[MUSIC] [Frederik Werner] I'm Frederik. I want to tell you a little bit about making the most out of Adobe Analytics, or at least how I find ways to get the most out of Adobe Analytics. And I'm very happy to make an announcement straight from the start, because despite all the trends, that Adobe started this year, this presentation contains 0% generative AI.

Yep. Yep. Because I'm just trying to create the content that I would like to consume. So that's the content I would like to consume. So the boring part first. This is me. So I like to still write content about Adobe Analytics. For some reason. I mean, there's GenAI for that. I still write myself. I'm dumb and backward apparently. So I'm very happy to tell you a little bit about Adobe Analytics today. And I'm working at a company called Accutics. We also brought some stickers, so feel free to take them from the chairs. Don't make me take them back home. That would be annoying. And yeah, what is Accutics doing? So Accutics, as a company, we make software. Software that works really nicely with Adobe Customer Journey Analytics, Adobe Analytics, all of those tools. So with Accutics, you can do things like create campaign codes, track URLs for your marketing, you don't actually miss any of your marketing data. Once we have everything in order, we make tools that help you to validate if all of your marketing is set up correctly in your platforms. Gives you like a nice head school, and you can set up rules to check if every campaign is CIDs or UTMs and all that. And then it gives your marketing team a nice list of things that might go wrong so you can fix them. And once all your marketing data is in order, we also have a tool to bring the data to your BI tool, like your Google Data Studio, if you like, but also, of course, Adobe Analytics if you want to have all of your offsite metrics, your Google Ads, your Google Search Console, native impressions, keywords, all that stuff in Adobe Analytics. We also do that. So if you want to learn more about Accutics, you can scan the QR code. Also, on the page, that you find behind that QR code, you'll find all the resources of today's session and also the uncensored full slide deck, because Adobe has not just made me remove one spicy slide, but also when I've done the dry run, I ended up with 30 minutes extra content. It's all in there. So if you want, it's all in there. And if you want to talk with me afterwards, the Accutics booth is right up there. We have more stickers and swag and all that. So please come by if you like. So, what do I have for you today? Don't worry, I'm going to bring the QR code up a couple of times. Don't need to scan it. And trust me upfront, you can just have it, when there's an actual topic to scan. So, 'what I have for you today' falls into categories. So something strategic, something long-term to work towards and also something to... if you want to open your laptop right now and put it into action, so something to put into action straight away. Also I'm going through this at high speed. I don't know those tech presentations that drag on for way too long. So if you want to take pictures, if you want to download the slide deck, if you want to watch the recording, please do, because I'm not going to wait. So, agenda-wise. We're going to start with some strategic considerations to get the most out of Adobe Analytics, then take a longer look into Adobe Analytics and some technical features. And then we're going to talk about creating the right environment for our companies and our business users. So let's get started straight away. We have like 62 slides to get through. So, strap in.

Starting with this strategic idea. Why are we doing this? Our companies are doing things. They try different things out. Some of them work, some of them don't work, some surprise them. And of course we want to know what to do, what to keep, what to let go. And we need to have information to decide on what to do. Makes sense? It does. It leads to something that's good. So, problem then is, I don't want to throw any shade at Adobe, but them as a vendor, they don't help with the situation, but we like to glorify data to be the one and the only solution to any type of problem-solving, when in reality, data is just one of many sources of information. Just common sense. Reading the news, googling, following best practices. All of those are things that companies can do instead of looking at data. When COVID hit, which of your dashboards gave you the answer on why your customers are behaving differently suddenly? You can't get everything from a dashboard. Just the same with following industry's best practices. If you're doing media, if you're doing video, if you do what YouTube is doing, you're going to be alright, probably. If you're doing e-commerce, if you do it like Amazon does, you're probably going to be alright. You don't need to have any data or analytics for that. So data needs to be incremental and offset at cost, but also in no small part to company, due to companies like Adobe telling us it's otherwise. Data doesn't actually tell us what to do. We can put our heads or our ears against the laptop as much as we like. It's not going to tell us what to do with our business.

