then we that way We we want to know whether people are dropping off in a linear fashion if it is linear then there's not a lot you can do but if it's not linear if it's dropping like this and Reaching a certain field in the form all of them are stopped dead in their tracks And you know that particular field in that form is causing the most problems Which turned out to be the case. Can you guess which form field? Is the culprit of stopping people from converting? Any guesses age Date No, because I don't really care. Oh, you mean date of birth No, the the field in question is actually income because people are usually very private about their income even to banks I don't like to review their income So that's the data. That's the insight over the action. How would you counter this? People's natural reaction aversion to revealing their income Several ways one is not to ask them About income. The second is to explain In a way that your audience would understand what feel that it's beneficial you would explain that You are customizing the offer based on the income for example. So anyway, you have to Address that objection another example. It's a local supermarket And one of the field that stopping people dead in their tracks is actually date of birth Because people don't want to review their date of birth and it turns out that they are recording date of birth only because they Want to send out birthday gifts So if for that particular reason you don't need the year number, right? You just need the month and and the the month of the day of the month So removing year actually boost the conversion rate. So that's a micro conversion again, it's a It's that coding part of this Presentation where you can assign different fields. So some of you are should be quite technical like in JavaScript You would be able to know whether a field is blur is focused or not Whenever that field is focused you can send a message through JavaScript to Google Analyze server To tell Google that this particular field is clicked and entered information if That field is untouched then you know that people are not going beyond that form, right? so The final screenshot in this section. I want to show you is what I touched upon just now It's number of pages per session It goes back to the Statement I made earlier that data can be ambiguous So if I am telling you that I have two websites or before redesign after redesign after redesign We are getting a higher pages browse procession Your natural conclusion is that it's a good thing right people are seeking more information But that's not always the case in under which circumstances do you think that a higher Number of pages browse procession is actually a negative thing Searching for something. Yeah, very good So you could mean that people want more information It also could mean that people are not finding the information they're looking for this is especially true for support site So you can designate the support part of your website to be tracked by analytics only the support pages If you see an increase in number of pages browse, then you know that your support pages are not doing an as good job So that's the data insight. What's the action? How can you? Make sure that you're tracking this correctly Have you been to one of those software websites the documentation where at the end they say do you find this article helpful or Is your problem solved now? I can bet that on each of those buttons. They have set up event tracking every time somebody clicks I find this article helpful or My problem is solved the analytic is the back end the analytics would know that those questions are marked as solved Instead so instead of looking at the average number of pages browse, they will look at that as a conversion Okay, so the final reason for analytics and I think is the most important aspect analytics is to Constantly improve your response rate now We are a marketing agency where obsessed with performance and here's something that outsiders typically don't know Let me give you an analogy. I used to do a lot of copywriting And as copy copy writers would know this that they will try to write the most effective headline from the start But chances are they will fail because there's really no reason no way to know that To guarantee that you will have the best headline from the start or there are copywriting formulas For example a typical headline that will work. It's the how-to headline like how to Make ten thousand dollars in five weeks how to attract beautiful girls Quickly and easily even if you're ugly and short. I mean I would respond to that headline obviously, but what if If other people's response are different from mine, there's no way of telling I so the only way of constantly improving your response rate is to split test and Putting on different versions of the headline or different elements of the website and let the data speak Good marketing agencies or good copy writers good ux designers probably won't give you the best version from the start but what separates from the Average agency and the stellar agency is their ability to constantly improve right which you should always Use analytics to guide your split testing always have multiple versions. Let me go back to this this example So we now know that an image-based version outperforms the text-based version What's the natural next step? What else would you like to try? Sorry video very nice. So text Image video that's something we could try what else what else would you like to try? Everybody's drowsy because of the food so if I know that Putting on a face to the website to the software helps us convert Then maybe the next thing we could try is to try different faces right now as you can see that Michael has the biggest lift in conversion rate and John has There's really no way of telling now. I think it's a t-shirt But that might not be the actual reason the only way to tell is through analytics right let the data speak Now the next thing they could try is perhaps the headline or based on one piece that we talked about earlier is to have all of them Turning right or turning left to look at the headline. That's something we could try I we could also try the call to action now typically the elements that lead to the biggest lift in conversion are Headline image call to action beginning paragraph these four things so Let me tell you just very quickly what other tools that we use for UX improvement One is called Quantcast or similar web now Google Analytics is very good or all the analysts of are very good at Extracting insight from your own data, but what about your competitors? I saw similar web and Quantcast all these third-party website. They track almost all the major websites out there on the internet so you can almost Get insights the same way that you can get insight from your own website Google Analytics is obviously a good one, but the problem with Google Analytics is that It's anonymous all the data are anonymized because Google doesn't want people to know who the actual visitor are You can only get statistical data meaning 10 out of how many people are doing this But you don't know what a particular person is doing so kiss metrics is and several other solutions will tell you What exactly a single person is doing? John has come to the website ten times each time He's spent two minutes on the third time He converted to sign up for newsletter on the eighth time he submitted a contact form So kiss metrics and several other solutions will give you that level of detail Another one is crazy egg crazy egg gives