PurePerformance - 006 How to Sell Performance to Marketing with Richard Dominguez

Episode Date: July 1, 2016

Have you ever wondered how to argue with a marketeer about not releasing a new feature or running this campaign? Or in the contrary: how can you show a marketeer that performance engineering and monitoring is as critical to the success of a campaign as the marketing campaign itself? Richard Dominguez, Developer in Operations at PrepSportswear, is enlightening us about how his DevOps team is cooperating with marketing to have a better shared understanding between business and technical goals!

It's time for Pure Performance! Get your stopwatches ready, big number 6 of Pure Performance. I am one of your hosts, Brian Wilson. I've got to think of a better way to say one of your hosts. But I'm Brian Wilson and as always we have with me Andy Grabner. Yeah, the other host. Hey, hi number six already huh wow uh time flies and i think based on what i saw on the internet as they call it uh people are actually listening to what we have to say yes and i hope it's more than just our uh co-workers well i hope it's more than just the uh the bots that I put out there to regularly

download the content, to fake the numbers, because we want to make our marketing happy that they see that we actually have listeners, right? Yes. And speaking of listeners, we do now have an official website, www.dynatrace.com slash pure performance, or that will redirect you to dynatrace.com slash en slash pure performance. But if you just do dynatrace.com slash pure performance, or that will redirect you to Dino Trace.com slash EN slash pure performance. But if you just do Dino Trace.com slash pure performance, that will bring you to our official podcast website where we update all the episodes and for the no prize winners, your names will be listed on that site. So definitely check it out. We're also up on Spreaker. And as always, make sure to contact us through hashtag pureperformance at Dynatrace. Or you can email pureperformance at dynatrace.com.

So any ideas or if you want to be a guest on the show and have something you want to talk about, get off your chest, maybe some personal issues that were related to performance at your company's website that are just stressing you out. We're here to help. Awesome. I think what you forgot to mention is that we are also, I think, on iTunes now. Is that right? Yeah. We had a little trouble getting that working, but yes, you can find our podcast now in iTunes. Perfect. Cool. Hey, I think it's time. This is the But yes, we are. You can find our podcast now in iTunes. So cool. Hey, I think it's time. This is the second time now that we have a guest with us.

Right. The first guest that we have was Mark Tomlinson, the famous Mark Tomlinson on episode number two, I believe. We had him in today. He was appropriately on number two. Yeah. Yeah. I'll let him. He was appropriately on number two. Yeah. Yeah. I'll let him mull that one over. Okay.

And today we have a special guest. And actually, I want to introduce Richard. Richard works. Hi. Richard, before I give you the time to actually introduce yourself to the worldwide audience, I just want to let the people know how we actually got to meet. the time to actually introduce yourself to the worldwide audience. I just want to let the people know how we actually got to meet because we two met, I would say, a year or two ago. Is that about right?

That sounds about right. For Seattle's first meetup. Exactly, the Dynatrace meetup that we did up there. And since then, we got to know each other i got to know what your company that you work for prep sports was actually doing i don't steal a thunder so i think you should do that but i was really excited in already doing two webinars with you and in total i think we also have at least two blogs where we cover some of the stories. I remember a memory leak issue that you found on your application. I also have a blog right in front of me that actually my colleague Klaus wrote on search engine robots, which is a topic that we're also going to discuss later in this episode.

But Richard, first of all, thanks. And for the listeners, if you want to kind of hear and see what Richard already had to say in the last couple of years, you can find most of the content if you open up your most favorite search engine and search for Dynatrace and PrepSports, where you will find the recent webinars that we did on. The last one we did was Six Ways DevOps HelSportsware move from monolith to microservice. And then you will also find the blogs. But now, Richard, who are you? Hey, thank you for having me here. This is an awesome opportunity.

Yeah. So, yeah, I'm Richard Dominguez. I'm the DevOps slash release manager slash insert here person at Prep Sportswear. Basically, I just make sure the site's stable and is released in a stable session. And if there's any performance issues, I'm the first person to kind of notice that and just basically raise alarms. And I'm even like a part-time dev. I'll fix some bugs here and there. It's just all over the place.