What companies then tried to do in the past is if the data doesn't speak for themselves, clearly, we require some expert data whisperer, someone who knows how to talk to the data, to actually get all the answers out of it. And when that doesn't happen, clearly, we just need to spend more. We need to invest more. We need to build data lakes and whatever. So at some point, the data is going to just speak for itself. In my personal experience, it never does. So maybe it actually won't do that.

What this understanding also leads to is to job descriptions like this one. So I'm not actually trying to mock this specific one from the Towards Data Science blog, but when you're reading through it, I want to point your attention to this part because if we follow this definition of what an analyst does, it is to understand what happens, why it happens, what to do about it, and then give some proactive recommendations. Sounds familiar to anyone? Okay. Some people are nodding. Very good. If this is your understanding of what an analyst does, then I want to ask you, why would you even have a marketing team in your organization? If the analyst knows what happens, why it happens, and what to do about it, and has solutions, well, they should just work in marketing. You don't need your marketing team anymore, they should just do it all. Or use GenAI to actually get rid of everyone in the company, as we've seen during sneaks yesterday. I mean, I don't like that future, but well, it's where everyone is going to. So what this also leads to is to processes like this. So I'm still in this slide from big gains from Adobe, which is one of the typical life cycles of how people try to get answers from an analyst. They put tickets into backlog, they wait for weeks for it to be prioritized and worked on. And this leads, of course, to a lot of delays into getting actual answers from your data. So a more, in my opinion, realistic and also optimistic view on data is that the reality of our business is too complex for anyone who doesn't work in that business to understand it and actually give recommendations. So you need to actually enable your business to answer questions themselves. As a psychologist, which is how I started in this industry, I also know that what we call insights are actually extensions to that mental map that anyone has of reality. We assume how things work, we don't know for sure. And then whenever we learn that it doesn't work that way, or like an exception to the rule, we like to call that an insight. And stakeholders are absolutely able to learn from their own work, and learn in an embedded way. Of course, what it all leads to is it really takes product vision, because the whole reason why I like Adobe Analytics so much is because the people who develop it have this understanding in mind, and it matches my understanding quite well, I have to say. So, yes, when you read through it.

I think everyone who comes in contact with your customer journeys, with your customers, with those experiences, should have everything they need to actually improve those experiences and thereby your business. So what do the best companies do? Some I've worked with at least, they like to create this holistic strategy for working with data in the company. Based on that, they build a small and highly skilled central team to provide capabilities to the business. They clearly separate concerns between stakeholders and analysts, and then build an environment for those stakeholders to be working in and what that means and how different you can approach that challenge, I want to see that on the next slide. So I need to see your hands for a second. Who of you... they had me remove the logos for copyright reasons, but you know what they're going to be. Who of you would be able to work with data in a tool like Microsoft Excel? Show your hands.

Alright, I'm in the right room. That's almost everyone. Very good. Who of you, who just showed your hands, would be able to do the same thing in the open-source tool, RStudio? It's just a couple of you. So, from a company perspective, what we might say is, well, RStudio, it's open source, it's free. We can just use this for everyone. We can just use self-service in this tool. What we would see in that company is nobody would be working with data anymore because it's so hard to get into and nobody knows how to use it. So it's actually a great barrier for entry. I've seen that everyone has shown their hands before. So, one more. One more. Give me. Who would be able to send an email in Outlook? Alright. I think that's everyone. So who would be able to do the same thing that you do in Outlook by opening a command terminal on your computer? I'm going to make you do it on stage if you put your hand up.

