Continual Optimisation Using an Evolutionary Process

Paul Boag

This week on the Boagworld Show, we discuss the case for continual optimisation and consider how this can be achieved by building an evolutionary process.

This week’s show is sponsored by Balsamiq and Fullstory.


Paul Boag: This week on the Boagworld Show, we discuss the case for continued optimization and consider how this can be achieved by building an evolutionary process. This week's show is sponsored by Balsamiq and Fullstory.

Hello and welcome to the Boagworld Show, the podcast about all aspects conversion rate optimization, user experience and digital strategy. My name is Paul Boag. Joining me on this week's show is Marcus Lillington. Hello, Marcus.

Marcus Lillington: Hi, Paul.

Paul Boag: Waving. I've told you about this in the chatroom. We are an audio podcast.

Marcus Lillington: Look, I can-

Paul Boag: Just because you used to be some TV star you think, "Oh, I've got to play up for the camera."

Marcus Lillington: No, it's just multitasking. I'm talking and waving at the same time. Look, even you might be able to do that.

Paul Boag: No, don't expect too much.

Marcus Lillington: Don't try it Whatever you do.

Paul Boag: Not today. Not today, definitely not. So how are you Marcus, you all right?

Marcus Lillington: Absolutely fine. I'm going to talk about the weather. So there, because I feel it-

Paul Boag: Nobody cares.

Marcus Lillington: No, but it's been amazing.

Paul Boag: It has been amazing.

Marcus Lillington: It's gone back to being chilly today but it was like the middle of summer over the weekend, which was great.

Paul Boag: It's lovely here.

Marcus Lillington: Well, it's beautiful but it's quite nippy today.

Paul Boag: Well because I insisted on putting so many windows in this new house, the house super heats.

Marcus Lillington: Yeah.

Paul Boag: The slightest sun and everybody cooks inside.

Marcus Lillington: What's it going to be like next summer, Paul?

Paul Boag: Well don't forget I have just gone through a summer, in the summer you can –

Marcus Lillington: Oh you can open the doors.

Paul Boag: Yeah, so it's fine but this time of year it's kind of the minute you open the door you get a freezing wind whistling through the place. So I was very much put in my place, do you remember last week we were talking about Doctor Who and we were talking about how Doctor Who, how it kind of dithers around with the way it deals with racism from the past?

Marcus Lillington: It uses its artistic license all the time.

Paul Boag: Flip me you should have seen Sunday's episode, Marcus. Unbelievable. So so good. So so powerful, it was set in, I need to be careful because I don't want to spoil it for people but it's set in 1955 with Rosa Parks who was the lady who refused to get out of her seat on the bus and it kicked off the whole civil rights movement. So they did an episode about that but it was just so well done and you knew it, I mean from the very very first moment of the episode it was like, okay they arrive in 1955, one of the companions who's black notices that a lady drops her glove so picks up the glove and gives it back and then the husband turns around and slaps him and that's, you just think wow, that is very non-Doctor Who-esque and that whole, and what was so good about it and I really don't want to give away the ending is basically the Doctor didn't get to swoop in and make it all better and it was just so moving and heart wrenching so that put me in my place. Absolutely nailed it, you go watch it.

Marcus Lillington: Okay. I will. One day.

Paul Boag: And talking about things that put me in my place, right? I was also moaning, wasn't I, I was moaning about home automation stuff and how it was a lot more difficult than it made out to be and it was a pain in the arse, do you remember that? You don't remember that, you don't listen to me.

Marcus Lillington: I do, because I'm quite interested in it but you didn't persuade me that it was of any use whatsoever, so I just ignored everything you said and carried on.

Paul Boag: Oh, from that point. Well this week I got a nest, you know, one of those thermostatty things? Talk about unbelievably good user experience. It treated me the prefect DIY level of complete and utter moron.

Marcus Lillington: Now open the bag.

Paul Boag: It was the level of details that the instructions went into about how to replace a thermostat. Remove the cover. Remove the back of the battery cover.

Marcus Lillington: Yeah, crosstalk 00:04:26 That one.

Paul Boag: Pull out this little. But the battery cover back. It was just, I mean it's really really seriously impressed. Good user experience there, so that was quite fun. So all in all, I'm having to apologize for everything that I've been saying the last couple of weeks.

Marcus Lillington: But all it does is mean that you don't have to get up to change the temperature on your thermostat, that's all it means isn't it?

Paul Boag: No no no no no no no no.

Marcus Lillington: No no no no, there's way much more.

Paul Boag: It's much more than that. No, it learns what you like.

Marcus Lillington: But does it really?

Paul Boag: And turns the temperature, and it'll do, you know with like a normal thermostat you have to like, if you say set it, and you want it when you get up, say if you get up at eight in the morning, seven in the morning, whatever, you have to kind of work backwards to work out how long it's going to take to heat up so that when you get up at bed at eight, it does all that for you automagically so that you get out of bed and it's the right temperature because it's heated up, it worked out when it's got to start heating in the morning. And then also when you go out –

Marcus Lillington: So it's connected with your alarm clock, is that what you're saying?

Paul Boag: No no no. I'm saying –

Marcus Lillington: How does it know when you're going to get up then?