But it's a really fun gig. So you check off all trades, huh? Pretty much. I don't think there's really a title for what I do specifically, but then that's not unique to me. A lot of the people there are the same way. There's a database administrator that's like everything else, really. It's pretty fun because when he actually does database administration,

hey, you actually have a job title. What the heck? And Richard, what did you start as before you became this very long title? Yeah. So I worked in, just to give a little bit of a preview, I worked at Microsoft. I was an I worked in just I guess you give a little bit preview. I worked at Microsoft. I was an asset software developer and test worked as a full timer for about a couple of years. And I worked on the Windows phone platform. I was on the services team. So basically all the endpoint services for the phone, not the client but the services so like location awareness and

all that kind of stuff um and then before that i worked on the hypervisor team so actually when hypervisor first came out on server 2008 i was on that initial team um i worked on oh i went blank right now i just worked on a whole bunch of little, just a little project here and there. Was PrepSports where your introduction to DevOps kind of operations? Yes. Okay. Yes, yeah. This was my first DevOps introduction.

I wanted to kind of get out as the pure QA tester, as that kind of role, and get a little bit more front end. And DevOps is definitely that. So, yeah, that's basically where I started kind of like my DevOps-iness in the prep sportswear. And Andy and I have talked about DevOps several times already in continuous delivery.

And some of the challenges that people face. Would you say it was a scary endeavor to go through entering into that world? It, I wouldn't say it was scary because I actually was doing DevOps-y stuff. I like the word DevOps, by the way. When I was in Microsoft without realizing it was considered devops you know i was actually doing a form of uh continuous integration i was already doing uh build verification deployment stuff like that but it wasn't labeled devops Microsoft. It was labeled Ops. There really was no focus on what DevOps was. And when DevOps started, at least the term started to come out of Microsoft,

it was, oh, so these are developers who just happen to write some stuff in operations. And it's like, well, maybe, yes or no. I don't know if there's really one particular statement for devops i think of it and i tell this all the time that it's more of a it's a culture it's a mindset it's more of kind of like the progression of how we push out new changes in a streamlined and confident way out the door and then and then at some point in time some marketing folks came in and said we need a cool term that we can use to excite everybody about something that is total common sense and we call it devops yeah most that's probably

what ended up happening i think so yeah i'm just you know i have to make jokes about it because uh i mean obviously we talk about devops all the time and i remember and i think i've told this story many times now that uh the first time i spoke about devops at a meetup uh people pointed fingers at me and said you don't get the point you are working for a tool vendor you just want to tell tool but you just put devops on everything you create to get the people excited because it's such a hot topic and i had to agree with them and but i i agree with you devops is a lot about just the way we do things the cultural change and um and yeah it's just interesting but you know what's more what's what's really i mean you you are in devops obviously that means you are as you said you are making sure that the stuff you deploy out there actually provides value and you can streamline process.

And what I found so interesting when we actually pitched the idea to you that we do a podcast, you actually said, I want to talk about not dev. I don't want to talk about ops, but I actually want to talk about how to sell performance to the marketing folks. Because in the end, it is the marketing or the business or the sales guys that tell us that we have to build a certain type of software. And they have obviously their vision about what the software should look like. And you work for perhaps Sportswear for an e-commerce site. So obviously you're heavily marketing driven because marketing drives the people to your site. And I thought it was very interesting

that you actually spawn the bridge from DevOps to include obviously business, which is a very critical thing. So I think the term DevOps itself is not including everybody that is really involved. It also doesn't include tests. I mean at least the name doesn't include it. We know it's part of it.

But also the name doesn't include business or marketing. But it should because obviously they are driving what we're doing, right? And so I thought it was very cool that you said, you know what? I'm struggling all the time with performance, and it's mainly because of marketing. They have some idea on what they want, and then they sometimes go off and do stuff that then impacts performance again and stability. And then I'm the one who needs to kind of pull them back and say, you know, this is not going to work. So, Richard, you proposed a topic, and I would like to ask you the question. So what does marketing actually want that might even be a little bit contradictory to what we as performance engineers, as DevOps engineers want to achieve?

