Is your design providing significant value?

Paul Boag

This week on the Boagworld Show we have talks on value-based design, chatbots and why software development is getting harder.

Skip to talk 1, talk 2 or talk 3.

This weeks show is sponsored by Fullstory and Teacup Analytics.

Paul: Hello and welcome to the Boagworld show the podcast about all aspects of digital design, development and strategy. My name is Paul Boag and amazingly Andy Clarke has actually turned up this week. Nice of you to bother!

Andy: Well, you know! I had better things to do last week, you know. I had to wash my hair…

Paul: At least… Oh, I see rub in the fact that you’ve still got hair. That’s nice, thanks!

Andy: Genuinely I buggered up the time difference. Because I forgot whether it was nine hours or ten hours or whatever. And yes, sadly, I missed it. You talked to yourself like normal! Just like being at home.

Paul: This is true. Yes.

Andy: Did you laugh at your own joke at the end?

Paul: Well of course. It was the only way of being sure of intelligent conversation to be honest. I purposely…

Andy: Did you do the fantasy joke?

Paul: The fantasy joke? Yes I did.

Andy: Yes.

Paul: What was wrong with that? It was a good joke.

Andy: Yes, it was a very, very good joke actually. It was one of the better ones.

Paul: Yeah.

Andy: It’s the way you tell them.

Paul: So, how long do I have to put up with this misery of time zones and bad connections and crap like that? When are you going to get your arse back to a decent country? I mean, the weather here is phenomenal at the moment.

Andy: I know, there’s been people… Do you know how Al Power? Lovely Al Power?

Paul: Yes, I know of him.

Andy: Yes, lovely fella. I follow him and his wife Zoe on Instagram and they’ve been posting some lovely, lovely pictures of the sun and the sky and all of that stuff. Mr Hicks had a good… It was Father’s Day this week apparently, in the UK. It is a different Father’s Day here, we have a different one. Perhaps, in case you have a different father, I don’t know. And…

Paul: Hang on a minute, I don’t like the way you started referring to we. Have you gone native already?

Andy: (Laughter) Well I suppose so, yes. So yes, I have been following… I’ve been vicariously living the UK summer through John Hicks and Al Power and you and various other people that I follow on Instagram.

Paul: So what’s the weather like with you at the moment temperature wise?

Andy: It’s winter so it’s been grey and cold and it’s been about 14° and it’s been raining. There’s been a bit of low pressure off the coast of the Tasman Sea and so yes, therefore it’s been wet and rainy. But it is winter here so you know, mustn’t grumble.

Paul: (Snigger) Serves you right.

Andy: But to answer your question about whether I’m coming back to Theresa May’s kingdom, not any time soon is the honest answer.

Paul: Oh, right! What happened to the six months thing then?

Andy: Well, the six months thing came and I realised that I was enjoying it way too much for it to just be a six-month thing. Apparently… This is the thing, this is the thing that you will not be able to understand. The listeners will have no comprehension about how this has possibly happened but they think I am actually doing quite a good job!

Paul: Yeah.

Andy: So they offered me a job and I decided that yes, I was going to take it. So I’m going to stay in Australia, in Sydney for the foreseeable future.

Paul: Well, I mean to be fair, you know, they are a very long way away from civilisation so I imagine their choice of, you know, designers, developers must be fairly limited so, you know, you end up scraping the barrel. I mean the people that are there are convicts and, you know, the rejects from society. So yeah, it makes sense really that you managed to get a job there! Well done.

Andy: I am the webs Laurence Llewelyn-Bowen.

Paul: (Laughter) What!? What does that even mean?

Andy: Well, when was the last time you saw Laurence Llewelyn-Bowen on TV in the UK? Apart from reruns of changing rooms from the 1980s. When was the last time? You know, you don’t see him. You don’t see him at all. But he is on telly here. He has rekindled his career in Australia. He’s got a new lease of life. Boy George the same, Boy George is on the voice here because he got kicked off the voice in the UK so basically old has-been’s come to Australia and here I am!.

Paul: So I’ve probably got another what, five years then…

Andy: Five years and you’ll be here, yeah.

Paul: Good to know.

Andy: You’ll be here. It’s like the kangaroo sanctuary for old British web designers.

Paul: (Laughter) We could start a seniors home. I’m sure we talked about that before once, a web designers retirement home.

Andy: Yes we did. And it would be somewhere nice and warm obviously. But yes, I am staying in Australia.

Paul: Oh, well, congratulations. I’m very pleased for you or something. Actually no, I am deeply envious if I’m honest.

Andy: I’m going to be chief, well in fact I am, chief colouring in officer.

Paul: Well done! Do you get like a gold plated box of crayons?

Andy: No, they haven’t given me that yet. Maybe I should. Maybe that’s what you get after 50 years in the business, like a gold box of crayons.

Paul: Yeah, I wondered if you got it when you retire instead of a gold watch?

Andy: No, no. So what do they call them. Golden hello? Which is not at all like a golden shower. Anyway, moving swiftly on.

Paul: I was going to say, it sounded like something a bit dodgy didn’t it. You are a bad man.

Andy: We ought to talk about value-based design is what we ought to talk about.

Paul: Yes, we ought to. From someone who actually knows what they are doing. Now, do you know this guy?

Andy: No I don’t know this guy.

Paul: Okay. I’m trying to work out how to say his surname. Disabato?

Andy: Disabato? I would say, yes.

Paul: Yeah, so this is Nick Disabato. We are going to go with that. I’m sorry Nick if we’ve just absolutely ruined your name. It’s part of the course if you’re going to be on this show. So he is a designer and writer from Chicago who does a lot of A/B testing on online businesses so he’s got a whole load of different people he works for. He’s written a small small book about interaction design and does loads of really cool stuff. You can find out more about him at Now I’m really interested to hear Andy’s comments on this one because Nick is talking about value-based design and how we need to measure the economic value of our design. Yes, I think it will be interesting to hear what Andy has to say on this subject because I’ve heard him say some different things on this. So let’s see what Nick has to say for himself.

