E013 - Is Data Privacy Exposing the Uselessness of Digital Data with Timo Dechau

E013 - Is Data Privacy Exposing the Uselessness of Digital Data with Timo Dechau
Life After GDPR Podcast Episode 013 with Timo Dechau

In this episode, I interview Timo Dechau, founder of Data Implementation Agency Deepskydata and all-round data enthusiast that loves to share his insights and challenges with others online.

If you don’t already, you should follow Timo on his LinkedIn, check out his website. Also, if you’re interested you should check out Timo’s free course showcasing how to build a tracking plan.  

In this episode we discuss:

  • The importance, value and requirement of documenting your data implementation
  • How data implementations have changed over the last few years
  • The questions to ask yourself when you’re deciding upon a new tool
  • How to change Marketing Analytics to something that yields results and is privacy friendly
  • How we can make server-side implementations more privacy friendly
  • And much more

Some of the resources mentioned in this podcast:

Make sure you follow the show:

If you want to help us out, please share the link to this episode page with anyone you think might be interested in learning about Digital Marketing in a Post-GDPR world.

Life After GDPR EP013 Transcript

Transcription Disclaimer PLEASE NOTE LEGAL CONDITIONS: Data to Value B.V. owns the copyright in and to all content in and transcripts of the Life after GDPR Podcast, with all rights reserved, as well as the right of publicity.

WHAT YOU’RE WELCOME TO DO: You are welcome to share the below transcript (up to 500 words but not more) in media articles, on your personal website, in a non-commercial article or blog post (e.g., Medi), and/or on a personal social media account for non-commercial purposes, provided that you include attribution to “Life After GDPR” and link back to the https://lifeaftergdpr.eu URL. For the sake of clarity, media outlets with advertising models are permitted to use excerpts from the transcript per the above.

WHAT IS NOT ALLOWED: No one is authorized to copy any portion of the podcast content or use the Life after GDPR Podcast name, image or ness for any commercial purpose or use, including without limitation inclusion in any books, e-books, book summaries or synopses, or on a commercial website or social media site (e.g., Facebook, Twitter, Instagram, etc.) that offers or promotes your or another’s products or services without written explicit consent to do so.

Transcripts are based on our best efforts but will contain typos and errors. Enjoy.


[00:00:00] Rick Dronkers: Hey everyone. Thanks for tuning into the “Life After GDPR” podcast, where we discuss digital marketing in a post GDPR world. In today's episode, I'm joined by Timo Dechau. Timo is the founder of deep sky analytics, a analytics implementation agency which he founded and Timo is a fellow data enthusiast. He's always sharing knowledge about new tools

[00:00:28] he's experimenting with. He recently released a course about how to build your own measurement plan. A free course, and he's always sharing knowledge mostly on LinkedIn, so you should definitely follow him there. So we explore a lot of topics about implementation documentation, about the value of how to structure your events and how to think about that before diving into which tool to choose. How to handle perhaps marketing analytics in a world where Google Analytics is no longer allowed. So it's a really fun podcast. Quick disclaimer, unfortunately, this is needed. I am not a lawyer. I am not a privacy expert. My guest today is not a lawyer, but even if my guest was a lawyer, nothing on this podcast is legal advice or should be taken as legal advice.

[00:01:14] You should always evaluate with counsel before you decide upon any route you want to take. In this podcast, we simply try to explore the topic of digital marketing in a world where data privacy is now a more important topic than it was a couple of years ago. And we try to figure out how this influences our field of work. And what we can do, how we can help. And we try to invite as many guests that can shed a little bit of light on this topic from their point of view. So take that and enjoy the podcast.

[00:01:51] Normally, I ask other people to describe themselves, but I'm gonna describe you as the guy who has tested every analytics tool out there, or at least tries to. Is that accurate description?

[00:02:02] Timo Dechau: I think that's a valid description. It's how to say it's cool. And it's a pain on the other side as well.

[00:02:10] Rick Dronkers: Yeah. You take it upon you.

[00:02:13] Timo Dechau: [Laughs] Yeah, kind of, no, I think I'm, I'm a very natural, curious person. And so of course,  when something is new popping up, I basically have to discipline myself not to immediately leave everything beside and do a full setup with this new thing. So, but yeah, I still try it. So I have a backlog of tools that I want to try out. And then, yeah, try at least to do one test a week.