So, who would be able to do that by connecting to the email server directly putting on all the commands to actually send an email? As soon as you put that practice in your company, you would see nobody would be sending emails anymore. It's the same with data because us, writers can't be working with data. They don't understand it. They don't speak the languages. But also, if you would be putting this in front of your users, nobody would be able to send an email. It's the exact same thing. That's why I'm always saying why should working with data be any other than sending an email in Outlook. Nobody is born with the capability to send an email in Outlook. We all learn that. People do it better. People do it worse, but we all get the job done and that's what's important to me. So my favorite KPI for any analytics team is this one. User adoption. Monthly active users in your Analytics tool. This is a report that you can create for free in Power BI. So scan the QR code if you want to have the blog post that shows you on how to do that, you can do it in Adobe Analytics. Customer Journey Analytics. Doesn't cost you anything. You can just do it.

So yeah, this is my favorite KPI for any analytics tool or analytics team as well, because it actually shows you your impact to the business. So, summary for this very first part. To get to the right strategy, you should just define that analytics is part of everyone's job. You don't need to be an analyst to be working with data. Get to that understanding and enable your users to do it in the right way. So, by creating a rich environment for those who are incentivized and motivated to actually work with data and improve experiences. And then you also need to trust your stakeholders to be working with data, because what they always see is as soon as someone makes a mistake, they lose their access and they're like, "No, no, no, you can't be working with data anymore." You need to give them room to actually learn that. Makes sense? Very good. So I sped through that first part. I need to drink something.

So, second part. The perfect setup. A little bit more on the technical side, a little bit more hands to arms. We are switching to dark mode, of course, as developers do. So we're going to talk about three talking points in this larger topic. One is implementation frameworks. And we're going to talk a little bit about the actual Adobe Analytics setup. And then some frontend considerations. So when I implement Adobe Analytics, what I always have in mind is the core principles of Adobe Analytics. Also goes for Customer Journey Analytics as well. So it's one interface that does it all. There's no dedicated solution for dashboarding deep dives like reporting. It's all in one. So we need to be able to tell our users, "You don't need to have a separate tool for this. You can do it all in one. You can turn the dashboard into a report into deep dive. You can just do it. It encourages us to always start on a high level, then break things down. Go from a high level further down. Don't do it the other way around and we can just flexible. Use flexible filters, metrics and all of that. So first part. Reducing the complexity and cognitive load through frameworks. So what is a analytics framework? If you implement any analytics tool, you can do it in the old ad-hoc way. So in that world, you have so many dimensions and metrics and events in Adobe Analytics. Because whenever somebody tells you, I want a new thing about this small interaction, you are like, yes, I'm going to dedicate a new event to that, and that way you can run out of events very quickly because you're going to set them up once, you're going to forget about it, they're going to stop working at some point in the future. You can't do any long-term analysis on them. So it's very narrow and a very technical understanding. So the opposite of that which I propose and which helps me to keep those numbers down is through frameworks. So frameworks help you to just use a few dimensions, metrics and events to cover all of the use cases of your business. By doing that, you reduce complexity everywhere, not just in Adobe Analytics, you need fewer rules in Adobe Launch. You can get the library size down significantly and you can reuse stuff. So this is a holistic UX mindset. You actually care about the actual user experience of your Adobe Analytics environment. How do you get to that? So, what I like to do is first we start by just listing the things that matter to the business. Like our entities. We might have something like users, products, we make it, something I could purchase. So we need to list those out and see what matters about those entities. Next thing is then to cluster the interaction. You might be able to click a button, download a document. So what can you actually do with the things that matter to your business? That's like your verbs in the sentence. That's like the actual interactions with your content. Once you have that, you try to abstract it to the point where you can create a set up in Launch. I'm still going to call it that. Sorry, Adobe, I'm not going to call it Text.

And then actually set up all of those capabilities in Adobe Analytics and of course document the whole thing. So, how that looked...? Sorry, I can go back if you want to take a photo of that. I always need to pause a second. So sorry for those watching at home. People are taking photos right now. Also those that are online. I'm going to show the QR code in the end. So how that looks like in an actual implementation then is in the ad-hoc world, for example, you want to track clicks on your links and your buttons, and that might mean you need to have an event for link clicks and button clicks, and you need to have an eVar for your link name and your button name. You might have the media that's on your page, videos and audio. So you need to have your video name, your audio name, your start, pause, reverse, whatever you have. Now, in the framework world, how that would look like? You would cluster all of those into on-page interactions because it's things that people do on your page, and you can just trick them in a very general and abstract manner. You can then say, what's the type of the content? What's the name of the content? What's the interaction? And that way you save so many more things. And for example, if media matters to you, you can just dedicate an interaction just to media interaction and that way, save a lot.