Value Based Design

Play talk at: 07:23 – How do you measure the economic value of design – and use it to not only grow a business, but further your career? This talk will cover some high‑level methods to measure design’s impact – and the operational ramifications of pursuing a value‑based design strategy.

Nick Disabato: Value Based Design
Nick Disabato is a designer and writer fromChicago who uses research to drive A/B tests for all sorts of online businesses. He’s probably most known for writing Cadence & Slang, a small book about interaction design; and for Draft Revise, an ongoing A/B testing service for growing businesses. Visit Nick’s agency at

Hi there, I’m Nick Disabato I’m a designer and writer from Chicago and recently I saw two blog posts in the middle of April that came out from two different people that talked about kind of the same topic. They were pretty, you know, well-known people. One of them was Jared Spall who runs you i.e. and he is extremely famous speaker and author across the world and he wrote about how to justify the value of design in an organisation. Then four days later on April 17 Julie Zhuo who is the VP of product at Facebook she wrote a blog post on medium about how design often feels like an uphill battle in an organisation and that the fundamental issues that you have in improving the value. And I must admit that I am surprised that design hasn’t proven its value in 2017 because by this point you have dozens of organisations, I mean Apple is the biggest one but you also look at B&B, al their cofounders I believe are designers. You have dozens and dozens of organisations based themselves on having good quality design both graphic and UX, very good design research patterns. If you look at an organisation like Uber they have redesigned their application three, four or five times and have done wholesale refactors of them each time and yet organisations are still suspicious of what designers can actually bring to the table. I think the onus on this is mostly on designers. We don’t do a very good job of proving our value, we either came from art school or we fell into the field from some sort of other profession and in neither of those circumstances are you actually taught about the business ramifications of design decisions. So this is going to be a very brief summary of how I do that in my own design practice and how you can do that in organisation of any size.

So the number one thing that you need to be thinking about when you are proposing a way to calculate the value of design is by measuring it. The best way to measure it, in my experience, is by taking a little look at analytics. Now, Google analytics is one of the worst applications of all time. It is a besantine horror, I get it, I know. But there are a couple of things you can do if you kind of go step-by-step through it that will help you prepare analytics and install effectively and focus on the right things. Why would you want to do this? The goal is to connect a design decision to a corresponding economic impact within the organisation. You are there to bring more value than, frankly, you cost. So you have a salary, you have benefits, you maybe have square footage in the office, maybe have a laptop that they have given you or something like that and you need a way to actually prove that you are belonging their. Otherwise people aren’t going to believe that design has any value and you are going to be struggling to have these sorts of conversations. So with that in mind the two things that you should be calculating in Google analytics are revenue and conversion rate. Depending on what business you have, obviously I’m going very, very broad with this, you should probably have a revenue and a conversion rate. So on the thank you page you add basically a couple lines of JavaScript that tell Google analytics how much revenue this person made, if it’s an online store or something like that is usually called e-commerce analytics and the goal there is to calculate the average revenue per user. So if you have, you know, 1 million people coming in and a hundred thousand of them are buying and they are all buying the same $20 product then it is $2 million per million users so you have a two dollar average revenue per user. That is a very simplistic calculation. Correspondingly if it’s a hundred thousand people it’s a 10% conversion rate. A pretty high conversion rate, is not bad. But you want both of those numbers to go up so you want the average order value to go up. People should be buying more stuff. You want the conversion rate overall to go up and you want the overall revenue for the business to go up because if you have a higher conversion rate but you have lower revenue then that’s not so good either. There are a lot of ways that you can do this you end up, you know, creating all of these goals and then you are doing design, you are listening to your customers seeing what it is that she motivates them to buy and in a non-skeezy, non-manipulating way make it easier for them to buy the things that you are selling. So that partly involves promoting the value of your product, partly involves making it easier for the usability standpoint. So a lot of my work usually involves cutting out unnecessary elements, simplifying unusable forms, removing extraneous form fields, improving page load time, improving mobile which is always a perennial issue. In doing all of those things usually the conversion rate goes up because it is easier for people to load pages if they are on the subway or something like that and it’s very hard for them to get on a connection then they can actually download a faster page, it makes more sense. So doing all of those things you have configured conversion rate as a goal, you configure revenue as a goal, you create, maybe, a funnel from one page to the next if you are running a software business or something it could go home, pricing, sign up then the beginning of an on-boarding process and once you have all those things prepared then the next thing that you do is research. So you are continuing to pay attention to analytics, maybe there are geographic or demographic things that are interesting. Maybe traffic sources are being a huge influence on your funnel. You also run heat and scroll maps to see where people are clicking, where people are scrolling, where they are not clicking, how far they are not scrolling. You can run surveys on people both post purchase. So on the thank you page write, you know, just a single fields that was like “Anything that held you back from purchasing today?” or “How did you feel about our checkout process today?” And just one form blank and you will get a lot of interesting things to be mining. These are all important because research is obviously a fundamental component of any design process and so you want to be continually listening to your customers and building it in as a holistic capacity with the organisation. So you are saying “Okay, it’s not just me sitting in photoshop doing design every day, it’s me listening to these customers and figuring out how to synthesise that into design.” This sounds more like a data driven version of what an organisation like IDEO or Cooper or adaptive Path would do. Which are all pretty well know in the design field. They spend most of their time doing fieldwork with existing participants or customers and spend a lot of time just sitting and listening and researching. And then they go back and try and translate that into something that has a positive upside for the business. In this case you are sitting there and looking at what people are doing in analytics or in surveys. You might even want to recruit people for interviews on the phone or go to a website like and do usability testing or whatever have you. You are trying to do that in a way that has a palpable economic upside for the business. So that results in “Okay, where are we leaking revenue here?” “Well lots of people are dropping off the checkout form right here, that’s bad let’s try figure out a way to improve that.” I can’t tell you how many times I have encountered forms that there was like one busted field in it, it didn’t have the right mobile keyboard or it wasn’t doing validation correctly and it was costing the business like a 5% bump on their conversion rate. It is worth getting a developer on that to fix it, right?! So, in this way you can justify design decisions and focus critique around what value it might have for the business. So if it’s something that happens to be high risk like you are cutting out billing address from a form which might be a lot of development effort, it might be very hard to justify from a risk standpoint or a fraud standpoint or whatever have you, you can maybe cite other studies, you can maybe run an A/B test or something like that that determines “Okay well, we are going to make this much money by actually cutting this out, so therefore it’s worth it to us.” Having that sort of database justification is very, very good for focusing critique. It is great for helping people understand why it is we are actually trying to make these sorts of design decisions. It also kind of de-fangs the what is called the Hippo or the highest paid persons opinion. So if it is the CEO that recommend something that is just gobsmackingly ill-advised you can go back and is say “Well, the research doesn’t support it.” That way it makes the suggestions and critique a little bit more egalitarian and focused and rational. With that in mind if you don’t have the traffic necessary for A/B testing can still measure step function changes in Google analytics. So if you have analytics, as I mentioned before, when you hit a redesign you can add what is called an annotation on Google analytics which says “Okay, on this date we will launch the redesign.” and then you can say “Okay, we looked back and here is the history of all design decisions that we launched for the site.” If you can integrate Google Analytics with any sort of git pushes to actual marketing pages of your site or the front end of your site or whatever have you, you can say “Okay, we did this, this and this on these dates and the conversion rate has been going steadily up ever since.” This allows you to go back and look at the impact of your work and it also keeps you focused and accountable to that work, right? Because if you are focused on the conversion rate and your presence there and people taking you seriously is based on the conversion rate it’s like operating on a deadline, right? If you program just for fun it’s great but you become a great programmer by programming in a more stressful environment on deadline! Same as design, right? If you are designing so that you can improve the conversion rate then suddenly very few other things matter. The goal is to do this in a way that is ethical and still upholding the values of design that you have come to hold so dearly to yourself. I don’t advocate any sort of dark patterns or any sort of negative, you know, manipulation of the customer, I don’t think that makes any sense and it’s not the kind of line of work that I ever wanted to get into. But the goal is to listen to your customers and trying to create something that meets their needs. It also allows you to do this in kind of a more continuous way so that you can continue researching. You know, you’re never done running heat maps on things, you are never done surveying people postpurchase or maybe every quarter you survey all previous customers or you ask them for reviews after they have received the product, a few days later. You are creating a weekly research cadence where you’re gaining more data, more insight and trying to figure out a way to act on it. This works well for anybody as a full-time employee and also works really well if you are a consultant or an independent worker and you want to go on an ongoing retainer. So you might be hired to redesign the site as an agency but then a very good way to continue working with a client is not to find other places where you can redesign but find ways to optimise the design that you have already created. And recognise that you might not have got it fully right the first time but you are already fixing a lot of major problems ahead of time. So that is it on the economic impact of design I hope that this was useful for you.