you a heat map. It generates heat map Based on where people are clicking on the page, but that's the next best thing the real Good the really good heat map is generated by tracking the Eye movement the people movement. I think the cheapest device is like a hundred dollars Next one is called inspect let again This is their whole whole load of software that works like this how it works is it records? User screen many of you probably have used similar solutions. So basically let you observe The user's interaction with your website as if you are looking over his shoulder. So that's very good at deducing primary data and if you're selling an app or if your Business is app-based then you obviously want to use mix panel Which has a big focus on mobile analytics and app any as the name suggests it gives you the the data on the App landscape what other people are doing versus you are and if you want to improve your split testing Here are the two industry standards basically optimise Lee and visual website optimizer it lets you Concurrently test multiple version multiple elements of the page So you can conduct not just a B testing, but also a B C D E F G a testing. We call it multivariate testing All right, so these are the Types of software that we use on a daily basis So with that Let me do a quick recap. So what have I talked about today? First thing is that opinions can be risky opinions your opinions versus mine is not a very Productive discussion, so we don't know whose opinion we should rely on the only way we can Get actionable insight reliable inside is through data, but data themselves are ambiguous You have to think in terms of data inside an action always let action guide your action should be the design decisions and always let it guided by the insights user dry from numbers and Finally, we talked about the four areas where analytics and help you it will end all the debates It will help you conduct user research It helps you measure your UX success and it helps you improve your website your user experience Over time so with that I want to end my Presentation if you have any questions right now, I would like to take them You can clap questions Yes, please Hello. Yes recently I had a client who Actually had Potter in the travel advisor. This is he's the other travel agents basically they did a Social like SMO sort of like digital marketing so we help them and we targeted them and everything was you know on track what they KPI is and everything But they complain and they say that oh, there's no conversion Nobody's calling and everything like that. And so they blame us for basically that's happening Because you're not doing a good job and you're supposed to do that. So how how do we actually? Yeah, try to explain to them that is a great great question so the question is if you are doing social media marketing and Often that you observe there is no conversion from Facebook Let's say you post constantly on Facebook, but nobody is actually converting from the links there That's a typical problem. It goes back to the The product funnel the marketing funnel that I mentioned people usually go from awareness to consideration to Loyalty so we call it the top of the funnel and the bottom of the funnel Social media is typically a top of the funnel element meaning to say social media is very good at generating awareness But visitors has to go through multiple steps before they are converting now The analytics tracks in such a way that if I come from Facebook the first time and land on your site And I remember your brand ABC right the next time I search on Google ABC I either click on your ad or I click on your listing. That's no longer a social media visit That conversion will be counted towards either SEO search engine organic traffic or that will be counted towards ad Which is why across all businesses that Facebook or social media have very low conversion rate The only way to explain to them or a very effective way It's to use something called a multi-channel funnel which you can't really see on this page But some here down here some a part of Analytics down here. There's a section called conversion. You click on conversion. There's a section called MCF multi-channel funnel What multi-channel funnel is does is to analyze the entire user journey the entire path So because Google has access to everybody's cookie, right? So if the first time I came from Facebook the second time I came from ad the third time I came from organic search the fourth time is a direct Typing URL into search engine and then I convert Google will not only tell you that yes the last Direct visit resulted in a conversion, but it will also show you the entire path Facebook ad Organic direct then Google also does this statistical analysis on what the typical path is It will tell you out of all the sessions resulting in a conversion of all of these How many of them in how many of them does Facebook or social media play a role so you can actually tell your clients look Out of all these 100 conversions. You're right. Only 10 of them was the last step where Facebook played a role, but actually in 60 to 70 out of them There was this element of social media Facebook was present Throughout this entire journey throughout this entire conversion path. So that's what we usually use Any other questions? If not, you can always email me at There's another question over there like You can always email me by the way or add me on LinkedIn Let's say if a client and they are having a sale campaign and then They come to us and then they are they want to us to improve the Conversion rate right and then so how we do the testing So without without affecting the sale campaign without affecting the sales now that is So if I understand correctly, you're saying Let's test multiple versions of the website While sustaining the sales. Yes That is kind of tricky. It's more like a chicken and egg question Because if you do not test then you wouldn't know whether there's opportunity for you to improve But if you test then chances are it can go up or you can go down. So your sales would be affected That's how testing works. So you have to Let the client knows and the client can actually Decide together with you. How much of your traffic you want to test? So what typically happens to these? Multi-regional brand right MNC's they do not test with 100% of the traffic that would be very risky What they do is to test only 10% of the traffic or These big brands they will start with a particular country But let's say if yours is a smaller brand you can say let's not go crazy and test the 100% of the traffic Let's test with 20% of traffic 10 so 80 Sorry 90 and 10 Now actually you set aside 80% and you only test the remaining 20% is 10% versus 10% Then you can already see some some improvements But here's another thing I want to bring up The benefits the major benefits of working with an agency with analytics experience is not just the approach They adopt but also the insights they have gathered throughout the years Let's say that line of sight thing right most most people will not know it because they have never made such tests But if you are working with an agency with such experience then based on all the past tests that they had conducted in the past They would choose the version that is likely to result in an uplift instead of a down Downgrade in performance. So to summarize don't test 100% of the traffic go with an agency who is versed in testing Okay, no more questions. Like I said always you can always email me or at me on LinkedIn. Thank you very much