Paul Boag: Well it knows because you tell it but you don't have to kind of go, oh well it takes two hours for the house to heat up, so therefore I need to set the temperature at 6:00 –

Marcus Lillington: So I don't have to do a number minus two or minus one. Yeah, okay. Fantastic, wow.

Paul Boag: And I'm just ignoring you now, you're just a cynical old man. And it'll also, what else will it also do? When you go out it knows to turn the temperature down.

Marcus Lillington: How does it know that?

Paul Boag: Through your phone.

Marcus Lillington: Oh, so your geographical location has changed so it, yeah, okay alright that's good.

Paul Boag: And if you've left someone at home who isn't on the system it's got motion detection.

Marcus Lillington: They freeze.

Paul Boag: No no, that's what used to happen with my old one, I used to have a to-do, and that's what happened, they started going, I'd get these calls going the temperatures set off and but no it's great. So it's really good, I was seriously impressed by the nest. So there you go. So there's one thing you can get.

Marcus Lillington: I think if, sorry, don't read the notes Marcus, don't read the messages.

Paul Boag: No it's a good question, Kate's asking a good question. How high is it on the wall, if it's child height be prepared for unpredictable changes. Which is an entirely fair comment Kate, and we don't have children in our house so it doesn't matter. Other than me, who is a big child.

Marcus Lillington: I think if it can stop me using as much energy then that's got to be a good thing.

Paul Boag: Yes, and it does that, it's got all that kind of monitoring and it learns what you want when and it's great.

Marcus Lillington: I have a three year old granddaughter and I can imagine how that would happen. I come in to have the entire Netflix or Sky or whatever it is now all in Japanese. How did that happen? I dunno, she says.

Paul Boag: Yeah, that's fine. You want Japanese TV you just don't know it yet.

Marcus Lillington: So, hm.

Paul Boag: So yes, the one I want next while we're on my home automation thing is on bloody hot days like today I want it to detect that and automatically open my skylight that's above me, that's the next thing on my list.

Marcus Lillington: Oh I see what … but not keep the heating going, turn the heating off and then …

Paul Boag: Yeah, turn the heating off, open the skylight.

Marcus Lillington: Okay. Yes, I have in answer to Malarky's question, Andy, hello Andy. His question, have you replaced all your bulbs with LEDs? Yes I have.

Paul Boag: Oh so you're at least that far along the road then.

Marcus Lillington: Mm-hmm (affirmative).

Paul Boag: So, but they're not smart bulbs one presumes. See. Amateur. See I can sit here and I can go Hey Google turn off the lights in the study and it'll turn off all the lights.

Google: Okay, turning off five lights.

Paul Boag: Which you probably don't notice because it's so frigging bright in here anyway.

Marcus Lillington: As I say, when I can't use my legs anymore, that'll be useful.

Paul Boag: Alright, you're just. So rude. I'm just trying to drag you into the 21st century.

Marcus Lillington: Yeah but we mustn't all be like sheep and just do everything that everyone says that we should do. You need to look at these things critically, Paul.

Paul Boag: Alright. Here's one. This one's gonna convince you.

Marcus Lillington: Right.

Paul Boag: Get a robot vacuum cleaner.

Marcus Lillington: I've heard that's fairly good. We've got lots and lots of little steps in our house. Some of them only like an inch, maybe.

Paul Boag: Yes you do.

Marcus Lillington: Some of them like proper steps from one room to the other.

Paul Boag: I mean it won't go over the steps, it'll detect the steps so it's not going to fall down but it's not going to be able to then, you're going to have to pick it up and move it from one room to another. But they are pretty good I have to say.

Marcus Lillington: Bearing in mind we have two big hairy hounds, hoovering is a thing that happens a lot in our house so that would be good. But our floors are all different types and they're different heights and that's the –

Paul Boag: The heights is a problem, the different types is not. So the one that I've got actually will detect a carpet and turn up the, you know the –

Marcus Lillington: The brush.

Paul Boag: The suction, that's the word.

Marcus Lillington: Oh, oh suction, alright. Normally just the brushes go up a bit.

Paul Boag: And also when it's on hard floor it will mop as well if I want it to.

Marcus Lillington: Mop? It's got water in it?

Paul Boag: Yeah. Yeah.

Marcus Lillington: Hey.

Paul Boag: I know right, see? I'm winning you over?

Marcus Lillington: Hmm, I quite like the sound of this. So yeah, and everyone on the podcast must be really into this. This is more boring than web design, don't you think?

Paul Boag: No. No. Because it's all about in my opinion, bringing this back around to our subject matter.

Marcus Lillington: This futurist kind of thing.

Paul Boag: It's the future of the user interface, which is no interface. For example with that vacuum cleaner I can just tell Google to start the vacuum cleaner.

Marcus Lillington: However, and there is always a however, coming back to this idea of getting a nest would make me use less energy, exactly Andy, what a riveting podcast this is. It's high brow Andy, you wouldn't know what that means. He posted a two and a half thousand pound suit on Twitter earlier and said that –

Paul Boag: I know. I know and you think that I'm wasting money on a vacuum, no a suit's better than a vacuum cleaner I can crosstalk 00:11:35 but it's cheaper as well. But you could get four vacuum cleaners for that at least.

Marcus Lillington: How much energy is the Robot Vacuum cleaner using compared to a standard one. If it's more, you probably shouldn't be using it.