Paul: So, that was Nick. Andy, what did you think?

Andy: (Sigh) Yes, I’ve got to be honest haven’t I. This is absolutely not the approach to design or the type of design that fascinates me, at all. I mean, that’s not to say that it isn’t valuable because design is there to solve a purpose in general. And I’m not suggesting that this isn’t valuable or that it’s not… You know, every design can benefit from various different types of input. You know, you can’t design something in isolation, it isn’t about one person’s vision of a thing. You know, the stuff I am working on right now, you know, I have a very strong idea about how we should accomplish something but that something has to be aligned to the strategic goals of the company. But, but, but, but I get incredibly frustrated when, you know, some people (cough, cough) Kenneth Bowles, think that this aspect of design is all that matters. The idea that, you know, you should base your design decisions on what people do rather than what you necessarily are trying to get from them, I find difficult. You know, I’m in a business where we design products and the idea that somebody needs to use these products is obviously incredibly important. If they don’t use the product and they can’t use the product then we don’t make any money, you know, they’re not going to stay as customers. But the idea that that kind of metric is the only aspect of design that matters to people is incredibly frustrating and, you know, I’m not saying this is bad, obviously, but we need to balance it with, you know, he mentions graphic design at the beginning of the talk, you know, good graphic design, design with personality as we are going to hear in a talk in a minute, is as important as validating design decisions based on data.

Paul: I mean… hmmm. This is a tricky one for me, okay, because I probably fall somewhere between the two of you, okay? I totally agree with Nick and I think there was a lot that he said that was very valuable. So for a start he talked about design needs to bring more economic value than it costs, right? And that to me is the difference between art and design. So it does need to provide a return on investment to it but I think the line there is that there are different ways of doing that, okay? So I think someone like yourself, Andy, you do a lot of things intuitively based on years of experience that provides that economic value, okay? There isn’t necessarily a need for you to measure every single element and every single thing that is going on but I think there are, there is real value in doing that measurements and the reason… One of the things that I thought provided real value which is something that you picked up on as well, I’m going to steal one of your points in your notes, is that it’s a great way of de-fanging the highest paid person in the room, it’s a great way of making the business case for what you have done. So it is not saying that all of your designs should be led and dictated by the numbers but it is a good way to either validate your design or alternatively to boost the case for your design, which I guess is the same thing really. But I think there is another really important aspect in it and it is that Nick was focusing very much on e-commerce sites and that kind of thing and many of the sites that I work on are not e-commerce sites, they have got other calls to action. So sometimes you need to be a little bit more broad in what metrics you are measuring and even in his situation, he talked about using non-sleazy manipulative methods and that it was really important not to do that but if you all you are measuring is the economic return on investment then you are inevitably going to fall into the worlds of sleazy, manipulative methodologies to get there, dark patterns and all that kind of stuff. So I think it is about measuring the right things. You know, you do need to measure things like delight, customer satisfaction, those kinds of metrics as well as the purely economic one. If you’re trying to build a long-term business, you know, if you’re not, if you’re just trying to win a quick buck and screw a load of people over in the short term then by all means just focus on getting that conversion as quickly as possible. But if you are building a long-term business that wants repeat customers and that kind of thing you’ve got to think more broadly than just conversion rates and those kinds of metrics. And to be honest, I think Nick was saying that he was just trying to introduce the subject.