What do they want? So the base term that they want is increased conversion rate. That's the single metric that really is what they look at and what that is is as a user comes to the site what's the percentage of these users actually buying something so the idea is okay well there's multiple ways how we can increase that let's increase the amount of users hitting the site. To do that, we might have to make the site more interesting. Let's add social aspects to it. You know, your Google+, your Facebook, your Twitter, all that kind of stuff. Let's add, you know,

additional knickknacks to the site that will entice people to stay on the site as what we think they should want and continue clicking through the site as what we think they should want and continue clicking through the site to hopefully hit that coveted add to cart button. The other part of it is, of course, do they actually click the add to cart button? Do they complete the process? Because that's a sub-process behind that. And that's really at the end of the day what they're looking for how do we increase conversion rate if if we provide a proposal that will increase conversion rate removing a lot of these extra stuff they'll

they'll go through they'll they'll be more than happy to um to agree with that it's just that what do the what do the current metrics of you know of the current situation dictate you know how are the big boys really doing it and how should we mimic that as close as possible that makes sense for us so that's basically what we're doing right now so that means in the beginning um marketing and we we talked about this as a preparation to this talk basically marketing as you said they like to put stuff on there to keep people on the site so they are probably adding as you said the widgets they i know a lot of sites now if you stay on the site for longer than 10 seconds there's like a little pop-up coming up and say hey do you want to chat with us we can help you and, adding a chat widget on it.

I'm sure there's a lot of other stuff, like a lot of tracking code that makes it on the site to actually figure out who are the users, what are they doing, where do they move the mouse to, what is interesting for them, and all this stuff. There's a lot of tracking pixels.

A lot of it all over the place. User analytics are getting really, really big. I don't know if, and just to say, I know you're going to a specific point, but I was curious if, I guess it really wouldn't apply to maybe a prep sportswear so much, but another avenue that people are going now with is user analytics. To make it even more complicated is to start customizing the content for the users right so if you come in and it would be more of a personal experience yes the personalization yeah and so that even gets even more complex but that's just wanted to throw that in there for a second actually uh the customization part is actually a

really big component to our site. Oh, okay. We actually have, we give users the capability to create their own stores. And they can share these stores with everyone else to buy stuff, essentially. So, yeah, we try to do that. And the way it's currently implemented, it's not the best way because it kind of, it creates a lot of additional load. I don't want to get into the knickknacks and the details. But we have to do a lot of, let's just say, SEO filtering to make SEO happy. Because apparently Google doesn't like it if we just implement IDs in the query field.

They want these nice, clean, pretty route maps in the URL. You know, if you're going to a college, going to a school, where is the location, blah, blah, blah. We dynamically render that, actually. And because we have millions of stores, obviously we're not going to have a million controllers in our code it's just a single controller that's just grabbing the stuff from a database but there's there's an interesting disparity between how we're dynamically rendering this content and how google wants it perceived which is really interesting so uh can you tell me because i want to make sure that the audience understands,

when they are in a similar situation like you, working with a marketing, a business team, and if you let them do their own stuff and if you give them full control of what they put on the site, what are the things that could happen? What happened to Prep Sportswear where you in the end had to step in and said guys i know your goal is to make users happy but the way you're doing it right now might not be the most efficient way because with the stuff you do you're actually impacting performance and we know that performance has an impact on user experience and user experience has an impact obviously on user and user experience has an impact, obviously, on user behavior.

People may not consume the content or they don't click through it. So what are the biggest mistakes that you would say that you have seen that people do? Let's see. I'm trying to think of a more recent thing. And while you're thinking was, you know, Andy kind of indicated that marketing had the ability to put things on the site without going through, or at least the way I heard it was marketing out of the ability to put things on the site,

different tracking mechanisms and all that without going through kind of like an official release cycle or gatekeeping. Was that the case? Not for us. I mean, I heard that happens a lot, but didn't know if it was. This is true, and it depends on, I'm assuming on maybe those other sites that are using some sort of CMS that basically allows them to add anything they want automatically um and marketing loves cms because

it's just it's it's it's development without having to understand code right they just add they just add stuff to the side you know click here click there add that you know in five minutes they put this beautiful looking widget on there and don't realize that this particular widget is being called 50,000 different times and different other third parties. And there's a lot of repercussions that can happen just by adding that one widget. There hasn't been anything like that since I've been here. There were instances before I started where they had a lot of third-party software on the site that was slowing the site down considerably, and it wasn't known until much later, the real user impact of it. But as it is right now, I think we're in a pretty good stance.