Paul Boag: Ah yes, but in our house we have an entirely renewable energy company.

Marcus Lillington: That doesn't … no … you're still using …

Paul Boag: What do you mean no?

Marcus Lillington: You're still using more energy though?

Paul Boag: Yeah but it's all coming from solar. It's all coming from solar or coming from wind fongs, or water.

Marcus Lillington: Yeah, they say it is. I don't believe them. Mine is too.

Paul Boag: What do you mean you don't believe them? I think possibly somebody might have prosecuted them if they were just making that shit up.

Marcus Lillington: But what happens if, there's only so much power produced through wind and sea and whatever. So if that runs out they'll take it off the standard grid, surely. They won't just, crosstalk 00:12:41

Paul Boag: No no, basically it still comes off of the standard grid. So my, they don't ship in special electricity to my house.

Marcus Lillington: Exactly, exactly.

Paul Boag: But what they mean by that is that every kilowatt or whatever that they sell they produce the equivalent kilo wattage via renewable sources.

Marcus Lillington: Okay. I still think we should be, yeah I still think we should be using less of it.

Paul Boag: Pffft. I'm looking forward to how the transcriber is going to transcribe pfft.

Marcus Lillington: Brilliant. Anyway, so I'm bored of that conversation now, Paul.

Paul Boag: How could you possibly be bored of it?

Marcus Lillington: Pffft.

Paul Boag: Alright, we'll talk about web design. What boring arse thought for the day have you got then?

Marcus Lillington: It's a continuation from last week's. I've got to find it now, what have I done with it? Somebody asked, Ryan Taylor asked the question, how do you … I was talking about doing website reviews if you remember, Paul. It probably fell out of your brain immediately after the call, the podcast. And Ryan said, but how do you get people to pay for these things. So I thought about it.

Paul Boag: With money is normally the preferable way. Or bartering with chickens. How many chickens do you get for a discovery project?

Marcus Lillington: That would be quite a lot, actually. Come back to that crosstalk 00:14:05.

Paul Boag: I would have thought so.

Marcus Lillington: Yes, that's –

Paul Boag: It's got to be at least a cow, wouldn't it?

Marcus Lillington: I'd reckon a cow, yeah. That's pretty good, yeah.

Paul Boag: I don't know how much a cow is worth. You carry on I'm going to find out how much a cow's worth.

Marcus Lillington: Lots of money.

Paul Boag: Is it? How much is a cow …

Marcus Lillington: Well bye, bye Andy. Right, so.

Paul Boag: He got so bored that he's leaving.

Marcus Lillington: It's because I insulted him. I didn't mean it. Anyway he's gone now. Right, so I think it's a little bit more than just website reviews, I think we need to, it's a bit wider than that. It's about convincing clients to spend money on research generally or even harder, potentially convincing them to spend money on scoping, specification, that kind of thing. But I'm going to talk about that next week, probably. I'll get my thoughts together on that. So I'm just going to talk about getting clients to spend money on research. The first thing you need to do and this is, we're assuming that we're talking to a potential new client here.

So the first thing you need to do is ask them what research they've already done. Have they done a competitive review, have they done any testing? And if they have, please can you have a look at it? That's really important. Because if that research is good, don't propose to repeat it. You obviously need to acknowledge it and say that it will make up part of a more complete package that you're going to propose. And I suppose if all the bases that you're going to propose have been covered, then you would just move directly on to the next phase of the project. But that's never happened in my experience. There's always something that you need to investigate further.

But it might, it's quite possible that a client will already have done a bunch of research and you need to kind of bring that on board. And the reason why I'm saying that here on the subject of convincing clients to spend money on reviews is they almost certainly, they will have put effort into those previous research they've done. Even if they did it in house they'll realize that it's something that takes effort.

And if you're talking about it as a general thing, then you're making the case for doing research as part of this new project, we're going to take on board the stuff you've already done and you're going to have to pay for this bit we're proposing, okay? So that's part of the convincing exercise.

So, specific to a site review, I'll add one to a proposal if I think it's an appropriate thing to do. I often do. And I sell it, and this has kind of come about over the last few years, we've been doing site reviews for decades now but over the past few years, I've been kind of selling it in air quotes as part of a three pronged approach to research, reviews, recommendations, that kind of thing. And those three facets or approaches or whatever are one, users. IE, we want to get input from users, interviews, surveys, testing and it's obvious why we'd want to do that. Second one is internal stakeholders. We want to understand objectives, the character of an organization, understand it's aspirations and get their take on users as well. Everybody knows that.

The third one is to get kind of an expert user viewpoint. Someone who's been doing this kind of thing and is into UX and has been doing that kind of stuff for years and years, get that viewpoint. Kind of you know, the best practice viewpoint if you like, because we can bring that. So it's providing a critical eye on things like design and copy that you won't get from the first two groups. Maybe a little bit but we're looking some more detail on those kind of things so again, you're selling it as part of this kind of package of research, some of which might have already been done, other bits won't have been and that's a site review we kind of bring it as you're adding a third input into the research process.

We also look at analytics and competitors as part of our expert review. And I guess it's just another part of the discovery phase armory that we have and it's something that our clients actually have been honest doing. I can't think of … I can think of one recent project where people have said I don't want you to do the expert review, but you can do all the other bits. But that's really rare so it's something that it seems that I don't need to do a lot of persuading on but maybe that's because of all the things I've just said.

