Personalisation - Big gains from small steps

And you add in some components for the other types of users lower down the page and those at the bottom of the page are a little bit neglected. So what personalisation allows us to do is A to understand the needs of our user, who they are and what they're after, and B to present them with a message that's more relevant to them, that's going to allow them to convert, that's going to allow them to achieve their goal as quickly as possible. And previously we couldn't do it, we used to have this one size fits all approach and in trying to talk to everybody on a single page we ended up not really talking to anybody. So personalisation fills that gap. So what is it? What exactly is it? There's two types of personalisation really. Explicit over on the left is the kind of personalisation that a user is expecting to see based on some action that they've undertaken themselves. So they've created an account. They've logged in. They've set some user preferences. They're now expecting to see slightly different content on the website, maybe personalised newsletters, campaigns, that kind of thing. They fill in a form and then they see a message with their name on it and something that's appropriate to them. That's explicit. The user's expecting to see it's based on something that they've done. The personalisation that we're talking about today is implicit. It's behavioural. It's implicit because it's based on an implied understanding. The user isn't expecting to see it. We are presuming to understand the needs of the user based on some kind of behaviour that they've undertaken on the site. And therefore we're trying to drive some messaging, some relevant, contextually aware messaging to them. And that's the kind of personalisation that can really help drive our conversion. So it's the implicit type that we're talking about today. Wikipedia defines personalisation as... So it's a service or a product to accommodate specific individuals, sometimes tied to groups or segments of individuals. So, in digital, in websites, we're talking about tailoring a page suited to a specific segment or an individual. So let's give that a go. Let's try it out. So we have a separate offering, Mayfly. We have our Mayfly media offering, which is very much a digital agency kind of project-based approach. And we have a separate team, separate developer, separate account manager. account manager, separate processes for support and maintenance of existing Umbraco sites. So we have a Mayfly support website. Now I think that some of our users might be interested in hosting too. That kind of sounds like a nice fit for our Mayfly support offering. So I'll go ahead and implement this little JavaScript plugin from a personalization tool and create a variant of that page. So I now have this support and hosting for Umbraco variant for this segment of our users that I think exists. We have our default variant at the top here. So two variants. So is this personalization? Yeah, I've created a variant for a segment that I think exists. But what have we achieved? I don't know. The tool was so easy for me to integrate, I just plonked it on and gave it a go. So I'm not quite sure where we are with it. Do I actually want to increase hosting sales? As a business owner, is hosting something that's going to drive our growth and our targets for this year? I don't know. I haven't thought about that, really. I just wanted to give it a go. Are our users interested in hosting? I don't know. I didn't look into that either. I kind of thought it was a nice fit, but I didn't really speak to them and understand what they were after. Has it been a success? I don't know. I didn't really know how many hosting sales I was doing before, so I can't really compare whether I've increased it or not. Have we actually decreased something else by creating this variant? Have we decreased our support and maintenance sales? Again, I don't know. I haven't planned it. Shall we continue with it? Not sure. Shall we try something else? Okay. I'm stuck. Take it off. Personalization hasn't worked. I don't know. I'll do something else. Maybe try it again later. This is where 90% of personalization efforts fall down, with a lack of planning and strategy. I didn't think about it. I stuck the tool in. I created a variant. And off I went. And got myself into a bit of a muddle. I didn't know if it was working because I hadn't planned it. One of the other problems you can run into is increased complexity. So here's a wireframe of a charity website that we built recently. Now it's a component-based design. These four components, a banner and these three content components here. I've decided that they are going to be suitable components on my webpage for personalization. I'm going to create two rules per component. So there's going to be a default rule. So users that don't match any of the personalized conditions when they hit our website just sees these variants of the components. But I'm also going to create a personalized variant of the components as well. So you will see something else in these areas. Now, very quickly, I've created 16 different experiences. I've gone from one homepage to 16 different experiences that it's possible for a user to see. Now, I've exponentially increased the management complexity of this page. I've gone from one to 16. And you multiply that by the number of pages, the number of possible components or personalized spots on your website. And you can very quickly get into a situation where you've got thousands, if not tens of thousands of different experiences. Which is unmanageable. For a small team going from a marketing exec to implementing a page, implementing personalization, you're just going to tie yourself in knots. So you create what's called a filter bubble. There's a great TED talk on filter bubbles. Filter bubbles effectively, because you've got all these conditions and so many components that are personalizing different experiences, users can see really weird variants of pages. They can see really weird combinations of content that only they got into because of the rules that they hit. Because they lived in Germany and because they were using the site after five