How that looks like on your actual page is you might have things like, Jim Gordon is not in the room, right? So I can actually rant a little bit about Event-Driven Data Layers, but I'm not going to, I used it just for him. So in your current implementation, you have like, your data layer interactions for something like your button clicks. So you have an event that is called button click. And then you're going to track the name of that button that has been clicked. You link click event and the name of the link that has been linked. What you end up with is two rules in Launch, two eVars, and two events. With the framework approach, what you have is a more abstract way of tracking that. So you can see, we're tracking both of them as a content interaction. We say, alright, our content type is Button and Link. The content name is Subscribe and Read more. And both of them are clicks. So with only one rule in Launch one eVar, I'm going to show you how to do that, and one event, you can track the same thing, and you save half of your complexity. That way you can just cut down so much on whatever you're tracking in the frontend.

Some tips for actually making that work. One of my favorite features of Launch is custom debug output. So when you sent those events, when you actually track them in Launch, you can just write directly to the browser console. This thing is just being tracked. If you require your business users and developers to use any browser plugin, in my personal opinion, you're doing it wrong. You should directly tell them what it means, what they've just been doing in the frontend without knowing what eVars have been distributed or whatever. There's no need for that. If you have the option to track IDs and classify them later on to it. I'm going to go over it in a little bit, but just do it. It's also more examples in the in the slide deck. And also one of my favorite things, you can just combine attributes into dimensions and then classify them later through Classification Rules.

So, all of this was surprisingly independent of Adobe Analytics to this point. So let's talk about Adobe Analytics. What do we need to be aware of is all the things that you can do to your data as it's been collected. So many people are not aware of all the features, from processing rules to marketing channels to classifications, and how it affects what you get out in terms of AA. It can uncover all of this. So I'm just going to focus on a few things and just give you a list of all the things you should be aware of. Again, here's the QR code, I'm going to let you take a photo first. Everyone who wants to scan the QR code, can then scan the QR code, because otherwise, it's just going to mess up your photo. But yeah, some of the features that you just need to be aware of. If you're using Attribution IQ you don't just use it for marketing, you use it to attribute, as we've seen in the demos on sneaks yesterday. Use it to attribute your content directions to all of your other success metrics and pages. Use things like custom events for built-in events, custom occurrences, custom page views, custom link clicks so you can use them with Attribution IQ. All of those things. Be aware of them and use them. Because if not, you've already bought the tool, you already have the product, you might as well get the most out of it.

How it then looks like in terms of actually implementing Adobe Analytics? What I like to think of as a mature architecture. My boss just walked in. Hi. It's my actual CEO, Victor.

It's not intimidating at all.

Well, thank you. So, a mature architecture for Adobe Analytics. The no-brainer part, of course, you're going to track events in the frontend and send them to Adobe Analytics and store them there. What does a couple of companies do is just take the data as it is through data feeds, and just drop it in your data lake. My favorite thing, just throw it in the lake, and not even use any of the Adobe frontend features, you're just using it for data collection. It's the most expensive and cumbersome way to do that. There's better ways to do that, if you just do that. What you should instead do is not just use Adobe Analytics in the frontend, but also augment everything you have in Adobe Analytics. For example, with your backend data, you can bring your product classifications and your product returns, your offline orders, all of those important metrics to your business to actually know what's going on and bring context to the data, because otherwise people are just going to wonder, why has revenue suddenly gone up? If they know there's a promotion going on or some returns and offline, they will know why. Of course. Thank you for coming in right now, boss. My company also makes tools to get more value out of Adobe Analytics. So of course, if you manage your classifications in acute standardized for your marketing, you can, of course, get more value out of Adobe Analytics. And also what we do is, they again, had me remove the Google Ads logo and all that, but imagine it being there. So if you want to get your Google ads, your Facebook ads, your Google Search Console data into Adobe Analytics, you can do it manually, you can buy a tool to do that. Choices. Yes. How interesting! So, again, QR code for those who like QR codes. But again, find out more about the sponsor of all the stickers, on this website. So, the next couple of slides are actually the slides that Adobe does not want me to show you, because I submitted those next two slides to Rockstar this year. And Eric Matisoff was like, no, if I show this, nobody will buy CJA anymore. So, Adobe actually does not want me to show them, I'm still going to do that because screw you, Eric.