Paul Boag: I think another thing, sorry, I thought you were wrapping up, Marcus.

Marcus Lillington: I am wrapping up. This is my last sentence, then you may talk Paul.

Paul Boag: Thank you, thank you for your permission.

Marcus Lillington: Anytime. But going back to what I said last week, however, it's important, A, not to state the obvious and B, to provide recommendations on the issues that you highlight, otherwise I guess it's possible if you just provide boring expert reviews that just say "your site is bad, this is really obviously bad stuff" then you might get a reputation for producing site reviews that aren't work spending money on. You need to make sure that they are actually good quality if somebody is going to recommend you to a colleague or a friend or something like that.

So that's it.

Paul Boag: I mean the thing that I find works well, when any of this kind of upfront discovery work whether it be you know, a site review or indeed anything else, is I position it as this is, I do it as a project in it's own right is the way I approach it, right? So we're going to do, before we do the main contract and the main agreement we're going to do this upfront piece of work. I position it as a chance for you to try the relationship with me. Right? And see whether you think I've got, you know, have a look at what I've produced, is it good quality, is it useful?

so it reduces your risk because you have a small outgoing upfront and a chance to walk away at the end if you're not happy. I also say, it's … I always make it very clear what the deliverables are going to be and that those deliverables have value in their own right, right? So, you know, they could take the site review and go and get another designer or developer to implement it if they wanted to. Or they could, if I'm helping doing some discovery and some background, you know, the deliverable there will be a better brief that enables people to quote more accurately on your final development cost and you'll have a better idea of what's going to be delivered at the end of it rather than a wooly brief where it's all a bit vague, do you know what I mean?

Marcus Lillington: That's my … that was the other point that I might talk about next week if I can think of enough to talk about it which is that … paying, a client paying for you to scope their project, or specify that project. Because that's different to doing a review.

Paul Boag: Yes, it is.

Marcus Lillington: And that is a different thing and I think when you're going to do that you have to make it a separate thing, like you're saying, Paul. Otherwise it's just fantasy, isn't it? We can price this bit but all the rest of it we've got no idea because we haven't done the specification yet, though I guess it depends on how good a brief is when it comes in. Usually it's, these days briefs are getting better and better in my experience.

Paul Boag: They are, yeah that are, but It's sometimes when that brief lacks things like user research or prioritized business objectives, those kinds of things, which they still often do if we're honest. So it's not necessarily a big project to help them scope it but there is some work to be done there. But we won't get onto that because you're going to cover it next week. You have to now, you've got no choice.

Marcus Lillington: Now, mind you it's going to be weeks before we record the next one, so.

Paul Boag: Yeah, yeah because I'm jetting off around the world, that's how rock and roll I am or something. Although actually I've got two weeks of jetting off around the world and the third week, where I think we are actually recording a podcast, I'm officially on holiday so I can play Red Dead Redemption 2 in my pants, probably. Allegedly.

Right, let's talk about our first sponsor, which is Balsamiq and actually talking about the idea of site reviews, if you would like some feedback on your website and are too tight to pay Marcus for it, you can actually get me to do it for you, for free! Well I say for free. Balsamiq are talking about paying for it, which is very nice of them. Basically they've asked me to review websites and give my first impression of those Websites, so I've already recorded one for a web design agency in Australia. And it's actually, it's amazing how much you can cover in 20 minutes of me waffling on a video.

So we do publish these afterwards so that everybody can benefit from them. I don't think that one has been published yet but they're very keen to get me to do some more, so if you would like me to review your website then go to It's really great the way that Balsamiq are indulging my rants. So they are effectively sponsoring me to rant, both through the podcast and videos, which is very nice of them.

Okay, let's talk about this week's topic which is really the kind of case and implementation of continual optimization. So it's this idea that we should be continually optimizing our sites rather than doing periodic redesigns of them. And nowhere is this more important than when you're talking about conversion rate optimization which is an incredibly important part of the equation when doing conversion rate optimization is to ensure that you are continually optimizing your site.

I think probably the best place to start on that kind of conversation is by looking at the damage that periodic redesigns do. So most organizations do a redesign of their website every, what, three to five years? Maybe a bit shorter than that, it kind of, it will vary a little bit but something along those lines. And this is a bad idea. It's a bad idea for all kinds of reasons. For a start, your existing customers –

Marcus Lillington: Sorry Paul, I've got to interrupt. 12 years, current, we're working with.

Paul Boag: 12 years since they've redesigned their site.

Marcus Lillington: Before the iPhone it was. Before president Obama and there was another before. Well … crosstalk 00:25:11

Paul Boag: That is unbelievable, I've never come across one that's quite that long. But that depends whether have they been maintaining it in the in between time and evolving it?

Marcus Lillington: A bit more of just like bolting on extras because it's so old and unusable.

Paul Boag: Right, okay. So, anyway.

Marcus Lillington: Sorry, crosstalk 00:25:28.

Paul Boag: It's generally not a good idea to do a redesign but it's also not a good idea to leave your website for 12 years without really doing very much at all. The problem with periodic redesigns for a start is that it can really throw existing customers. So if you have repeat traffic that's coming back to your site regularly, people are going to learn the eccentricities of your site and even if you improve it, if you do a big redesign you change everything in one go it will absolutely throw them and actually will damage your conversion rate.