[00:02:39] Rick Dronkers: Well, that, that's a, that's a nice way to keep on learning , to make sure that you continue.

[00:02:44] Timo Dechau: That's true.

[00:02:45] Rick Dronkers: Besides testing new tools. What else do you do?

[00:02:48] Timo Dechau: I mean, it would be great if you would just make a living out of testing tools and I still have some ideas to do that. let's say my daily job is really help different kind of companies to basically get there tracking or AKA data collection. Right. And so, I mean, there could be different use cases.

[00:03:08] So some come to me and say, okay, we haven't really started out. So we need to get an idea what kind of initial stack we should build up and what kind of events we should collect. These are the easier ones, because you basically start on the green field, the more complicated ones are the ones that come to me say, yeah, actually we track 400 events and we don't really use them and we need to clean up.

[00:03:29] And so how do we clean up? And this of course, definitely doable, but of course,  requires more work because you have to incorporate all the legacy things and how to handle these kind of things. But these are basically the things. So at some point I decided I was working on the full data stack.

[00:03:45] I also didn't data warehouse. And so, but at some point I decided to just focus on where the data's coming from. Cause I spent 70% of my time also on data warehouse to fixing these things. And so I thought, okay, maybe we fix it at the source and not down stream.

[00:04:00] Rick Dronkers: I can really resonate with that. That's also part of what we end up doing, trying to make sure that it's right, when it goes in, right. Garbage in and garbage out. So it's probably worthwhile to invest in the collection phase Making sure that it's properly set up.

[00:04:18] Timo Dechau: And the problem is, really the garbage really has different kind of levels, so it can be technical. So this is often  how to say a, an obvious problem that for example, someone forgets where an event is triggered. This is already also garbage. So when you don't really know where your newsletter subscribed, the event is triggered.

[00:04:35] So just when someone clicks the submit button or when say the subscription is successfully resisted in your system, you like the obvious point. And so, but you don't know it. And so this is already garbage too. And another garbage. What I already mentioned is, if you just track everything or a lot of things, and then you have, I always love these kind of striking setup that are, I dunno, eight or nine years old.

[00:04:58] And they sometimes have six or seven events that are basically the same thing, [laughs] but no one really knows, which is the right thing. So it's really an excavation. So you basically remove one layer of the next layer and just cover new things. Sometimes we also just decide, okay, we build something new next to it and then try to see what we can move over.

[00:05:18] Rick Dronkers: Yeah, that has been my approach lately, as well, because, well, we mainly do Google stack implementations and because of Google Analytics for migration and most cases makes sense to just leave the old container set up service side tag manager, set up a new tag manager client site and build GA4 in that.

[00:05:39] And then once you're done there, migrate all the other tags over from the old container think are necessary. And then you just throw off the old container. It's an easier way than to try to work in that old container where like nine years ago of employees didn't look at the old tax and just implemented new tech on top of each other.

[00:05:58] Timo Dechau: And you will find some old Facebook image tags in there still lingering around for five years already. And still, I think the crazy amount is  when, when you look into tech management setups, I totally understand why it's, why it is a problem. I mean, it's often really  kind of an ownership problem.

[00:06:13] And tech managers move through ownership over time. So there are tech managers who had 10 or 15 different owners. But when you look into it,  you see it from the perspective where you as a business voluntary share your business data with all these different kind of companies. And you're still sending out all your revenue, your orders to advertisement platform you worked with eight years ago, or sometimes even to agencies you worked with eight years ago.

[00:06:40] this perspective it's quite interesting. When you tell people this, then they start to think, maybe we should really do a check in every two months to clean up things. but that's a nice idea to basically do a new set. I think you should maybe also do a post about this.

[00:06:56] I think this is something that people would love to learn because I think a lot of people now struggle how they approach GA4, they just pop it into this overcrowded container. Yeah, it's a good possibility to clean up now.

[00:07:08] Rick Dronkers: If ever right then this is the moment, right? Because you, you can't inherit anything from universal analytics anyway. So you might just start from scratch and clean up and figure out what you actually use. I think tag manager having a lot of owners over the years, I think that's a really good point.