Right now, we're just trying to clean up the site as much as possible. Our code base right now needs to be improved, and I think that could be true with any company. But as far as third party, as far as what marketing is adding to the site, it hasn't been currently, fortunately, an issue. But the idea of performance, the idea of how it impacts the site is definitely the topic, is definitely the discussion because even though right now there hasn't been an issue, it doesn't mean that they don't understand by adding something there couldn't be an issue. And this is more of a just in case, you know, don't add whatever you want to the site. We need to understand the metrics behind it.

Yeah. And I think what's key, I mean, it's great that you actually, you know, have a great, that you control the whole process and then you make sure that people don't make mistakes. But I guess what happens before you start it and also what I see out there a lot, if you start, if you allow everybody to do things like use the content management system, put stuff on there, you end up adding dependencies to third parties that – I mean they may be necessary. But the thing is the more third parties you add, A, obviously the more complex your overall website gets. But the more dependent you are on these third parties. And that's the critical thing to understand.

If we – as Standard Trace, we always do these site analysis, like web performance analysis when big events come up like the Super Bowl or the Olympics and stuff like that or Black Friday. And what we always find, obviously, most of the sites out there, they use similar third-party components whether it's the social widgets whether it's any recommendation websites all these services so if a big event like break friday comes in or the super bowl and you're sponsoring then you are depending on the same third-party providers that all your competition is depending on as well so if they and then obviously third-party providers they're hit the most and if you pick some third-party providers that are hit pretty hard and they don't scale well it impacts you as well so i think that's just it's a very interesting aspect that you need to think about don't make yourself depending too much on too many different players because you cannot

control them especially in certain high peak load environments. So that's why I think you need to make sure that you're not adding too much on the site. You also want to figure out which other options do you have? Do you really need to go with, let's say, the Google APIs, even though maybe somebody else has a similar offering? So I think this is a big piece. And it's great. So what you are doing is you're actually making sure to educate your marketing team and actually

help them implement all these changes. So you can actually have a handle on how many domains do we depend. So this is, I think, one of the key features or key metrics to look at when you do web performance optimization. How many third-party domains do we have? How many JavaScript files do we download? How many CSS files do we download? How much content gets downloaded after the fact?

What does the caching look like? All of these components. And actually, what is the percentage of content on the whole website that we control versus what is under the control of third party. So I think these are some of the best practices that I have. So look at these metrics. And it's great that you're doing this basically already, not letting them do something. No, absolutely. And Dientris does a really great job of kind of showing you what's third party and what's not.

It's really easy to kind of go through with UEM. And this is what I do. You know, a good portion of my time is just figuring out, okay, let's see how this user's experience. Why are they frustrated? Okay, is it us? Is it their connection? A lot of times it's, oh, okay, you're on a cell phone

in the middle of nowhere in Europe. Yeah, we don't serve that. I'm not too concerned with that, they're like you know in washington in the middle of the city with a bad connections like okay what why are you having a bad connection you shouldn't be so and you can kind of see you know where is it coming from is it because some google commerce.com widget or tracking pixel for some reason took too long. You know, what's the reason behind that? So what's the kind of, you know, you said, you know, how to sell performance to marketing. Are there dashboards that you have and what metrics do you have on there that you show to your business people?

That would be interesting for me and the audience. And adding to that, with that data, how did you start the conversation and sell that idea to marketing? That's important. So what data and how did you approach that topic and really get them to listen? Right. So the dashboard that I have currently in place,

it's currently focused on SEO, as that's 90% we're getting all of our marketing through SEO. We don't have any salespeople. We're a company that strictly goes through these search engines. And the performance metric that I have is, are our Googlebot servers operating healthy? And when they're accessing our particular landing sites, are they experiencing any issues? So I think this is a little bit more detailed, specific metric. And the one metric that shows that a lot is the document object model load time. I think I showed,

I remember showing Anthony this. Andy, sorry. You officially have a new name, Anthony. I know, you're Anthony. Anthony Cracker a new name, Anthony. I know. You're Anthony. Anthony Cracker. Your name is Anthony Cracker? I know.