Sorry. I love you, Eric. If you ever see the recording of this, I'm sorry. So the beauty of classifying everything. As you might be aware, classifications allow you to translate any value we have in Adobe Analytics into any other value. And if we have that power, why shouldn't we just do that with every dimension, like your page name, for example. To actually do that, you could just say I'm checking my page name as normal. Then go into your classification settings. Create a classification like your consolidated page name, and create a classification rule that always pulls the page name that you drag into a classification one-to-one. No conversion. No, any type of processing. And then you can just use your consolidated page name. Why should you do that? And it's just more work for the same thing because as things go wrong, your authoring team puts a typo in the page name or creates a new version of the same page under a different name, we've all been there, what you can just do is you can just go into your classifications, manually upload a new version of the same page name and be like, no, this thing they called New Homepage, it should just be called Homepage. You can use a tool like Accutics Standardize to do that and a nice frontend, of course. But this way you can just retroactively fix your data in Adobe Analytics today. Free of charge. Doesn't cost you anything. You don't need to buy CJA for that. So I think that's a pretty cool use case. I don't know about you. Do you like it? Some do. Okay.

Second thing, because Customer Journey Analytics has always advertised with you, "We have unlimited dimensions." And yeah, you can just track everything. They pull ahead of amplitude and all that. We just show them everything. I don't like that tool. But wait, they just say you can have an unlimited number of dimensions in Customer Journey Analytics. And I'm like, I have that in Adobe Analytics already. So the way to do that is you can use one of your list variables and create a list of just custom attributes of all the interactions that you have. So, how you might do that is you just go into a list variable set up and create classifications on a list variable for the name and the value of any attribute, and you can that way just create a list of keys and values.

What you then do is... to use the laser pointer, again, I was told not to, but I'm going to. So what you can then do is in your frontend, just track a list. So you see how I'm tracking the authors of a page. And also that's text of a page, completely different types of information into the same variable in this delimited format, like you see this key and value format here. Then again, I'm using classification rules to take those values apart, which just gives me a list of keys and values to use in Adobe Analytics. I don't need to set those up upfront. Your developer can then just go and just enhance that. Use it as they like. If your product team creates a new version of the checkout and they want to track the checkout variant and you're like, I don't want to dedicate an eVar to that. You can just throw it in your list of attributes, and that way your developers can just do whatever they like. They can just add attributes, remove them again.

It's entirely flexible. And by doing that, you practically get a list of unlimited dimensions because technically you're limited to 2 million unique values per month. But I'm like, that's pretty much an unlimited list. So yeah, again, going to QR codes. If you would like to read more. Did you like that tip? Okay. Tell that to Eric and then we'll see what happens. I'm never going to be allowed back here.

Also, if you haven't noticed this little icon on the top always shows you the things that work for CJA or Adobe Analytics. So this is actually Adobe Analytics stuff. Doesn't work in CJA right now. The beauty of data sources. Because with data sources, as we do with our company as well, you can just bring summary level data into Adobe Analytics. Your ad clicks, your ad impressions, your Google search keywords. In some session yesterday, I was hearing that you can't get the keyword data into Adobe Analytics anymore. Yes, you can. You can just import it.