But there are more kind of fundamental problems, even than that with this kind of boom bust cycle of periodic redesigns. For example, it's typically incredibly wasteful because you're throwing out everything, right? When you do a redesign you normally throw out everything, the good, the bad, the indifferent, it all gets chucked out and you start again, which, okay you might migrate the content and the information architecture but even that, it's probably that's the last bit you should be migrating in most situations in my experience. So you're throwing out perfectly good stuff along with the bad. The other problem that you've got is that for a big chunk of it's life cycle, the website isn't optimized, isn't running at peak performance. It is when you initially launch it and then very quickly it gets abandoned and it becomes less and less fit for purpose, content becomes out of date, the technology stack doesn't age well, the design begins to creak as more and more stuff is crowbarred into it.

So, most of its life, your website is not really very efficient, it's not working as well as it should be. And then finally when you do do these redesigns, they are a capital expense. In other words they're an unpredictable expense that comes along every three to five years and they're a big cost for the business. It's much cheaper and much more predictable if there's an ongoing operational expense to maintain and keep the website up to date.

So periodic redesigns are not a good idea. But you know, incremental improvement, improving your website over time, now that makes a lot more sense because your website's always going to be operating at peak performance, you're going to make significant cost savings because you're not throwing everything out and starting from scratch every time. And you're going to know what it is that you update that's actually going to prove beneficial.

Again, another problem with throwing out a whole website and launching a new one. If that new one outperforms the old one you're not going to really know what it was about the new site that's better than the old one or heaven forbid that the website performs less well than the previous one you're not going to know where the problem is because you've changed too much in one go.

so incremental improvement allows you to know what is working and what is not. So obviously better for your cash flow as well. And the decisions that you're making about the website are data driven rather than just on opinion and senior management shouting about it. And of course when you're updating a website and refreshing it on an ongoing basis, it keeps you coming back, didn't it? It's like every time you see a major app, one of the apps on your mobile is a major release to that app and it updates it you go back and have a look, don't you? You go and see what's changed.

So, part of the problem is that we need to be kind of optimizing on a continual basis but there's a kind of another part of it as well which is that the way that we traditionally run projects is failing in the digital world. And the problem is that we think in terms of finite projects with a beginning, a middle and an end, right? So that's part of why sites aren't continually optimized but are rather redesigned periodically. But alongside that project mentality is that kind of waterfall approach. That somebody in senior management decides the website needs redesigning. A big specification is drawn up and sent out to tender. Companies like Headspace tender for it, there's a competitive tender bill, a supplier is selected, a statement of work is written, and then when the project begins, it's all about avoiding scope creep because the budget's been set, there's a contract in place and nothing can change at that point.

And then when the website's delivered, everybody goes on to the next project and kind of walks away. And that is just not suited to the digital world and there are kind of three reasons for that. One is that that methodology of avoiding scope creep and doing a big specification up front is kind of comes from this world of I don't know, building something physical, like a building, a factory, right? You don't want to get halfway through building a factory and suddenly go oh shit, we're doing it wrong, we've got to change>

And that's because of the cost that goes into building that kind of thing because it's not just labor, you've also got the cost of raw materials, your cement, your girders and whatever other things are involved in building a factory. Showing my ignorance. Well, by contrast in the digital world, pixels are free.

Marcus Lillington: The walls and the roofs and things like that.

Paul Boag: The raw material is free. What's that?

Marcus Lillington: The walls and roof and things like that.

Paul Boag: Things like that.

Marcus Lillington: Gates and doors. They're real things, aren't they?

Paul Boag: In the digital world, all those kind of raw materials are free. And then secondly, once, imagine a factory where all the raw materials were free and you were designing a factory and then when you built the factory you then got real time data flooding back about how efficient that factory was, right? See if you got that in real time, you'd start tweaking the factory, wouldn't you? You'd move a wall here, and change the production line and see if you couldn't increase the data that you're getting back and that is where we're at with digital. We get so much data yet the stupid thing is we stop paying attention to that data once the website has been redesigned and relaunched because everybody moves on to another project.

Marcus Lillington: I don't think it's as bad as it used to be, Paul.

Paul Boag: No, it's not, it's not. But you still see it, because even you can look at this on a macro level, right, so an entire website redesign but I was meeting with a large pharmaceutical company this week, right? And they of course, they're constantly iterating and improving their websites so you think great, they've got it. Until you really dig into that and what you discover is actually what they're doing is running one campaign after another, right? So they're replacing redesigning the whole website with we're going to create a landing page and do an email campaign, we're going to send it out and then we're going to walk onto the next campaign and never actually monitor whether that campaign was effective or not. So it's not as bad, but it's kind of happening at a lower level if that makes sense with you know, individual parts of the site. There's no, they're not paying attention to the data in quite the same way.

The final problem with this kind of waterfall type approach is that the web is bloody complicated, right? And involves a lot of different specialists working incredibly closely together, right? There has to be a lot of integration between the developers, the designers, the content people. And if you're just flinging the project over the wall to the next person down the chain, that's not really going to work. So what we need to be doing instead is a more, an approach that's more similar to the scientific methodology of, we form a hypothesis about how the website could work or how the campaign could work or whatever it be, you then test that and then you iterate upon it based on what you learn.