I think I'm only on my second coffee. That's the problem. Are you sure it's just coffee? Exactly. Well, I might have had those three or two shots, but who's counting? Exactly. So I'm going to move it from anthony to tony now so now you're not tony there you go tony so yeah the i was talking about the the metric probably the

most important metric that i like to show them is the the loading page time uh this is from the end user's perspective and this is a a big one that i know our analytics expert looks at all the time as well as our seo manager um which one is probably more page page load or uh procedure this is uh when the when the dom is complete so the dom load time yeah okay uh it's you know it's it you know this is all from stem from you know uem and when that is when we notice when we see spikes in dom a few days later we'll start getting alerts from google um it's almost almost always what happens and the reason why i say a few days later is because anyone who's experienced with google analytics they're not exactly they don't exactly tell you when the issue happens as it occurs they tell you later at some point in time and they don't give you specifics maybe

we have a more expensive package i i've never really worked with google i know there's a lot of specifics but at least from our current usage, we don't get much granular information. It's just on a daily average, that's all. Are they looking at the DOM load time as well, or are they just looking at the overall page? I honestly don't recall. I think it's the overall. Okay. That'd be interesting if they were looking at DOM load as well.

Yeah. That would be interesting if they were looking at download as well yeah yeah that would be that would be interesting yeah but i mean what one one thing that i want to highlight here i know you're looking at the download time uh and i think bryden mentioned that uh one of the metrics that i know some of our clients are using in uem by the way uem is our user experience monitoring uh where we load a little javascript file on your page and then are able to trace all of the users in every single click. So one of the things we also do is the perceived render time, basically telling you, well, when is the website rendered above the fold so that the users actually perceive it to be completely loaded. And this might be independent to the DOM load time because depending on how you structure your website,

the site may still load if you have a very long page. But what really matters to the end user is the stuff that I see after I enter the URL, which is the stuff that I see on my screen above the fold, is that already there and I can go on. So that's why, Richard, maybe also a hint for you, the perceived render time is an interesting thing to look at as another metric. I just had a Dynatrace user group meeting this week in Raleigh and the client down there that gave a presentation, they also say they look at both. They look at DOM load time and the perceived render time,

which is very interesting for them. And the way their site is actually structured, perceived render time is faster than DOM load times because just the way... Yeah, I would think that. That's interesting. Let me bring up another metric because I might understand it wrong,

but maybe you can correct me, Tony. Document interactive, right? My understanding is that's when you can actually interact with the page right now. That's when the DOM should be fully loaded and all the dependent resources are loaded and the DOM is basically ready for user interaction, exactly. Yeah, it's like the browser fires. Okay, the page is ready. The user cannot interact with me. It depends on the browser. Some of them allow page is ready. The user cannot interact with me.

And it depends on the browser. Some of them allow you to interact. Some of them don't. So I won't say this needs to load before you can. Okay, so there's a browser dependency. Because to me, especially on a site that people might go on all the time, if you're going to prep sportswear and buying a lot of gear, you might start entering a search field

before the page finishes, um, or at least the perceived page finishes, but as long as that search field, but I guess what you're saying is some browser that's a little less reliable because some browsers will let you interact before. Yeah. Yeah. That's, that's the nature. Uh, it's all based on the browser, you know, it's just a standardization model that we all use and we all think of how it should be used, but the browsers can do whatever they want in the end of the day. Well, also, if you look at the definition, for instance, I'm just Googled DOM interactive. Basically, it's a document-ready state, and it has three states.

Loading, the document is still loading. Interactive means the initial HTML was loaded. The sub-resources have been parsed, but they may have not yet been fully loaded. And then there is document complete, which means everything is fully loaded, including all the sub-resources. So that's why, I mean, interactive means, yes, the initial HTML is loaded. The browser knows what else to load, but it may not be there yet. That's why I really like the perceived render time, which is great. And then coming back to the marketing team, I think that what the marketing

really wants is, you know, when do people see what we want to show them? And I think that's why this makes a lot of sense as well. And Richard, you said in the beginning you were looking at this data broken down by SEO landing pages. Is that right? Mm-hmm. Okay. Yes. And because you have different SEO managers, do you have different marketing people that are responsible for different websites

or for different aspects of the site? Well, we have a single SEO manager and he's responsible for everything SEO related for the company. But as far as how it is, are you asking how it's monitored and how we provide that information to our SEO manager? Is that your question?

Yeah, my question is, do you have different stakeholders where you say, well, for you, these are your SEO values and these are your SEO values, but it seems you have one SEO manager. All right. So, yeah, we have – so some of the metrics I am getting is some – what is it so we have a lot of schools that we represent uh colleges schools golf courses military but the three like 90 percent of the time usage is schools and colleges and everything else just kind of falls in beneath those um So I have a metric that gathers the general landing page of the product page. This is the page that first displays what is available to purchase for a particular school.