Andy: Hmmm, yes.

Paul: So actually I think it’s a balance isn’t it really it’s like anything.

Andy: Yeah, it’s absolutely a balance and I think that the pendulum has swung a long way in the kind of metrics and UX area of design over the last couple of years and the example that I like to use is toothpaste. You know, people keep banging on about digital products as if they are like real products. As if it’s a tube of toothpaste! How efficient is the tube of toothpaste. Do you remember when the tube of toothpaste used to be made of metal?

Paul: Yeah.

Andy: Do you remember that? And you had to roll up the end to get the last bit out. Not an efficient tube of toothpaste. The bloody thing used to split down the seam and you would get toothpaste all over your bag or your toilet bag or something. Not a great tube of toothpaste. And you would lose the because the cap wasn’t one of those kind of fixed in plastic caps it was like a screw-on cap. Not a great thing. So we have obviously improved on the tube of toothpaste and that product design.

Paul: Mmm Hmm.

Andy: But that’s not what I do! I couldn’t give a monkeys about the tube of toothpaste and about designing it because that’s what somebody else does. If people are fascinated about making a tube of toothpaste excellent, but what I’m interested is in is why you buy Colgate rather than Maclean’s. What makes you pick up Aquafresh on the chemist or supermarket shelf. And that’s the aspect of design that I deal with so, you know, there needs to be room for everything.

Paul: But there is overlap in those areas because part of the reason why someone would have chosen Colgate over Maclean’s at one point was because “Oh, Colgate have come up with these great new plastic tubes which don’t split.” So for a while that gave that… that product design gave them a competitive advantage that did drive sales. So it does kind of overlap with the area that you are interested in. So, it’s not a black-and-white thing and I think particularly at the moment with user experience, you know, user experience is becoming a primary driver for many sales. People want the easiest to use experience. So, think of it like Maslow’s hierarchy of needs, right? So the kind of stuff that I maybe spend a lot of time on, or Nick spend a lot of time on, even Kenneth Bowles spends a lot of time on is the more base of that pyramid. It is the kind of fundamentals, okay? Of “Is it usable? Is it accessible? Does it meet my base needs?” You are looking at the top end of the spectrum at the emotional elements of it and the self-actualisation and those kinds of parts of it. So it’s all part of the same kind of pyramid, we are just looking at different parts. That’s how I kind of view it.

Andy: Yeah, I completely agree. You know, I am working with product designers every day that want to know about interactions and they want to know about what the colours that we are using means for discoverability and all that kind of stuff. Brilliant, love those guys, not what I do for a living.

Paul: Yeah, yeah. And that’s what it comes down to and I think it’s good… I think it’s nice that there are people focusing on different parts of the experience. By the way, just before we move on to the sponsor because we really should, we spent far too long on this. I do just want to say that you used toothpaste as an analogy I use Maslow’s hierarchy of needs. I think that makes me cleverer! Right, let’s talk about our first sponsor. All the more pretentious, probably is the truth.

So, we’ve just heard about value-based design so who better to talk about at this point than Fullstory.

Andy: Ooo, tell us about Fullstory Paul.

Paul: I really like Fullstory. So Fullstory is a tool I use all the time and I think it is… My personal opinion is that it is the best analytics tool out there for measuring what is going on in your site and knowing how users are responding to it. So, they asked me this week to share my own thoughts about the platform. So they gave me no talking points at all, they just said “You say whatever the you want to about it.” So instead of doing that, mind, what I want to do is just read you something from our slack channel this week. Because somebody happened to give this product a go and they were giving a bit of feedback on it. So this is straight from the slack channel all right? “Fullstory actually needs very little configuration to test anything. Pasting in some JS into the head is basically it. It is a broad-spectrum tool that monitors everything that you don’t exclude using CSS selectors.” Now this is the really key bit, this is the bit that I think says it all “From a two hour playing around that I wasn’t asked to do or authorised for I got evidence for 12 problems I’ve mentioned to others that I can now replay to them and say it’s not just in my head. For example even repeat users are leaving the site multiple times or unsure how to reach the cart despite the notifications.” All right? So for me that says it all. It almost goes back to what we were talking about previously. The area I really like analytics in is that ability to backup what you know is true. The evidence, you know, it is evidence for what you know you have already been banging on about and haven’t been able to convince people of. So definitely check that out. You can sign up today and see if it provides you with the same kind of value as it did the guy in our slack channel. You will get a month for free on their pro account, don’t worry, you don’t need to give your credit cards over so, you know, they’re not going to start charging you at the end of the month. If you get to the end of the month and you want to pay then that is great and you will get the full suite of stuff. If you don’t want to pay then you still get the full suite of stuff but they only record a thousand sessions per month. But, to be honest thousand sessions per month is often enough to get an idea of the kind of problems that are on your site. So there’s no reason not to. You can find out more about them, sign up and all the rest of it by going to, Boag.