Now I know what you're thinking, you're thinking, oh that's agile, right? And we do agile already and I get this all the time when I talk about this at the moment, oh yes, we do agile. But do you really? Right? Because in my experience, there's a lot of companies that say they do agile, right? And do not. What they mean is oh yes the development team does agile. But if you, every other part of the process is not agile, if there's still someone in senior management who is defining the project, someone then creating the specifications, someone then designing it, then you have a bit of agile going on when it's developed and then people pouring the content in. That's not agile. That's just waterfall with agile kind of shoved in the middle somewhere. So I think there are more fundamental problems going on here organizationally.

But Marcus I fully accept what you're saying, it is better than it was five years ago even.

Marcus Lillington: Yeah, it's kind of like what does a redesign mean these days? I mean we've redesigned the Headspace site, we kept most of it but it looks very different and I think that's because a redesign of a 12 year old site does mean throwing it away and starting it again but I think redesigning a two, three year old site doesn't have to. It might do. I'm not disagreeing with any of what you've just said. I just think it's not as maybe as black and white as it once was.

Paul Boag: Of course and sometimes a redesign is absolutely necessary. If your foundations are not firm, then yeah, you're going to have to change everything in one go.

Marcus Lillington: And I think that, sorry, one of the things that I think is really important here because what I think what you're saying about real, proper agile, about the idea of having a digital team that has the power in itself to say we're going to, the data is telling us this so therefore we're going to go and work on this. I, yeah. It takes … a lot of things need to be in place for that to happen all the time. For starters, senior management needs to be brought in, if it isn't, it isn't going to happen. Because they'll go, what are you doing? How much are you going to spend on this? When is it going to be finished and all the usual kind of, those kinds of things. And if that is the case then that shouldn't necessarily mean that no web development shouldn't happen, do you see what I mean?

It's kind of like if you've not got the perfect team structure where people could work … that's where agencies like Headspace tend to come in. Because we'll come in in situations that are politically difficult and that people aren't maybe getting things done that they want to get done so they'll hire the likes of us to help them make that happen. We're not needed in a situation where there's a fully specced digital team that have their futures mapped out by the data they're getting. But unfortunately that's not the way it is in many many places in the world.

Paul Boag: No, absolutely and you're entirely right, you know. You make do with what you've got don't you and you work within the constraints that you've got, this pharmaceutical company that I was meeting with is a classic example of that, that they had certain constraints placed upon them via global and in fact they'd had too many consultants come in that basically told them all the theory like I've just spouted now and then are not better off at the end of it. So they have to work within the constraints they have.

But there's a balance to be struck there, you know. If you just always work within the constraints placed upon you, you're never going to get the best.

Marcus Lillington: Yeah, and you're always gonna get a website that dies at the end of the once we say bye, we're done.

Paul Boag: Yeah, absolutely.

Marcus Lillington: It has to, and that's not good for us. So we need to be encouraging people crosstalk 00:38:04.

Paul Boag: Because the agency often gets the blame for that.

Marcus Lillington: Yeah, oh yeah. And I don't mind it if we are used as scapegoats and that's fine to a certain extent, but you don't want to get the blame when it really isn't your fault. You know? And that would be a case of tha.t

Paul Boag: So it is a matter really of moving … you want to move things on a bit. You can't just sit in the status quo and say this is the situation we're in, we need to live with that. I think you need to know where you're heading, right? And you need to take little steps towards that and incrementally improve things over time.

So you're totally right pulling me up on well this is all well and good in principal but it's that balance isn't it?

Marcus Lillington: crosstalk 00:38:50 yeah absolutely.

Paul Boag: And that's a fair comment. You're absolutely right in doing that but you also need to be pushing forward. It's getting the two working simultaneously. And I, for me the key to that is starting shifting the decision making to be based on data rather than on senior management decisions. And even that is a slow road because it's how you show it, bring the data back to management without turning around and saying your idea was stupid, we should be doing this instead, look the data says so. You know? You've got to do that in the right way as well.

But talking about data, that brings me on very nicely to Fullstory which is our second sponsor who is a session recorder and it allows you to replay sessions so it's just an incredibly useful tool in your designer's arsenal to give you a broad look at what's going on with your website, right? And yeah, great, analytics are wonderful, they certainly give a lot of insights but nothing beats that personal understanding watching real users interacting with your pages in real time, and that will uncover a lot of areas that you can improve on your website without necessarily needing to redesign the whole thing.

And it will give you evidence you can take to management and those kinds of things. Session replay uncovers a whole host of UX and general design issues that can't really be sussed out from really anything else, even usability testing often overlooks these things because people are giving pages a lot more attention in a usability test session than they are in real life.

In terms of why you should use Fullstory, well it offers the highest fidelity session replay available because it indexes absolutely every single event on a page and you can search and segment any action, so you can see users that furiously rage click on a button and then instantly view what they did and everything around that and you can get insights that overlays data on pages to tell you what's going on as well.