So that's a super cool use case for that because you can do it manually. You can do it through automations and APIs. And that way bring all of your relevant data into Adobe Analytics as well, and give your marketing team all the context that they need.

And again, a QR code for you to scan. So, talking about front-end features, because if you're using Adobe Analytics, you have the most powerful frontend in the world at your disposal. So let's talk about Workspace a little bit. And those are just five of my favorite skills for an Analysis Workspace. I always try to focus on them in my trainings. and I'm going to just run through a couple of them. There's more in the presentation. I don't have time.

So just a few things that I always like to focus on during my trainings. Trying to pad out the time a little bit to let people take photos. Thank you. So, most important thing, of course, in Adobe Analytics, in Analytics workplace - drag and drop. I see so many companies would just build-out pre-built reports for their company, at some dropdowns, but never actually show them how to drag and drop anything onto the canvas. I'm like, that's like the most important direction. You can see those drop zones right here, like those tips, everything they can drag and drop. So if you're enabling your users to actually use the interface, focus on that. It's the most important skill in my opinion, and not just drag and drop stuff from the rear into the canvas, but also, once you have it in Workspace at any point, do things like show them to drag something from a Workspace table to a filter, so many people just say, if you want to have any dimension item, go to the left-hand tray, click on the little arrow, search for the value that you like, and I'm like, no. If you see it in the frontend, you can just hold down the Ctrl key on your keyboard, start dragging and dropping it to a filter, and just do it that way. It just takes two seconds. If you see it, you can use it as a filter and once you have it as a filter, It's going to be turned into a segment up here. You can then drag it back to the table. You can just build a segment by just dragging it up and dragging it back down. No need to use the segment builder for that. So much quicker and easier.

So what you can also do is, of course, use it as a filter. So again here I'm dragging it right from the table as a filter up top because I'm like, I want to filter this table down to just those two items. And of course if I want to have a metric for just those two, I just highlight both of them. Right-click. Select Add. I mean, it's not the nicest naming for a metric, I'm going to admit that, but it works. If you just want to go into something, just do some quick aggregations, Why not show them this? I do this all the time in my trainings, and I'm like, once you see it, you can work with it. It's so much value that you can just bring to your users.

One important thing to also note, if you want to implement those custom attributes that I've shown you before, is how to differentiate between segments and sub-hit level filters in tables. Because once you have your list of attributes like I have here, you might be like, alright, I just want to filter for... I don't know if you can see that, I just want to filter for my Page Versions. So you would just want to see the first item in there. And what people try to do usually is they create a segment. But actually all of those have been set on the same page. So as you deploy that segment, you're going to see it's the exact same. You're not going to make any difference because it's filtering for the same hit. And so many people are like, no, we need to have segments that are un sub-hit level so we can only get this one value out of there. And I'm just like, if I just drop the value from the component rate, you have it right now, we can just do that. So this is just a classification. The Page Version of my keys and values that I have. And we can filter that way. You just need to know how to do it in the right way. So there's a difference.

Also, one of my favorite things to do is hacking visualizations for efficiency because for example, people need their conversion rates, their funnels and all of that. So once they build a funnel like this, which is my favorite way for them to see any conversion rates, just drop in the Fallout visualization. You can just add all of your touchpoints. And then as soon as you right-click on any of those items, you can just do amazing things. What I've done here is just break it down. So, on the left-hand side, you can just see a breakdown of the people who went through that funnel. By country, in this case. You can just trend your conversion rate and use it in other places, in Workspace. You would never need to go to the metric builder or the segment builder for that. It's all in the frontend. You're just right-clicking, and changing right things at the right place.

So summary for this part. First, sorry, I told you it's high speed. So, I'm just going to pause for you instead.

So, first rule, of course, make sure that you understand what you can do in Adobe Analytics. Look at those features, see how you can use them for yourself. You pay for them. All of those features don't cost you anything on top of what we're paying. So use them. It's there. Base your implementation on those few and strong principles. So actually remove the cognitive load both for your developers and also for the people who need to be working with the data, and don't work against Adobe's own workflows. Start on a higher level breakdown. From there, don't try to use Analysis Workspace only for BI-like reporting. Get some interactivity and show your users how to use it. And that way just get more value. Doesn't cost you anything.