Okay, next up we come to “Designing personality in chat bots” which is a really interesting subject. This is from Mike… I’ve no idea. I’m not even gonna try.

Andy: You are rubbish with names aren’t you! Is it Jongbloet?

Paul: Jongbloet, may be. I don’t know. You see I am, I’m terrible. It strikes fear into me. It’s Mike, that’s who it is! It’s Mike from Deeson. So, he’s the head of user experience at a digital agency called Deeson where he uses open source technologies with in-house UX design and development teams and he has built solutions for clients such as Robbie Williams, ooo, National Army Museum and the brewery Shepherds Neame whatever that is, neem? Is that a thing? Should I recognise that name? Anyway, I recognise Robbie Williams!

Andy: I’d like to hear you commentate on a World Cup football game at some point! It would be hilarious!

Paul: No, this is terrible. I don’t know what it is, it just creates fear in me.

Andy: Serbia versus…… Nigeria. There you go, you’d be great that!

Paul: Okay, thanks. Anyway, Mike wants you to know that they are looking to hire people which is good so you can go to their website which is D E E S O if you want to find out more about their job opportunities. But right, actually none of that matters, what we are really interested in is what Mike has got to say about chat bots so let’s listen to his talk.

Designing Personality in Chatbots

Play talk at: 32:43 – I was a chatbot cynic until I tested the bot Channel 4 developed to support their marketing campaign for TV show “Humans”. I’d like to share why it changed my mind and how we can design engaging chatbot experiences.

Mike Jongbloet: Designing Personality in Chatbots
I’m Head of User Experience for Digital Agency, Deeson. We’re open source specialists with in‑house UX, design and development teams and we’ve built solutions for clients such as Robbie Williams, National Army Museum and the brewery, Shepherd Neame. We’re always on the lookout for great people to join us! Visit Mike’s agency at

What I wanted to talk about today was chatbots. The reason I wanted to talk about them was because as a lot of people are I was quite a sceptic for quite a while. We’ve been doing quite a lot of work with them but I kind of saw them as a bit of a gimmick and I couldn’t really see the purpose for them. Then my creative director sent me one to try out. It was the “humans” bot that Channel 4 created to support the marketing of their TV drama “humans”. Humans is a show about synthetic humans and the premise of this bot is that you are contacting their support staff because your synthetic human has gone wrong. The reason this bot changed my mind about bots is because it was really engaging. Previously I have spoken with many chatbots when testing them and got bored very quickly but with this bot I found myself drawn into the conversation. I found I was almost building a bit of a connection with it. The narrative is designed to try and build a bit of an emotional connection with you. It uses imagery and names of the support staff to give them a bit more human quality. It tells jokes, it shows emotion, it shows that it is scared at times, it shows that it is funny at times and it shows a lot of human qualities. The chat makes you build a connection with this bot just as when you are watching a TV series you build a bit of a connection with the lead character or another character. It is the same premise and that’s what I found really interesting about talking to this particular bot and the way it was used as a marketing campaign to engage someone worked really well. I think chat bots have some purpose and one purpose they can really excel at is to support marketing campaigns. This was only a 15 to 30 minute chat overall but I was really engaged. Another example of using a chat bot as a… To support a campaign is one that charitywater created in partnership with a company called Lokai. So charity water are a water charity so they map the journey that a lot of young African children take every day to collect water for their families. It turned out to be roughly a 2 1/2 hour walk to and from the water source back to their village. What the bot does is it encourages you to go for a 2 1/2 hour walk but engage with the bot as you do. As you go for your walk, Yeshi, which is the girl’s name, tells you about her walk and you can kind of get an idea of what it’s really like. It builds that emotional engagement with the charity much better than a donate button or an article does because you are actually living what is going on. During that walk she asks you “Where are you from?” And you say “I’m from London” and she says “Oh, send me a photo.” And you take a photo and it sends it over and she says “Wow, our worlds are so different.” It’s these kinds of things that make these bots more human and more real that will engage people. So I was trying to think afterwards about what it was, how could I summarise what it was that made the difference with those bots. I think it comes down to one thing which is personality. So I started with the thought “What is personality?” And I asked a few people and I got words like memorable, impactful, it’s what makes us unique and different, it’s our character, our thoughts and our feelings, and it comes out in our behaviour and our actions and the interactions that we have with each other. I also looked up some Internet definitions and the two that I liked best were “Its the combination of characteristics or qualities that form an individual’s distinctive character” and “The quality or fact of a person as distinct from a thing.” And when we think about chat bots, if we want to design a chat bot, we want it to be engaging with more than a thing. So that second one “The quality or fact of being a person as distinct from a thing” really helps. And that kind of confirmed that personality is the thing that will make our chatbot stand. With hundreds of them out there at the moment if you don’t have personality in your chatbot it is not going to stand out from the crowd, you are not the 1st to market any more. A guy called Aaron Waters wrote a book designing for emotion and in a presentation he gave he looked at Maslow’s hierarchy of needs and drew up his own based on designing for personality and emotion. The bottom three layers, functional, reliable, usable are things that if we are designing chatbots they have to have those otherwise they are no use to anyone. But the top layer and the one that is really going to differentiate us from the competition is the personality layer, the thing that makes it pleasurable to use. If we think about more traditional apps and websites and we think about where they have personality and where they trigger emotion it is through fun, pleasurable things so the MailChimp monkey sweating when you’re about to send the email, charity websites use imagery to trigger emotion, it’s this personality that helps differentiate and helps us connect on an emotional level with our users. So personality speaks to people and emotionally engages your audience. Traditionally in web design personality has come out through user interface design and the functionality. When we think about chatbots it’s actually what the box says thinks and what the bot does which can be likened to how you might describe it in a human. What you say or think or what you do. So I think in conversation with UI personality is the new UX, so how do we design for personality? The good news is that we already do design for personality so in web design and app design, as I said earlier, we are thinking about what neat little features, what quirks can we put into our designs that will trigger emotional reactions in our users. I think that a lot of the tools we use in user experience and web design are just as applicable for chatbots, we just use them in a slightly different way. Take personas for example. A persona currently, or the way we use them in web design, represents a user of our audience so that we can design for them. But we can have a bit more fun with chatbots, we get to design the persona from scratch, who do we want our chatbot to be? What mannerisms to be want it to have, what quirks does it have? What positive and negative personality traits does it have?We can think about questions like “If my bot was a celebrity who would it be?” “If it was a TV personality who would it be?” We can draw all of this together into a persona which describes our chatbots personality. Following on from that once we have decided what type of person it is we need to think about how it talks. If I was to write down three things said by some of your close friends on paper you would probably be able to tell which one was said by which of your friends. The reason is because we all talk slightly differently, we all have slightly different mannerisms, slightly different ways of saying things. To make it engaging we need our chatbot to have its own voice as well. So once we have a persona for it we can start thinking about the tone of voice, is it jovial, is it very serious, is it aggressive? We can start to draw up keywords of how we want our bot to behave. A challenge with creating personality for chatbots is the visual aspect. It is a different type of design because we no longer have control over the visual design. People are using these things in Facebook messenger or WhatsApp. But the “humans” chatbot that I used subtly used imagery and emoji’s and this gave it personality. As I kind of mentioned earlier when you first start the chat you get to choose who you want to talk to and each support staff has an image and a name which gives them a bit more personality and makes you feel like you are talking to someone not a robot. It is a really interesting challenge we have as designers for how we design chatbots and I have only really touched the surface because beyond this you then have to design the narrative itself which is a huge challenge. But before jumping into the narrative it is really important to think about who your chatbot is going to be, what personality it is going to have and how that is going to engage your audiences. Because without a personality your chatbot will be just a robot with a bland voice and people will not engage with it and will not connect with it. Thanks for listening.