o'clock. And because of some other condition. And that will have the opposite effect. It will decrease conversions and engagement. And massively, massively, massively increases management complexity. So if we take a step back to summarize really what personalization is, it's about context. It's about delivering the right message at the right time for the right user. And always to achieve a defined objective. So we need to understand the users, understand what they're after. Deliver them a message that's appropriate to them at that moment in time for their customer journey. Because they'll be after different things at different stages of their journey. And always to increase something. Always to increase this objective here. And we can only do that with planning and strategy. On the point about objective, there's a footnote on the Wikipedia article that adds to the definition of personalization which is about driving specific actions to achieve objectives. E.g. increasing sales, conversion, and conversions on a page. So, let's start planning. We've talked about what personalization is and some of the pitfalls. Let's start from the beginning and think about how we can plan it. And we start with the objective. We need to know what it is we're trying to increase for our organization. And on a website, there can be many different points of conversion. Many different goals. Selling something, maybe somebody commenting on a blog, maybe someone engaging in live chat. What you're seeing, signing up for a newsletter. These are all points of conversion on the website. And different goals will be strategically more important to your clients at different stages. Based on some initiative. Based on some strategic objective set by the organization. And to understand what's important for the organization, because we can't increase all of these at the same time. So to understand what's important for the organization, it's kind of pointless speaking to the marketing exec, the project manager at our clients. We need to go straight from the top and find out what it is that they're trying to do. So we start off by sitting down with the board, the key stakeholders, as high as we can go at our clients. And try to understand what their strategic theme is. Now a theme is an organization's mission statement. It's the kind of thing you see on a business plan or page one of their annual report. It defines them as an organization. It's the kind of things like we want to be the highest quality producer of a widget throughout the globe. Or the lowest value producer of a widget throughout the globe. Or to have the largest market share of a widget. And that defines everything that your organization does. We drill down a level further and we find out what they're doing about it. So what are your strategic objectives to allow you to achieve this theme? What are you as directors sitting down and discussing on a quarterly, annually basis to try and achieve that panacea? So if you're trying to increase your market share, you're probably trying to increase your sales. By what amount? By what percentage do you need to increase your sales to hit that theme? We understand that and we need the figures to work with because we need to be able to report back to them how many sales or how many sign-ups or cross-sales that we're increasing to try and hit that objective. Next we look at any marketing objectives that are in place. A website is predominantly a marketing tool. We need to know what the marketing department is doing about hitting that strategic objective. So if the company's trying to increase sales by 10% how much do we need to increase website sales for to achieve that 10%? So maybe increase it by 20%. We then go down a level further and find out okay, cool, so to hit that marketing objective what are the digital goals that we need to convert on to allow somebody to hit the marketing objective? Now there'll be many of these goals, many marketing objectives, many strategic objectives. So we map them out visually like this. Goals at the bottom, marketing objectives, strategic objectives. We then know that at a given point in time if a company's focusing on fast growth and trying to increase their B2B market share the goal at the bottom here is appropriate for them. There'll probably be tens of these goals but we need to know that if we're trying to increase this goal that's the one that's going to support the wider objectives of the organisation. What we do then is we benchmark. So before we've written a line of code, before we've done a concept design, before we've done any UX or any research work, we start measuring these important conversions however we can. Some of it's going to be measured already. We use Google Analytics, Adobe Analytics, email campaign tracking tools, whatever we can to start measuring. And we put this into some kind of spreadsheet organised by the strategic objective that we're trying to achieve, the goal that we're measuring, and a suitable measurement period. And it's really important to find the measurement period that's right for your client's organisation because they differ. So a charity might expect donations to organically increase around the Christmas period. Or somebody who, I don't know, sells Easter eggs is going to expect Easter egg sales to go up around Easter. So they can be seasonal, they can be campaign based. So choose a measurement period that's suitable and then start tracking your conversions in a spreadsheet. right at the start of the project. We define the objectives and we start measuring before we do anything else. The longer we do this for, the easier it's going to be when we come to personalise to see how successful it's been. Okay, so we've defined the right objective. Now let's think about who the right user is. So we know what we're trying to increase. Who are our users? And we do this through standard kind of UX tools. We create personas. Now, we've probably all seen these. We define the element of our audience. We give them a name. We give them a photograph. It defines their behaviour, their needs, the things that they don't want to do on a website, what device they're using, that kind of thing. There's lots of different ways to create