[00:07:26] And I think it also ties in really well with the topic of data privacy. What has happened is the focus was on making it easy for marketers to implement pixels. For years that has been the focus. There were some people who were thinking about privacy, but most people were just thinking about, wow, look at all these possibilities can just implement this pixel, then you can measure your advertising effectiveness.

[00:07:51] This is amazing. So the focus was on that. And then that manager was, now you can implement them even faster without bothering your IT department. Right. so that created this show that we're in right now, where, first of all, we, we went all overboard with all this third party, JavaScript, Facebook, Google, or even way more obscure players that we implemented.

[00:08:17] And  you,  you correctly mentioned, they harvest all data about all your visitors, which is really strange, right. It will be  you own a supermarket you share all the video data from within your supermarket about all your people, people walk around with these third party companies,  all the video data usually share with them willingly. So that's a really weird thing, but yeah, it happened because we got a little bit enthusiastic about the effectiveness of measuring return on investment. Then the ownership, you said, it was, you had people didn't see any harm in it. And then employee number one, implemented some tag employee, number two, implemented some tax.

[00:08:59] Nobody cleans up cuz who likes cleaning up. [Laughs] Right. Nobody does that. And then in nine years later you have a tag manager container with stuff that nobody, nobody understands. And, and nobody really knows why it's firing, how it's firing, what triggering it and what data it contains?

[00:09:16] Timo Dechau: That's true. I would still say it's, let's say in the defense for tag manager, the great thing about it all it's in one place. So, the worst case that you had before was  the scripts were scattered across your website and not even on the same page.

[00:09:30] So sometimes all on the confirmation page, of course. others were there. And so it was really hard to figure out where they are, but you mentioned something, which is, I think really important, which I still hope that it will, gets introduced in let's say tech management systems in a better way.

[00:09:45] It's  all this ownership topic. Cause I think ownership in general data is, of course,  relies a lot on ownership and, and tech manager is super hard. So even, for example, when I work on, on really big setups. So sometimes in, let's say in big organizations they have one tag manager set, which I think is in general a good idea.

[00:10:07] It's definitely better to have than to have 15 containers that kind of doing the same job. But then they even have a very good technical setup. So it's even that. You can even say. No really flaws in there. So naming conventions are fine. They don't really have, let's say tags that are not connected anymore, or some triggers that basically don't have tax.

[00:10:30] And so it's in good shape, but still has a lot of problems because in the end, they are basically how to say 40 or 50 different stakeholders who have some stakes on the tag manager and they have a stake on different levels. So there, of course there's a technical stake. So there's a team that runs the tech manager technically, but then there's, there are different business teams who rely on this kind of data and sometimes they don't rely.

[00:10:53] And of course they miss miss to telling them each other. So, no, actually we don't work with this agency for three years now. And so just it, and so I think this was really interesting to how to handle these kind let's say ownerships.

[00:11:11] Rick Dronkers: Yeah, we have an interesting project going on for one of our larger clients where are doing the migration, as I just said. So we are basically, we took this opportunity to deploy server-side, deploy new client side container, and then build out GA4 in there. And then we realized, okay, this is also a good moment to think about this ownership topic, because we saw what happened in the old container.

[00:11:34] Right. And we don't wanna repeat the mistakes, right. [Laughs] Learn from what we've been doing. and they are a client. So they have Google tag manager zones. That's already something that is interesting from a giving ownership perspective, restricting where people, where and how people can deploy, what types of tag.

[00:11:55] So you could give certain users access to a zone that can only deploy Google Analytics tag or floodlight tag, or maybe not publish, but only as for approval. One part of it. What is really valuable is  the integration with two things or they could be the same tool is a knowledge base it's docented what it is.

[00:12:18] So what we currently have is  a link. We use Coda, coda.io, which is Notion or all these, all these other things. But basically each tag manager tech has a link to the specific thing in Coda and in Coda, there's also a link to the specific tech in tech manager so that the IDs are shared and it works for the API. So you have the documentation is centralized sort of and then one next step that we wanna add is have  the approval queue. Cause we are basically, so, so for this client, we, we publish, they don't publish, they can ask for approval or they can just request to us, Hey, can you help me with this tag?