Paul: Okay, so that was Mike. I really enjoyed his. Chatbots have become a bit of a thing for me, if I’m honest, recently. That I’m not… That I was just as cynical as him but I’ve come at it from a different angle that for me when I was initially cynical about chatbots I was thinking about Siri and things like that. Those open-ended chatbots where you can ask them anything. I always used to compare it, because old like you, did you ever have a ZX spectrum or a commodore 64?

Andy: No. I was already working down the pit when those things came out!

Paul: Ah! So with those there were games… There was one game I used to play a lot called the Hobbit, right? It was a text based adventure game and in the hobbit you used to have to type in commands. You know, if you wanted to go east or pick up a hammer or something. And you’d spend half your time trying to guess the terminology that the programmer had entered into the text based adventure game. So is it “go west” or is it “travel west” or is it “travel to the left” or, you know, what is the combination that they want. And that to me felt like what chatbots were. You know? You had to guess what terminology they use and that just didn’t seem like a good user experience. But then I started to see a different type of chatbots and these were more like the choose your own adventure books. You must remember those!?

Andy: No, I don’t remember those. We didn’t have those in the North!

Paul: Right, well choose your own adventure books were basically you had a book and you got to the end of a page and you were given a choice. You could do this, or you could do this. If you do the thing a) you turn to page 18, thing b) turn to page 28. And you kind of travel through the story like that. So I started to see a lot of chatbots like that and they worked a lot better. But anyway that’s what got me into chatbots and that is not what Mike is talking about! He is talking about the emotional connection. I think that the advice that he was giving and the stuff that he was sharing really applies to any writing for the web because I think, personally, most writing for the web really lacks personality and sucks at it.

Andy: No, absolutely.

Paul: So I think all that chatbots are really doing are bringing to the fore a lack of investment organisations are making in their writen content and it still blows my mind to this day that people will pay a fortune for a content management system but won’t pay a penny for what goes in it because people think they can write. Which is bollocks.

Andy: No, I know.

Paul: Anyway, what did you think of it Andy?

Andy: You know, I would absolutely… If you are listening Mike I would really like to hear or see an expanded version of this talk because it’s actually one of the favourite ones that I’ve heard recently. Because he is talking about something which is very much up my alley which is thinking about giving designs personality. The idea… People would say “Yes but UX matters.” But do you know what? It’s about communicating information effectively. Getting people to understand a message and, you know, I loved the idea of somebody taking on a walk for what was it, 2 1/2 km or something. Interacting with this bot and comparing your experience doing a walk with the girl from Africa who is doing a walk to collect water and that kind of thing. It’s just, you know, there’s so much personality and opportunity for being creative there but that’s not just for creativity’s sake it’s because it’s communicating the messages. It’s getting you a deeper understanding of what that person is thinking and experiencing and everything else. No, that’s brilliant. I haven’t really had any experience with chatbots at all, I think, ever. I think it’s not anything ever crossed my radar but I just really love the idea that people are thinking about this as kind of another medium for creativity. You know, the other one that he mentioned from Channel 4, you know, I can imagine… One of my favourite ad campaigns is the meerkat, you know, compare the meerkat thing? It’s kind of gone off a little bit recently but I absolutely loved that whole kind of Alexander Orlof the meerkat in compare the I remember when that campaign started you could actually talk to Alexander on Twitter. And, you know, he would respond in the character and it was like a brilliant, absolutely brilliant kind of bit of advertising. I can imagine chatbots being used for that kind of stuff and again it’s not just there for designer’s sake it’s there because it’s helping people to communicate a brand message. I love it! I’ve got to disagree with Mike a little bit though when he’s talking about the fact that we already design personality into web design. You know, in my experience it’s quite the bloody opposite to be honest. You know, I think that most clients and a hell of a lot of UX designers, they are doing everything they possibly can to kind of eradicate any personality in a design because they are just so afraid of taking any risks. So it’s one of the reasons why so many websites look the bloody same. So I don’t think people have got this right at all in terms of web design. You made the point earlier on about copy. That’s another aspect where we should be really thinking about the message. You know, what are we trying to communicate, what is the best words to use, what’s the best medium to use, how can we art direct all this kind of stuff and bring people into the story that we are telling. Yeah, this kind of opened my eyes to some of the stuff we could be doing with things like chatbots and I’m definitely going to be looking at how we can use something similar at work.