a persona. When we're talking about personalisation, we think the most appropriate method is to use what's called a goal-oriented design approach. And a goal-oriented design approach defines personas based on what they want to achieve. Sometimes you see personas and there's a lot of overlap between what they want to achieve. They're more demographic-based or behaviour-based. Personalisation, we're trying to increase the conversion, so let's segment our users and create personas based on the kind of things that they want to achieve. And this really ties in nicely with some of the stuff that's coming up later. So we create a persona for our members, for our donors, service user, and events. It makes the whole thing a lot easier to manage going forward. The next thing we want to do is a rationalisation exercise. You might not always need to do this. But we've defined our company goals, company objectives, and we've defined our user objectives. Now the two don't always necessarily line up. We've done a lot of work in the airport sectors, and they're increasingly becoming more commercial organisations. So an airport wants to sell holidays, duty-free, and parking, all this kind of stuff. What a user wants to achieve is, yes, book parking, but see the arrivals and departures information, see what time the flight's coming in for the person they've got to pick up. So what the user wants to achieve can differ from what the organisation wants to achieve. We need to align the two. We need to make sure that we're building a website, yes, that's increasing the stuff that the organisation wants to increase, but that it's a usable website for our users as well. So it can just mean trying to be quite confident and strong with our client to make sure that those goals are aligned. So we've defined the objective. We now know who our users are and what it is they're trying to achieve. Now let's think about how we can deliver the right message to them at the right time. By the right time, what I mean is a user goes through different stages when they're converting on their goal. They're not always after the same thing. This conversion process can be quite condensed into a single journey. Or it can take weeks or months. If you're buying a car, the journey you go through as a user, and I'm not talking about a user journey on the website, I'm talking about me trying to buy a car, I first become aware that I need a car. Maybe some kind of life event happens, I have a baby. It triggers me to then start searching for a car, at which point I want to complete certain things. I then go and compare cars, and I decide I want to buy, and then I convert. So we map these out in what's called a decision journey. And these are the different stages a user will go through when trying to complete on that goal. We expand on that and create what's called a task model. And this starts adding some flesh to the bones to the decision journey. It just adds in the kind of things that a user's considering at each of these stages. You can't really see the text, but they're looking to see if there's a cheaper alternative, or they might be comparing two widgets against each other, or comparing supplies against each other. So it just starts to add some flesh onto the decision journey. Then we expand on that, and this is where we start to get into the real meaty stuff that's going to really benefit our personalisation. We create what's called a content relevancy map, or a content matrix. So we create one for each persona, for each person, for each goal that they're going to undertake. You can see at the top, we have the stages of our decision journey. And we start filling in exactly what it is the user is trying to achieve at each stage. So on the discover stage, they might be trying to assess the credibility of the thing they're trying to buy, or what to understand more about the widget. When they're at the compare stage, again, they might be comparing different widgets against each other, or comparing different supplies against each other. And on the conversion stage, they're trying to have a really, really simple conversion process and to buy the widget really quickly. So we map out their intent at each of those stages. We expand on that, and map our client's intent. So what is it our client, the organisation wants to achieve at each of those stages of the user journey? They might want to, at the comparison stage, to keep the user on site, and to prevent them from going off site and comparing with another retailer. They might want to compare and capture some information at one of these stages, because one of these stages might be prone to drop-off. So they might want to capture some information via a form. Typically, our clients are going to want to progress a user along that decision journey, closer to that conversion point. So we map out our user intention and our client's intention. Therefore, we can put forward what content we think will achieve those two goals. What content is appropriate for each of those stages of the decision journey, that will allow the user to achieve what they want to, and our client to achieve what they want to. Then expand on that, and define what call to actions we will use to move the user along that journey. So this is it. At this stage, this becomes the best piece of documentation available. Not only for a personalised project. Up until this point, this benefits any project. It becomes the perfect spec, effectively, for a UX. It tells the designer the personas, what it is they're trying to achieve, what content they should have at each stage, call to actions, all that kind of thing. Each of these blocks of content on here don't necessarily relate to one page on a website. One block could relate to multiple pages, or multiple pages could relate to one block. It's up to our UX guys to go and design some wireframes or some prototypes for us, based on that. And they can then start telling us, they can then start feeding back to us, what slots what components on a website are prime for personalisation. They'll create the prototypes for us based on the content relevancy map and tell us which components are good candidates. So up until this point, as I've said, we've planned