So third and final part. The right environment for your users. Given that you have all of that in place, like your custom attributes, implementation framework, all of it, how do you actually build this stuff for your users to work with it? So. Well, this was supposed to be animated. There we go. So how an implementation process, how I like to do it looks like is once you have all of those frameworks and all of that documentation in place, you start with your business deciding on, alright, we want to track this new interaction. We want to track this new piece of information and you attribute whatever it might be. And your stakeholders are able to just, based on the templates like in JIRA, for example, and based on some old tickets, to just replicate it, and create an implementation ticket on their own and just tell your developers this is how it should look like. Your developers then, of course, have their own documentation and they already know how to use it from the last time you've shown them, so they can just implement that, like your data layer pushes, whatever you're using. Custom events, if you want to have more control, you can do that. And once all said and done, your stakeholders and developers together can just validate it in frontend. If you put in those custom debug messages, there's no need to use a plugin, there's no need to consult you from the analytics team to actually get it validated. And of course, at the end of the day, they want to analyze it. So, if the analysis should be done, your users can just do it on their own. Do you realize what's missing on this slide? Okay. It's missing the part where you need to go to the analytics team and wait for a ticket to be worked on and got all of the information that is needed because it's not needed anymore. That's what it's trying to do. Your business should not need you as a bottleneck to get any implementation done or get QA done. If your business wants to know, is that button click, they should be able to figure that out on their own. You're going to have them do it the first time. Sure, we are there for them, but in the end, you shouldn't be blocking them.

Of course, if your business changes, if you're doing media today, if you're going to start doing retail tomorrow, then of course you might need to change your framework a little bit. You might need to extend it. So if your strategy changes as a business then you also need to do something. But otherwise, why should you be involved in this process at all? So of course, you need some good documentation for that. And the documentation I'm going to show you is something that I've used two years ago. At this point, at DHL, when I was doing analytics there. So this, for example, I would like to do my developer documentation to actually just describe, alright, this is how the event works, those are the slots you can fill in. You can see my list of attributes right there. They can just edit themselves. They can just put all of those values in there. And it's going to tell them, how do you send data to analytics. And this is where their responsibility ends. Once they set up this event, once it's firing, they're out. They are good. That's all they need to check. For your business users then, this is a conference page, they can just create tables like this. They're just like, all of the values that you've seen on the previous slides, so all of these slots they can fill, and they are just able to define them themselves to say, alright, if you click on this thing it should say this Analytics. It's their responsibility. It's their page. If it's messy, if it doesn't look nice, tell them to do better. But at the end of the day, it's their responsibility. You're not their parent.

So once we have that list of requirements in Confluence, you can then use an implementation ticket like this one. So this is just an example. JIRA ticket. And you can see here's the documentation for the developers. Again, here's just a copy-paste of the table from Confluence, this time with all the technical field names and all that, once they have that as a template, they're going to be able to fill out. Just take the table, just put it in a field that sounds very similar to what you have in your table. And that way they can just create a requirement ticket themselves.

Once they're done, put it all onto the page and they'd want to know, is it working? You can also give them some documentation on how to debug things. So this is the type of debug output that I like to output in Adobe Launch. So when an event is happening, just tell them, hey, you just sent me data, you just sent me this data. Then they compare that to the table that they've just created before. It's not hard. It's not science, to do this. So they can just do it themselves and you can just be like, yep, if you want to know if that button is tracking as you expect and intent, just use this way to see for yourself. Of course, do it with them the first time. Don't just give the documentation to them and be like, yep, here you go. Show them how to open the console fields and all those things. Of course, that's relevant, but yeah, this way you can just give all that power to them.