Paul: Yeah, yeah. Absolutely, I mean I’m a great fan of chatbots I think there’s so much could be done. Okay, let’s move on to our second sponsor and then we will get into our final talk.

The second sponsor is Teacup Analytics. So Teacup Analytics is a great suite of all kinds of really wonderful tools but the one tool that I want to focus on this week is their AdWords tool which makes dealing with AdWords effortless. You can run ad campaigns in Google AdWords in a much simpler way because let’s face it running AdWords campaigns can be a real head headache can’t they. You’ve got to do keyword research, you’ve got to set up your campaign structure, you need to create your ad copy then your landing page has to be made and it has to be relevant and then you’ve got to do bid management and optimisation and it takes hours of your time each week. So Teacups new tool makes AdWords effortless. Basically it’s really, really simple. So it starts by helping you create the right keywords. So all you have to do is answer a few short simple questions, right? Then Teacup creates both the ads and the relevant unique landing pages and the keywords all based on those few simple questions. Once the campaign is up and running Teacup then handles your bid optimisation, adding negative keywords, testing ad copy and a lot more. It is absolute witchcraft, I tell you, it’s a really weird thing to watch. So if you want to try out the beta, because they are currently looking for beta testers then you can sign up. And when they say beta testers they really mean it. They want your ideas, they want your feedback. That was the one thing… last time they advertised on the show they absolutely loved was the fact that you actively engaged with them as a community and you gave them loads of great suggestions and that’s what they are after again. So, they are offering a limited invitation access to the tool. If you want that and you want to provide some feedback then you can sign up to get early access by going to

Okay, we come onto our last talk which is Peter Pilgrim which I think is just an awesome name. He is talking about software development and how hard it can be these days. Especially he is really getting into legacy a lot and dealing with legacy systems. So here is Peter’s talk.

Why is software development hard and why is it getting harder?

Play talk at: 52:00 – Clean architecture for digital software development.

Peter Pilgrim: Why is software development hard and why is it getting harder?
Peter is an independent software developer, contractor and architect who works with clients to improve their enterprise level ecommerce systems. Visit Peter’s agency at

Hi I’m Peter Pilgrim I’m going to talk to you about why our software is so hard or why it’s going to be so hard in the future. Why is software getting ever more so harder to write and build and why are we getting frustrated with it. I guess I am frustrated with the software that we are constantly building and writing. The problem is is that no matter what innovations we build around the process of constructing software it never seems satisfying to both end-user, development, users, testers. There’s always something that we have to do to get even further into the software process or to get a reasonable result. We have tried agile and in the way that the minimal value release, the MVR, for the Minimal value product, we tried to build these kind of fine tune our process of building applications with test driven development and the peer programming code reviews and we still don’t get the answer of efficient, consistent software. It always takes longer and longer to actually build the software in the first place. I think the problem is that the majority of us, unless you work in a start-up, as developers we are building against legacy software libraries or packages. I would like to call these giant balls of wax. As you can imagine a giant ball of wax is actually disgusting and messy. Some of these applications, and I’ve witnessed a fair degree, have say half a million lines of coding. I seen a lot, I have seen Fortran77 20 odd years ago which has nothing to do with object orientated programming and even proceed structured function, structured programme with records and procedures. I’ve seen Fortran with GOTO’s and people have been stuck with it. So that’s the general worry is that… Okay, I haven’t explained that correctly. The general worry is that it is getting harder and harder to chip away at this mountainous ball of wax so much so that I think companies are going to find themselves in grave danger. I suspect this is even hitting the news. While I cannot prove it the recent failure to update software in the NHS, vis a vis the operating system, Windows 95 and then the actual fallout of British Airways. A data search destroyed their data centre. Okay, that is speculative so I wouldn’t necessarily add that but the point is is that our software is failing because of the technology and also the people that we want to maintain it. If you invest in a software platform or some kind of architecture I can pretty much say there’s no such thing as a silver bullet, hundred percent architecture that means that your applications are going to be sustainable for the next 20 odd years. Certainly programming languages can last that long but because we are buffeted by trends such as technology, JavaScript or Java or C sharp it’s actually the libraries around these programming languages and actually the concept of programming and developing software that is being challenged. So once upon a time there were such things as literate and structured software development programming which then I suppose evolved into object orientated programming and now we are looking at something called functional programming. Don’t worry if this is all over your head. The point is this, developers always like the nice shiny brilliant diamonds, always looking for the next best thing and that maybe something like Scala which is an object functional programming language, it might be closure, it might be something like F sharp or it might be something as funky as Russ. All these are far removed from the generous HTML, CSS and JavaScript and PHP. Even in the PHP world I suppose things are moving beyond the symphony frameworks. Not that I’m a PHP expert, even new ways of thinking of developing software are changing so I guess 10 years ago people wouldn’t have used have heard of user experience at all but it’s actually been around since I guess the 1980s with Steve Jobs and the Macintosh interface. It was human/machine interaction and now we are looking at the voice. The point is that we can’t keep adapting and adding plug-ins and bolting on and boiler plating legacy applications. Eventually these applications that we write or have written 10 years ago suddenly become susceptible to change and it might mean the original purveyor of the language or the framework has gone out of business, it might mean that the original chips that the languages or even the byte code that is running these languages is not available. And therefore I guess even with economic factors if IT directors and people who are the decision managers may be tempted to sit on their, I suppose, benches and not do a lot of change. But as they say change is coming. And I said software is becoming increasingly complicated. Now as a contractor I have really seen the green field. You are bought in as a contractor with expertise to really hit the ground running and also fix problems that somebody wrote in that code 6 months ago or to add new features or to remove new features. It is very rare that people build new architectures because people are just… businesses are hardly finding the money in this economic climate to invest in great new software that is guaranteed to be delivered and earn them a profit. In other words take up an acceptable risk. Consequently because of the time pressures, there’s that word again, with code with used pair programming with all of that good stuff eventually we are all humans, we don’t design or build software a hundred percent sufficiently. Obviously you can enforce things with code style, with lots of tabs or margins or whatever you write, I pretty much guarantee there will be code there that will have an error or a security violation inside it. I think the worrying thing is that in the future that we are struggling to write that software that is clean and simple. Every time now that I look at code it tends to be deeply integrated and complicated and not straightforward. I suppose it takes that enjoyment away. You don’t get to even come close to a tunnel because you can’t actually give, I suppose, reasonable estimates to managers because they don’t understand software development process unless they were software developers or architects like you. You can only guesstimate or make a guess if you don’t understand what the framework and order library is doing. So it throws designing and building software and these sort of information technology projects, they are getting more expensive and I wonder what are the factors that would cause a business failure. Sorry to an end on that unhappy note but, yes, we ought to be writing clean software, with a clean architecture. How we get there is still very much a mystery to the industry. I will stop there. Thank you for listening.