Fancy giving it a go, you can sign up and get a free month of the pro account for free, no credit card is required and you can continue for free up to 1000 sessions per month after that free month of the pro account or of course you can jump in and sign up. To check it all out and to give it a go, try

So. On the basis that we want to move to a more iterative process. What does that look like in practice? So from my point of view what you're talking about is a kind of five step process, and this applies whether you're talking about redesigning an entire website or doing an individual campaign or adding a new feature, or a new landing page. The process is basically the same. In all situations. The beginning of that process is to do a discovery phase. Or actually I don't like the word discovery phase if I'm honest. The problem with the word discovery, it basically makes it sound like we've no idea what we're doing or what this project is about so we're gonna flail around and waste some time working that out.

Marcus Lillington: Research.

Paul Boag: Yeah, research, specification phase, anything like that is better in my opinion than discovery, but that's …

Marcus Lillington: It's a bit Disney, isn't it?

Paul Boag: Yeah, let's do discovery! And it particularly falls down if you're an agency, right? Going to work for a client. Because if you're going into a discovery phase, you basically make it sound like you're going to be paying us money to work out who you are and what's going on in your company. And nobody ever wants to pay for that, do they?

Marcus Lillington: No.

Paul Boag: Right, so we're going to give it a different name. But essentially in that phase you are researching your audiences, understanding what needs they have, prioritizing those audiences. Doing a little bit of work around what the business objectives are, prioritizing those business objectives and then defining the shape of the project, what success looks like. I'm not talking about necessarily working out all of the functionality or anything like that but what should success look like? How do we know if this project has succeeded or not?

Marcus Lillington: It's gathering data, isn't it? That's what going back to the previous, it's all about making decisions based on good data. That data doesn't necessarily have to be numbers. It can be inputs of any sort.

Paul Boag: Yeah. And so often it's good to kind of define a project in terms of user needs. You know what does a user need to do for this project to have been a success and also what does the business need to have achieved in order for it to be a success? So doing that upfront. Then whether you're doing a big redesign, an email campaign, whatever else, then you move into a proof of concept stage or a prototyping stage. So this is where you begin to shape a possible solution based on what you've kind of learned in the discovery phase for want of a better word. So this is an interesting one because this could be a different levels of fidelity depending on what the project is and what you're doing, right? So if you're doing something like, let's say you're doing an email campaign. Your proof of concept phase probably would consist of doing two or three different versions of the email campaign and then sending those out to a segment of your audience to see which one performs the best before you send out the full one. But the principle is still the same.

If you're redesigning a whole website you might go through an iterative process of you do a low fidelity wire frame, maybe run a workshop that wire frames some key pages, maybe you do some sketches and then increase the fidelity over time. So step two, this build a proof of concept kind of merges in with step three, which is testing and iteration. So it's this idea that you prototype something very basically, you then test that and iterate and improve upon it. And this is an absolutely fundamental if you're talking about conversion rate optimization, right? Because it doesn't matter how long you've been doing this shit, and I have been doing it a long time, I still never get things right the first time. Okay? It just doesn't happen. You don't come up with the perfect design, the perfect copy, the perfect converting machine out of the gate, it does not happen, right?

So if you are not testing and iterating upon that then you are leaving money on the table, you are leaving conversion over the table. And I had the best most encouraging conversation I think I've had in ages with you remember I was, I talked to you about the company that I've been doing design for for the first time in ages, right? And I had a conversation with one of their directors and he came back to me and said I love what you've done on mobile because they've got really shit mobile experience on their current website and they know they do.

I love what you've done on mobile it's really really good but we're a bit worried about the desktop version because it's quite different from the current version and we're worried about the risk of our conversion rate dropping if we switch across and it was like yes, someone thinking about risk management. So we talked about well okay we can do some usability testing and we can see how it performs with the baseline of the existing site and then we can evolve it. So you kind of go through this and I was talking him through this and then that leads us on to step four, which is I said to him, well even once we are happy with it, we don't just switch it over one day and suddenly make this site live, right? Instead what we can do is we can roll it out to, we can put it online and send a certain percentage of the traffic towards the new site and see how they get on with it in a real environment. We can add a banner to the existing site that says hey, why not try our beta.

And see how you get on with it. And then on the beta site, on the new site we have a banner that says go back to the old one if I hate it. You know? And then you get an idea of how real users are doing this.

So it's that idea of selectively rolling stuff out and that doesn't, none of this applies just to a big site, it also applies to anything. You can do this on a landing page, you can do this with an email marketing campaign, even with a social media campaign, this idea of testing and iterating, rolling it out to segments of the audience. And then, once you've gone live with it, even once we've pushed this e-commerce site live that I've been designing, we don't even stop there. Then we start monitoring it. We monitor it over the long term, improving and iterating it over time. And that's the basic structure. But I think the important thing which goes back really to Marcus's more pragmatic view that he was talking about earlier is how much you will do every, you should do every single one of those five steps but to what degree is up to you, right? So a discovery phase, a research phase could be as simple as having a chat with a few salespeople that are meeting with users every day and that's enough.

Equally your proof of concept may never get beyond pen or paper, right? And just a little sketch. Or if it's IA work you're doing it might be an excel document, you know? It doesn't need to be rocket science. And the testing, that could be testing just design comp. It could be testing with anyone outside of the organization, three or four people, it doesn't need to be heavy duty. And the selective rollout, you might conclude well actually that's not a big deal in our case because we've got such a low level of traffic anyway that it doesn't matter, in which case, roll it out in one go but just have the ability to roll back if something goes horribly wrong. You know? So all of these things are scalable but it's that basic five steps, research, proof of concept, testing and iterate, selective rollout and monitor and improve which you kind of want to get into the rhythm of.