out the right message, the right time for the right user. So persona-based relevancy maps define the messaging, the content that's appropriate for them at that stage. So now we can start with our personalisation. Based on that framework. The first thing we do, and again with the relevancy map in mind, is we need to measure. And by measure I mean, we need to understand somehow, we need our website to tell us that a user is Robert, that he's interested in buying a widget, and that he's at that stage of the decision journey. Because we're going to implement some personalisation to benefit that. So we need to look at the tools that we've got at our disposal. Is a goal from Google Analytics going to tell us that he uses Robert and buying a widget and at the discover stage? Is there a query string parameter that's been passed in from a CPC ad? Is there some behaviour they're undertaking on the site that's going to tell us this? Is it the content that they're looking at that we can kind of feedback somehow and tell us that a user is Robert and he's buying a widget? We then look at how we can then nurture that user to convert. And I use the word nurture rather than personalise because personalisation is one of the tools at our disposal to try and move somebody along that journey. So we might use some personalisation on the homepage when somebody's at that discover stage. We might use some A-B testing further down the decision journey. We might send an automated email when somebody's at the comparison phase and they've been off site, their session's ended, they've been off site and they've not been back for two weeks. We might try and send them an automated email to bring them back. We might automate an SMS or a push notification. And when we're thinking about the personalisation, the types of message that we want to drive to them at those different stages, we've really got three types. So we have what's called symmetric messaging. Symmetric messaging is where we try to match the messaging or the content based on an action that a user's taken elsewhere to arrive at our site. So if a user's clicked on a Google AdWords advert and it says donate to my charity, when they hit the home page, they don't want to find it really hard to find the donation box. The top component on the home page should be a donation box with a really clear call to action. So we match the kind of thing that the user's expecting to see based on an action that they've taken off site. It's similar to creating a landing page for a campaign but we're using personalisation to do it. The next type of messaging that we can surface is what's called behavioural retargeting. And this is placing content in front of the user that we think is relevant to them based on their behaviour throughout the rest of the site. So if a user is on a charity website and they've spent most of their time browsing pages about running the London Marathon to support the charity and they then go off into other areas of the site, let's surface some information to them about running the London Marathon because they've spent a lot of time looking at it. That's behavioural retargeting. And the third type of messaging is what we call contextual empathy. This is about understanding the context of the user. They might be on a mobile device rather than a desktop, so we want to show them some condensed text and bullet points rather than reams and reams of text. We want to show them some nicely cropped images. If they're accessing the site out of hours, we want to perhaps tell them the office is closed and to email rather than call us, that kind of thing. We're understanding the time of day of the device they're using, that kind of thing, to make their user experience easier based on the context of their browsing behaviour. So we take all of that, we've defined how we're trying to understand that a user is at one of these stages. We've defined what tools we're going to use to nurture the user, whether we're going to personalise, and we've defined the type of messaging that we're going to display to them. And we feed all that back into the content relevancy map, these measurement and nurture rows. And again, this becomes our specification to refer back to later down the line. So just to recap, we've defined the right message at the right time for the right user, always to achieve a defined objective. Our content relevancy map is proof for that. We've started measuring our conversions, and that's been ongoing now for two or three months since we started the project. And we've defined how we're going to how and where at different stages of the journey we're going to personalise and the type of messaging that we're going to display to them. So we have everything at our disposal now to go and pick a tool for us to do it. And because there isn't at the moment an all singing, all dancing kind of experience management type tool in Umbraco, the world is our oyster when it comes to tools to help us. Some being SaaS based, some being Umbraco plugins, that kind of thing. So we can pick the tool that suits the type of personalisation that we want to do. Now, these tools work in very different ways. I've tried to kind of simply define how they work. So the user visits a page, a piece of logic then runs on the page. There's a rules engine that runs, similar to the rules engine in Outlook, when an email comes in and you say, if it's from Joe Bloggs, and it's a Saturday, put it into this folder. A piece of logic runs, runs a rule, and if a user matches a condition that we've defined using the personalisation tool, so we can create conditions to say if a user's from Germany and they're accessing the site on a Saturday and combine these conditions together, if a user matches that condition, we show them a content variant, else we show them a different content variant, or the default. And most of the rules have that kind of pattern, most of the tools have that kind of pattern in place. So, very simply, we create a rule in our tool based on some kind of conditions that we've defined. We create a content variant that we want to display to that user if they meet that condition. If a user matches the condition, they see the variant. Cool. So, these tools