Of course, you're going to say, alright, they can just put garbage value into my Analytics. And I've already told you, well, that's their responsibility. It's their page. Their product. If they want to mess some things up for themselves, they are free to decide to do that. But of course, people need some documentation. How do my page authors? In AEM, you need to fill those fields. How should my authoring team work with all of those? What are my naming conventions, if you need some? So this is more like business documentation, but also something that you can just help them get started with that, give them a template for that, and then yeah, your authors are going to be self-served. So, of the possible tasks that we do as analysts today, I would say tracking new interactions, add context to existing interactions, that's all your stakeholders and developers. You shouldn't be involved in it. You can help them. You can consult them but they need to be the ones to decide what's going to be tracked in which way. Also, if they want to debug something, their team. Analysis, of course, and reporting as well. What's then left for the analytics team? You're going to ask because yeah, we built a template, we built a setup, but also if there's ever going to be changes, as I said, media companies start doing retail, we need to have an extension to the framework. So that's of course on our table. If we want to have new metadata types available this way that we don't track today, we also need to put something into place for that. Documentation, and also monitoring adoption of documentation. Actually look at who is reading what. Do they actually do that? Super important at this point. Again, user adoption. Main KPI. And, this is Workspace, add templates. So if you know, a content interaction looks like this, create a template in Analysis Workspace that just shows people a curated list of components for that use case. So they learn how to use those things.

Talking about templates, curation, pre-made components. As I said, in Workspace, create templates for those interactions that you know they're going to use. Use curation, just hide all of the components that don't matter to a certain use case. If they're going to analyze page performance, hide all of the content interactions stuff and vice versa, so you don't overwhelm them all the time. Create some segments and calculated metrics for them. So if you know that you have content interactions for buttons, you can just use a button-click calculated metric. You can just create a link-click calculated metric, if you like. So help them with those pre-made components. And I like to at least try and make every interaction with the analytics team a learning experience. So put, in your documentation, in your templates, links to other documentation to help them learn more about how to do more with their analytics.

What do you also need to put in place is some way for them to get support in real-time, and with real-time, I don't mean go to GX, wait then to get in the right queue, wait for two weeks, but actually have Teams channels, Slack channels if you like pain, but also whatever your company is using, for them to just interact with your analytics team and just be like, hey, I'm trying to analyze this. I'm trying to debug that. How do I do that? Be open. Just show them that. One thing I always try to eradicate is any direct one-to-one conversations with the analytics team. Because it's not open, it's not transparent. You don't know what's going on. I always say it's never just relevant for a single person. Everyone can learn from that interaction. So just make it public. Make it open. Also, this icon, that was a recommendation from PowerPoint. I don't know, thanks AI, but it's apparently a good icon for that last point.

And yeah, prioritize those open discussions. And remember it doesn't need to be you who helps them. If marketing has a question in terms of how do I analyze this campaign, other marketing people can answer that as well. In terms of, I've just done it this way last week. It doesn't need to be you. You want to have that open community. It takes a lot of time for that to happen. You need to be patient. It's a lot of upfront work. But at some point they're going to help themselves. And that's what you want to foster.

So summary for that third and last part. Document everything in your company with self-service in mind. They need to be able to do that thing, whatever you're describing, without you being involved, that's the goal. You're going to help them. You're going to be there for them in those real-time support channels, but they should be able to just interact with everything on their own. And you should be that engagement, that community in your company just get to it a more active point and just help them to be more engaged with your users, more engaged with your stakeholders. And yeah, that way you'll get more out of your investment into Adobe Analytics.

Conclusion time. So, if you want to go get the most out of Adobe Analytics, take your practice to the next level, yes, I was instructed to call the whole thing, Those are some takeaways. So start with building that robust holistic strategy. How do you want to work with data and create value from analytics in your company? That's what you need to start with. Then get the technical setup done, get it all documented and live before you actually start migrating your users over. And build that iterative process for you as well, because you're not going to get it right the first time. You need to be able to iterate on your own practice. Learn. Learn from your community, and be better as well. So you create the most value out of Adobe Analytics.