Marcus Lillington: Cool.

Paul Boag: Have I been too theoretical and ideologically pure there Marcus, do you need to slap me down again?

Marcus Lillington: No, not at all. I like all of those things very much.

Paul Boag: I have Marcus approval.

Marcus Lillington: There is a link here, Paul in chatroom has a question. Quite a long question, I shall read it all out. Do you feel that ongoing iterative improvement means redefining the relationship and purchasing processes for web services? For example, maintaining an ongoing relationship rather than the typical purchase cycle and then abandonment that Marcus referred to.

Paul Boag: Yes.

Marcus Lillington: It depends on the team, well on the size and the skills involved in the team that have hired those web services. These, I mean and I think … it's rare that we'll come across a web team that has less than three people in it now these days. Normally it's quite a lot more than that. So therefore, we are able to be kind of hired for a section of the work, maybe part of what Paul's just described and then they can take it on board. If they can't, then they should be bring us on a more ongoing, I mean most of our clients are on an ongoing basis anyway, but it's who's got the skills available to ensure that all this stuff Paul's been talking about happens.

And it might be that you need to hire people external people, but if you've got skills in house you don't.

Paul Boag: The procurement process is another big part of this, as well. Because obviously you don't want to have to jump through a full procurement process every time you get somebody to do some work on your website. And that's often a big problem and especially with larger organizations they have the checks and balances in place to avoid this. So one approach is that yes you get them to do this for the first project and then they move into some kind of preferred supplier status or something like that in order to avoid this having to happen again and again.

And often times it's just a matter of sitting down with the procurement people and actually having a conversation about it and explaining the situation. Often I think people take procurement rules as being written in stone and nobody can ever change them and that's not actually true. The other big problem area Paul, often actually is the way finance operates within the organization. They think in terms of these capital investments all the time, rather than operational costs. Most organizations are trying to lower their operational costs and rely more heavily on capital investment and that can cause problems sometimes that again you need to have a conversation with them and go okay, how are we going to manage this in a way that allows us to be agile enough and responsive enough rathe than having to go through a three month sign off process to get budget every time we want to make a change to the website?

So absolutely of course there have got to be checks and balances in place to stop you know you going in and spending 10 grand with your mate that does websites. But on the other hand there's also got to be … we've got to change the culture a little bit to be more agile and more responsive than maybe we currently are. It's a good question.

Marcus Lillington: Just on the subject of procurement. I understand accept entirely that procurement departments are a fact of life. My biggest bugbear with them is that I can't talk to people properly before the process of pitching for a piece of work. And some large organizations who have very very stringent procurement processes do that have that. They will encourage meetings where you go and sit down and have a chat, you can kind of work out whether it's for us or whether we fit and get an idea of scope and all those kinds of things. They work. And because sometimes you can get one that actually it's not for us or whatever, and then sometimes you don't know and you put loads of effort into something that you don't know whether you're going to win it or not so. Or you don't whether you even have a chance of winning it so being able to have that chat upfront is what needs to change from my point of view.

Paul Boag: Yeah, you know me, I've got a step further, Marcus. I just won't have anything to do with it at all. I don't go through procurement processes, but that's partly my situation that I'm a one man band and can't afford to, to be frank. You know, you spend half your life doing pictures and not actually doing any bloody work. But it's a difficult one. It is difficult because I do understand the need for it but it is incredibly problematic. And the trouble is that half these times these procurement processes, people find work arounds for them.

How many times has Headspace been, well can we keep it under, can we keep each invoice under this amount because then it fall below our procurement threshold?

Marcus Lillington: It depends on kind of, if you're dealing with somebody who is really senior who has been around in whatever organization it is for ages then they'll just make it happen. It's, um, yeah. Nothing more to say on that. I think it's … and that actually comes back to this kind of idea of you know choosing not to go for, not to become involved in it.

We have to go for projects of a certain value and if you're over a certain value then they're all, well not all, most are behind some kind of pitching process or procurement process and as I say, it's just a fact of life, it just would be nice for us if we could do more just talking with people before hand. That's it.

Paul Boag: Okay, so I think that about wraps that up. We've gone on somewhat of a tangent here but I think it was a good tangent. If you want to know more about this kind of whole website redesign malarchy and my opinions on it I have written a post at if you want to go wave that under some senior manager's nose, not for all the good it will do. Marcus, do you have a joke to wrap us up with?

Marcus Lillington: This is a Bruce Lawson joke, so I imagine most of the listeners have already heard it but I'm going to say it anyway. Just spend 300 pounds on a limousine and discovered that the fee doesn't include a driver, can't believe I spent all that money and have nothing to chauffer it.

Paul Boag: Oh. Oh I quite like that one. No that's a good one. I like that. I approve this joke. Excellent. Alright, I think that wraps us up for this week. If you are interested in the masterclass from which this season is based, then go to, otherwise we'll see you next week for our penultimate episode in the season. But until then, thanks for listening.


Speaker 5:

Thanks to spawn101 from Shutterstock for allowing me to use this image.