And this is very important because this is the page that our Googlebot first looks at as it's trying to figure out what does PrepSports have to offer? You know, what has changed? What's going on? And this particular page is very important tomorrow to make sure that, hey, let's make sure this page is as fast as possible. Because if it's not, Google's not's not gonna have they might start retrying they might even start ignoring the page wow so that's actually a very interesting aspect now so what you're just telling me not only is your marketing team interested obviously to deliver the website fast to your end users but what's more important is actually that these landing pages are optimized so when the Google bot comes in, Google actually can load the page fast to give you a better ranking.

Is this correct? Exactly, yes. And we have dedicated hardware just for Google, for bot. Wow, okay. So that means you are – wow, that's interesting. And I know we wanted to talk about bots in the second part of the podcast. So let's save this for now. Okay.

But I think this is pretty cool. And I just want to maybe conclude this topic, kind of which metrics you actually deliver to your marketing folks. So you said you're looking at the loading time. Do you also do conversion rates per landing page? Because that's a feature that the UEM has. Per landing page, we do get the conversion rate. I don't think it's per landing page. I think it's just as per visit, I believe, is how the metric is currently set.

So maybe something for you to implement in the future and for others that are interested in this, because UEM has every single visit but the full user journey, so every single click along the way, we also know what was the landing page of that user. And what I see a lot of people do, they say, I want to have conversion rates per landing page. And the landing page could be obviously an SEO landing page.

So you can immediately see, hey, people that come in to landing page A convert much higher than those with landing page B. Could be greatly used for A-B testing, obviously. There's one aspect. But you can use the same concept if you have any marketing campaigns going on, right? You basically send out email alerts. You are putting – you're buying Google ads and then driving people to your website. And if Banatrace sees the landing page, including all your tracking URLs, and then you can perfectly measure, A, how many people actually reacted to that campaign or to that ad, and how many of them actually converted.

I think that's pretty cool, because then your marketing team can immediately see what the impact is of a campaign, and if you then additionally break it down into performance. So you know that we have, we break down user experience into satisfied, tolerating, frustrating, and the metrics behind it is performance and failure rate. But if you can then break it down into saying we hit this campaign, and overall this campaign was less successful than the other campaign that we did, and the major impact is performance because on campaign, we had 50% more people that had a frustrating experience versus the other campaign. And the reason is bad delivery of that critical page along the conversion funnel.

And the reason here is because we put too many third party components on that page. And the information is all there in Dynatrace actually. I think that's the cool thing. So conversions per landing page, per campaign, but also then broken into the impact of user experience to the conversion rate. I think these are some cool things that you should look into. Yeah. No, that sound very interesting uh we do a form of ab testing but it's it's very cumbersome and all we do is we collect a few button clicks from very specific locations it doesn't at all have like the history of the user like how they actually got there

um and it's very it's incredibly it's not i don't think it's really that conclusive uh this is something that google offers it's one of their ab split packages or something uh but no yeah uem and you can see exactly where the user first got to you can see what they clicked on you can see even what they typed um and uh you can just yeah you can go through the entire history and a lot a lot of this stuff is stuff that i've been really wanting to just immerse myself in for a long time um until we start stabilizing our environment but yeah it's yeah no that this this is great stuff and I would love not only our marketing folks, but even our CS, customer service, to understand this as well.

I think there's another metric that I've been playing around with a lot lately that I'm really digging, and I wanted to share it with you too. Exit page of an abandoned cart. So that's pretty, it's a little bit more of a couple of extra measures in a business transaction, but you're going to be basically

looking for a visit where somebody hits the add to cart but does not convert. And then you can look at what the exit pages are on those to help. Oh, so like where they actually left. Yeah, and that might be

to be able to tell you why they left. That's awesome. So basically what you're doing, because we have every single click, we know which visitor actually added something to the shopping cart. But then if they don't convert, meaning they don't clear the shopping cart by buying, that means these are abandoned carts, and then you can split it by exit page.

That's awesome. That's a cool use case. And you could just do, it's a simple one. There's another conversion measure, and you set the lower threshold to zero. You add that as well as the add to cart in your filter, and then you split it by exit page. Wow.