and these conditions typically evaluate two different types of data. We have what's called explicit data on the left. Now, explicit data is the hard and fast facts that we know about our user. It's not up for debate. It's the browser they're using, the device that they're on, the country they're in, the time of day it is, the clickstream, so, a page that they visited. It's not up for debate. We know it to be true. And this bottom one here, CRM, if someone's logged in and we're able to build conditions based on the data we have in our CRM, this bottom one is typically the most powerful to us. If we know what industry our user works in, what they've purchased from us before, the types of purchasing they've done from us before, that's the most powerful type of data that we can use to personalise from. Andy Butland and Theo are doing a talk after this, I think, in the back room about using personalisation groups with CRM and it's really like a really, really strong point, this CRM thing. So if you've got access to CRM data, that typically trumps any other type of data we can use to personalise from. In lieu of that, we have what's called implicit data. So if someone's not logged in, again we can presume to understand our users based on their behaviour on the site. We can presume that they are a persona because they're largely viewing certain types of pages that they've been using, that they're adhering to these content profiles or that it's been a certain amount of time since they last accessed the site so therefore we think they're in this certain stage of the site. We don't know it for sure, it's an implied understanding. Now in lieu of CRM data this can be really, really powerful. Again, if somebody's spent 70% of their time looking at pages to do with running the London Marathon on a charity website that tells us more than if a user has just viewed a single page. It's because a user's viewed a single page doesn't mean they're interested in it. They might have clicked on it by mistake. So we need to know more about the types of content that they're viewing. This is really where the biggest gap is in the Umbraco personalisation package landscape. CRM data is typically the best to personalise from. In lieu of that this implicit information is really, really good and that's where there's a gap. So, look at some of the tools. We've got personalisation groups by Andy Butland. It's an Umbraco personalisation package. It's got a tonne of explicit data points that we can evaluate. Day of the week, the number of site visits, the number of pages that somebody's viewed, the time of day, that kind of thing. We then create content variants based... content variants that we want to display to our users if they meet any of these conditions. It's in beta. Andy said that there are a couple of sites that are using it in production, but it's kind of ready to go. I think yesterday, I didn't see the keynote yesterday, but I think as well actually there's some personalisation stuff coming in V8 as well, but you probably know more about that at the moment than I do. Next up, there's a couple of software as a service based tools. So, Optimizely, which has previously just been really good at A-B testing, now have a personalisation suite, and Marketo. Both dead easy to integrate. A little bit of JavaScript on your website. Then you log in to these tools, build up your rules and conditions, and start creating your content variants. They both have, again, really, really good conditions built around explicit data. Fantastic reports that are really going to help with your measurements to let you know stuff's working later on. You pay for them on a monthly subscription and they're quite expensive. They're thousands of pounds a month. Generally based on the number of visits that you have on a monthly basis. Marketo probably trumps in my opinion Optimizely. It does all the implicit stuff as well. It allows you to categorise your content into different types. Marketo will tell you that we think a user is of this type of persona because they're mostly browsing this sort of content. That's really, really powerful. Marketo even can automate personalisation. It will troll your site, or categorise your content automatically, and then start surfacing automatically content variants to your user based on what it thinks they want to see. It's an incredibly powerful tool. The last tool is Spindoktor. Spindoktor was around a long time ago. The documentation is not great. The demo site is in Umbraco 4. The official line for Spindoktor is that it's officially supported but it's not very quickly developed. Now I've mentioned it because it does the implicit stuff really well. This is again where I think the gap is. This filled the gap. This is more of a plea for Spindoktor or a similar tool to make its way into the Umbraco space. I'm not saying it's not for them to open source it or something because it's incredibly powerful and it could do with being developed and have a roadmap that's continuously updated. Once we've chosen our tool, we build our conditions, create our content variants, etc. based on our content relevancy map. Now we're ready to launch the website. When we launch, we launch with absolutely no personalisation in place whatsoever. Why? In launching a new website with personalisation, we've introduced two variables. We've changed the UX and the design, which is going to have an impact on conversions. We've implemented personalisation. We've got no control mechanism to measure against. We've got this measurement data that we've gained throughout the course of the project. We now want to launch and see the impact that our new design and our new UX has made on the website. Run that for a suitable period of time based on the organisation. And then we have enough data to know if our personalisation efforts are going to be successful. When we're ready to personalise, it's about keeping it really, really simple. We don't go in and create four variants on the homepage on day one. We look at our measures. We look at the data conversions. We choose a goal that we think can be improved by personalisation. We go, okay, this goal probably could do with some help in any way. We look at the relevance. We