Paul: Now, unfortunately we don’t have a huge amount of time left because we’ve run over a bit, Andy. Unsurprisingly! But I thought this was really good. The aspect that really resonated with me was the idea that we are a bit rubbish at communicating the need to address legacy, right?! All the companies that I work for have been around for donkey’s years and they’ve all got legacy IT systems that are groaning under the weight of their patches and updates and expansions and all the rest of it. They really need sorting out but we are so rubbish at communicating that to senior management. Communicating the business benefits of dealing with legacy in terms of time saved in the future, in terms of the ability to add new functionality to be more agile as an organisation, all of those kinds of things. We are even terrible at focusing management of the dangers of not addressing legacy. You know, security, vulnerabilities, privacy issues and all these kind of things. I think the problem comes from something that Peter said himself actually in his talk. He said at some point that “developers are always looking for the next big thing,” I think the problem is is that the management knows that, they know that developers are always looking for the next shiny thing and so they are cynical when we turn round and say that we need to replace all this legacy because they just see it as us wanting to play with some new toy. So I think we need to get much better at dealing with how we present this kind of stuff and these kinds of requests to management. Which is what my upcoming workshop is about. Convincing clients.

Andy: Oh, what a coincidence!

Paul: Which is weird isn’t it? But actually by the time this comes out that has already been and gone. So it’s a bit late to promote that now, damn. So I’m guessing Andy that you have got nothing of value to add to this one because it was a very techie one wasn’t it!

Andy: It was a bit of a techie one. But very nicely presented got to say Peter. You made me feel old, talking about way back in the 1980s! You know…

Paul: Ah, well you are old so that…

Andy: Yeah I know, there is that! So no, I thought was very nicely presented, yeah, very well spoken.

Paul: Ooo! So there you go, that’s all he can say about you Peter because he didn’t understand anything that you said.

Andy: No, well you see I can’t help it! I can’t help it. I’m thick as pig shit!

Paul: No you’re not! You’re a specialist.

Andy: (Laughter) Well you see the interesting thing is is that, you know, I seem to have been able to kind of blag my way through this career for nearly 20 years and I know very, very little about the technical background to a lot of this kind of stuff. You know, when I was growing up we didn’t have computers, you know. When I was at school we had one video recorder, one VCR for the whole school and it’s got, you know, wheeled around on a trolley and the jimmies would steal it at the weekend. So, we didn’t have a computer department and I never grew up tinkering with bits of code because it was never my thing. But you know, I like crayons!

Paul: So, Andy are you going to give us a joke to end with?

Andy: I got a couple of jokes actually. Obviously Australian themed…

Paul: …I bet you a tenner they’re both Australian related.

Andy: Yes they are both Australian related and surprisingly they are both repeatable before the watershed because most Australian jokes I think you and I would get into trouble. If I was to tell a joke like for example how many Australians does… How many Australian men does it take to change a lightbulb. If I was to say a joke like that where the answer is: none because it’s a woodsman’s job, they obviously… We would be in serious, serious trouble. So I’m not going to tell any jokes like that…

Paul: Oh, Andy!

Andy: … So I’m going to tell a joke which is “What kind of music to kangaroos listen to?”

Paul: What music do kangaroos listen to Andy?

Andy: They listen to hip-hop.

Paul: Of course they do.

Andy: They do, they do. And I will save the other joke for when Marcus is on the show.

Paul: Okay, thank you. I can’t believe you said that! I’m leaving it in, I’m leaving it in. You can suffer the consequences of your own despicable behaviour.

Andy: I didn’t tell that joke! That was not the joke I told. I told the joke about kangaroos and hip-hop. I refuse to tell the other joke.

Paul: Ah. Right, so, for everybody else that hasn’t turned off because of Andy I would just like to remind you that you can submit a talk because we are still looking for talks believe it or not. You can submit it at and we would love to have you on the show. We have got loads of great talks coming up but we are always open to more. Alright, so that about wraps it up for this week, thank you very much for listening and join us again soon. Goodbye.