It's a real fun one. That's a great argument for, because we always get challenged why we capture every single click and every single transaction because people typically buy performance monitoring tools for figuring out performance problems. But I think the whole APM space has moved towards doing exactly these things, user analytics, digital performance management, whatever term you use it. But these are use cases that you can now do with tools like Dynatrace, UEM. And now, Richard, here's a strange question to you now. Why do you still have Google Analytics? Because they've had that for a while. Because we always had it, so we don't want to remove it.

Of course, I understand. But it seems what you are telling me, A, Google Analytics doesn't give you the data fast enough, B, not granular enough, and C, they don't have the full user journey, so you cannot answer some of the questions we just answered here or that we just discussed here. And C and D, the fourth thing is,

you have Dynatrace already on there because you need dynatrace why do you have another tracking software on there that doesn't even give you what dynatrace gives you that's the 35 million dollar question because it's it's google analytics that's what everyone uses right yeah but then maybe you want to bring up some of these questions to your seo managers to your marketing managers say look at this what we can actually do with uem and and one another cool thing and let's see it is um if it's a blog it's already out there We just posted something on GitHub last week. It's the UEM Click Heatmap, and the blog post is also out. It's called

Using Heatmaps to Obtain Actionable Application User Insights. It sounds really interesting. It is actually more interesting than the title makes us assume, but what we are doing, if you go to blog.dynamatrix.com and search for UEM Heatmap, because we have every single click available and the whole user story, the whole user journey, and we know what type of users they are,

like which campaign did they start from. Are they abandoning the website or the card? We can actually visualize their click. They click heat map for a particular page for a particular set of users. So you should definitely check this out, the UEM heat map. It's really cool, the click heat map. I'm going to check that out myself. Yeah.

I remember you showed that during, I think it was our webinar, and you were mentioning the heat map. I thought that was a really cool thing to look at yeah because what i do i mean i'm yeah yeah but because what i do i'm interested in hey um on our community website we have free trial users we have customers and partners and we have employees so i'm actually analyzing on a single page what content to free trial users consume versus partners versus employees? Or are people changing their click behavior if they have a good or a bad user experience? And it's very shocking actually what we found.

We found that more people are clicking on directly to give me support through your support system. If they have a bad experience, then those people that have a good user experience, they typically do self-service and click on community content like our documentation and discussion forums. Those that have a crappy user experience tend to more click on give me support and basically meaning engaging our backend support team,

which is not what we want because we want people to self-service themselves, right? That's what it is. So check it out. I still need to upgrade to 2.3. I mean 6.3. 6.3, yeah, yeah. Yeah, and there's a lot of other cool stuff coming.

But so anything else you want to add to to this uh kind of obviously conversion rate you want to make your seo people happy i really i'm looking forward to talking more about bots all right that sound means it's time for our trivia piece if you get the trivia question correct you will win the no prize knoww-prize which just means your name will go up as the winner on our podcast page at www.dynatrace.com slash pure performance this episode's trivia has to do with toys because everyone loves action figures, right? Dolls. Yeah. Oh, yeah, definitely. Everyone's like, yeah, mm-hmm.

So this one, it shouldn't be too difficult, but a little bit challenging. And it's really going back to a performance action figure. In my experience, there's only been one performance-oriented action figure. by performance, I mean, you know, computer performance testing and all that kind of stuff. Not, uh, you know, Iron Man's performance or whatever else. But I really regret not having taken this one from my boss. He offered it to me. And in my mind, I was just thinking like, how can you give this away? It's a cool action figure, you know, and I didn't take it and that was an idiot. So I really wish I had it, but what is the name of this performance action figure that I'm talking about? Um, goes back to performance testing. And again, just referencing, I started doing performance testing in about 2001. So it goes back quite a way.

And if you can figure out the answer to that, tweet it to hashtag pureperformance at Dynatrace, and you'll be the winner of the no prize. All right. Hey, so I thought it was a pretty cool discussion about selling performance to marketing, making them understand what the performance impacts are. But I think we went a little further than that. So to wrap it up from my side, hopefully I got everything right. Richard, you said what you are doing, you are giving,

first of all, you're making sure that the marketing team does not just blindly add a lot of content to the page because we know it's getting more complex. You're depending on more third-party components, and that actually makes the performance that you control, actually you can only control a small part of the whole page load performance because most of it is coming in through your social widgets and all that stuff. So you basically said, let's make sure that we have a handle on it, and you as a DevOps engineer, you know a lot about performance. And let's work with them to implement stuff that they want to do, but make sure it's not impacted performance, right?