go back to our relevancy map. We look at our Google Analytics funnels and see where there are drop-offs on our website for achieving a goal. We go, okay, there's a drop-off on this page here, which equates to the comparison stage on our content relevancy map. Our content relevancy map tells us that we're showing a user this type of content and this type of call to action. Therefore, it's these areas that we're going to personalise. We find the suitable spot in that relevancy map that's going to help that user in their conversion journey. We then go back to our prototypes, our wireframes or whatever, and choose the slot at that suitable place within the user journey. We design our assets. We get a designer and a copywriter to mock up the new banner, the new call to action, the new really, really engaging copy that we want to show to this segment of our users. We define the rules. We go into our personalisation tool, define the rules that we want to implement to show this variant of copy. We implement it. We've picked one goal. We've picked a spot in that user journey to personalise. We've designed the assets and now we implement it. Then we wait and we measure. We wait for a suitable period of time. If we've been measuring conversions for three months and we implement personalisation for two days and there's a bit of a drop off, we don't revert there and then. We need to wait a suitable amount of time to see if it's been successful or not. That being said, if we see conversions absolutely drop off the cliff, very very quickly, then fail fast. Get rid of it and revert. Working in these small, iterative cycles allows us to fail fast if it's really not working for us. If we've reverted or we've found out that it's been successful by measuring, we start it again. We pick a single goal, we analyse the relevance and we go through the whole thing through a cycle. It can take quite a while to start personalising all the stuff that we want to, but only doing it in this way, comparing it against the data that we've recorded so far, do we know if it's been successful or not. We start recording the conversions again. We have our strategic objective and the goal over on the left there. We record visually the variant that we've put in place for our users. We have our benchmark, 50 conversions a month and the result was 60 conversions a month. The variance was 10, so we keep it in place and then we pick another goal. If it's failed, we log it too, because we'll forget. Our clients will forget a variant that they've tried and staff will come and go and they might try something that didn't work again in the future. So we log everything, whether it's been successful or not, and the number of conversions that we can expect to receive over there. At the end of it, we can expect to receive between, I don't know, a 1 and a 20% increase in conversions over time, probably around 5-6%. That's a big deal, actually. If your clients are going through this process, they're spending a lot of money anyway, they'll have a high turnover. A 5-6% increase on conversions that they weren't receiving before is pretty big. It's not the 400%, unfortunately. There might be a few cases where that's the case, but typically I think we'll be looking at between a 1 and 20% increase. So in summary, it's all about figuring out what your objective is, right message, right time, right user, aligning your user goals and the organizational goals, measuring, measuring, measuring from day one, all the way through and continuously recording that data. Create that content relevancy map, whether you're personalizing or not, that document becomes the best specification, for want of a better word, for our UX guides, for our personalization strategy going forward. We then go and define our personalizations based on the stuff we've got in the content relevancy map. We integrate it, and then going forward, we're just working in these small, simple, iterative cycles so that we can tell if our work's been successful or not. And that's it. Thanks for listening. If anybody's got any questions, I'd be delighted to try and answer them. You mentioned three months as the waiting time a few times. Why three months? Is that based on a particular sample size or data you guys have analyzed? Is that a hard and fast rule? Could that be changed if a site is high traffic or low traffic? The question was why wait three months when we're measuring? It's an example I used for this. It depends completely on the organization that you're working with, the type of conversions or the type of goals that they have, and whether a conversion is affected by seasonal variance. So three months was just an average amount of time for any organization. Some measurement periods might be a year. You might want to ask your clients to start measuring because their conversions go like this throughout the year based on the seasons. Some could be shorter, some could have very distinct seasons throughout the year. It's a suitable amount of time based on the organization, and your clients will know what's the most appropriate. Thank you. I was just interested. It seemed like you'd had some contact with Spindoktor. Do you know what their plans are in terms of development? I've contacted Spindoktor twice. Spindoktor twice over the past two years. The first time took them two months to get back to me. The second time, it took them about two hours. So the line I got back to them from them was, it's officially supported, so if you have it, then they'll support it. But it's slowly developed. They described it as a bedroom project to me. So they're kind of doing it evenings and weekends when they can. The Spindoktor piece was really about that's where I think the gap is, the content profiling. And if somebody, or if they can pick up Spindoktor, or if they can open source it, or something like that, I think it would be a fantastic tool to have in the Umbraco space. But they didn't suggest that there was a potential for them to open source it or open it up? No. I didn't push them on it either. So that's, that might be the case, but I didn't push them on it. Okay. Thanks everyone. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. More questions? Thank you.