Yes, that is correct. Just basically give a metric to numbers. if I start seeing any kind of degradation. We'll do like some quick load testing and stuff like that to kind of figure out if this is going to really impact performance in any way. Then the other thing that you said, you have some dashboards that obviously contain a lot of metrics and the dashboards that I thought are very interesting

is you are splitting it up by SEO landing pages because you have your SEO manager who is interested in how fast are your pages that you have out there. And this is a segue over to the next topic, which we will cover in a second episode. But it's basically you optimize these pages on the one side for your end users, but also for the bots that are crawling these SEO landing pages because you want to basically show Google how cool and fast you are. And I think that's going to be a very interesting topic. But to wrap this up, you are providing page load time.

And we talked about you using the DOM load time. We also talked about perceived render time as a great metric to use, which is something that Dynatrace UEM gives you. And I came up with some additional ideas on saying you may want to also provide conversion rate not only overall, but by landing page, by marketing page. And then Brian came up with a great idea on also looking at abandon rates on abandoning shopping carts and on which page do people actually drop out because maybe people fill out their shopping cart

and then they all drop out on a page because you have a strange widget on a page and that breaks the page or something like that. So I think this is cool that you're delivering this data. Yeah, this is kind of my wrap-up. I think a lot of cool metrics so performance splitting it up by the different landing pages user experience

using UEM then I raised the question and maybe Richard this is something you need to bring back why do you still use Google Analytics and why don't you use UEM for everything yeah because we because we've always have used it that's pretty much what I get.

There's no, I think the idea is if it's from Google Analytics, well, obviously, well, we want a higher ranking from Google. So GA should be able to tell us that whether we're doing bad or good. It's like, yeah, I guess it could. It just doesn't do a really good job in giving you actionable data it just gives you an idea yeah and i think um you know adding to what tony was saying is that um the the most important part of all this right there's all these metrics all this data you can have capture but the most important part is having that conversation with the marketing and business teams and saying, Hey, we have information that might be interesting to you

and might help you. And it kind of seems like that's what you, you, you broach that topic with those teams to make it available to them and let them know that it's there. And it really kind of sounds like from what you're describing they've kind of become you know one of your best friends in a way because you're providing them with all this critical and essential data that they right didn't have access to before yeah they see uh just to kind of give a quick recap the the main i guess three pieces of data they really look at is the the response page time the dom and the server uh it's actually one chart that lists what our server impact is

and what the end user DOM impact is. And then the other two charts is just general conversion rate by how Dynatrace sees it, which is through meta visits. It's actually different through GA. They're using a session, and there's a different number for that we can talk about that later um and also add to cart how many times has that add to cart been clicked uh that's not a metric that we provide them right so that that brings it back to the beginning where you know you were saying you've always thought of devops as more of a culture and this is that culture you're you're integrating the technical side with the marketing side and

bringing everyone together so i think it's a great accomplishment would you have been able to to get going on over there yeah it's uh it's definitely an ongoing process i think i think changing cultures anywhere is always going to be a little difficult. But I think the mindset, at least in preps, is fairly forward thinking, which is a requisite for when there's ever a changing culture set. And I think we're definitely on that track. All right, anybody have any other final thoughts thoughts i'm just looking forward to the next discussion on the bots i'm really the bots yes they'll be the attack bots which will be uh coming

out right after this episode we're recording the back-to-back everyone because we went so long with this topic that we're gonna split them up into two shows. So we'll see. Maybe we'll have Bender from Futurama listening in on the next one, hopefully. And that's it for me. I'd like to say goodbye, and thanks, everyone, for listening. Again, you can contact us through Twitter at hashtag pureperformance at dynatrace.com, or you can email us at pureperformance at dynatrace.com, or go to our website, dynatrace.com slash pureperformance. And we really appreciate everyone listening.

This is Brian saying goodbye. Everyone else say goodbye. Goodbye, guys. Thanks for having me. Thank you. Thanks, Richard, for being on the show. This is Andy, and goodbye to the audience out there.