[00:12:56] Or they can set it up themselves and then say, Hey, can you approve it? And then we go through it. But getting that setup in  something  in our case Asana, which we use,  getting that streamlined and synced with the client Slack channel. So basically there's one place where they can see, okay, we request tags we are gonna deploy them. And then I think the thing that should be added is after a tag is deployed a year later, it should be put into the system again  if this tag is still being used. Right. Or, and if it's still needed or whatever. So, and then each year each tag pops up again, is it still being used?

[00:13:37] Is it still being owned? Who is the owner? Because otherwise you're, yeah. You're never gonna do it. You need some trigger, oh yeah, okay. We added that tag a year ago. Is that agency still working for them? Are they, are they still the owner of the tag or it's  the other agency now the owner of the floodlight tag.

[00:13:53] And, that requires a lot of thinking and mapping out that has nothing to do necessarily with  data implementation itself. It only has to do with, I don't know what to call it. Quality of work or something. Yeah.

[00:14:06] Timo Dechau: No process operations, I guess in the end. No, this is really interesting because I have a quite similar project and so we settled first, not on quota. I mean, we should have used maybe something AirTable or Coda, because it's definitely better to work with, but we started out with Sheets and we did the same thing.

[00:14:22] So we basically implemented an operational and ownership mapping layer on top of GTM also  communicating via API. So thinking via API different kind of information. And, and we, for example, we even took what we were not sure if this really works out, but even took the Hardaway, we even put expiration dates on it.

[00:14:41] So it's even not it's up for notice. So basically at some point, at least this is how we planned this attack basically enters a let's say a process where it fades out automatically. So it's, at some point it gets an expiration date when it gets set on pause. After this expiration date automatically.

[00:14:59] And then basically the owner has a possibility before, of course he, or she gets emails and notifications. So here it's a domain. So basically, yeah, your tech is about to expire. So either you basically and you cannot set it on auto renew, so you have to proactively say,

[00:15:17] And then it goes on pause and some place, of course, it goes really out and out of the system.

[00:15:22] Rick Dronkers: That's a great idea. Because, with our approach, what will happen is people will be, yeah, sure, it's good. Just click done on the task. Right. And not check it actually. [Laughs] And by pausing it, you have to take action. You have to

[00:15:37] Timo Dechau: Which is, I think totally normal because in the end, you just disrupt their day with that. So, okay. They are thinking about marketing campaigns and then someone comes around and say, yeah, actually, do you need this tag? And so it was usually requires at least some thinking.

[00:15:50] Timo Dechau: Sometimes it's straightforward. So when, okay, this is tag from. Say an old agency then. Yeah, sure. You just take it out. But sometimes it's complicated to decide.

[00:16:00] Rick Dronkers: So what I've seen for instance, with affiliate networks, right?  there's gazillion of them and then they seem to die and start new ones every day. Right. So, so those pixels, it's definitely worth to evaluate each yeah.

[00:16:15] Timo Dechau: I would just say, so if you really want to reduce or increase the quality of your tag manager and reduce the workload of the team that is handling the tag manager, just forbid to use affiliate programs. Cause I would say looking back on  extensive tech monitor retainers, 70, 80% affiliate.

[00:16:33] Everything else, I don't know. Google ads, analytics always pretty straightforward, not so many changes. It's affiliate us constantly changing. It's s ridiculous.

[00:16:45] Rick Dronkers: Yeah, maybe that changes when they all move to a server-side API, but , not too hopeful for that. [Laughs]

[00:16:52] Timo Dechau: Yeah, let's see. I mean, in the end, affiliates always struggles that they don't get. I mean, they are really, depending on getting this final signal. And so I think this drives them to always come up with new ideas to get this signal so they can make money. So the good thing for Google, I mean, Google already had, and Google and Facebook already have the money once people land on your website.

[00:17:17] So of course they would like  to learn a lot about conversions and so on. So, but they don't really have a lot of stakes in that. Affiliate has a lot of stakes because basically that business model depends on that.

[00:17:27] Rick Dronkers: Yeah, that's a great point. And I haven't thought about it like that. Yeah. So Google needs the conversion data to prove to you that their advertisement was useful. But in the end, if you would do incrementality tests without any tracking, you would also figure out, Hey, Google is it's redirecting some valuable traffic.

[00:17:46] Whereas with the affiliates. Yeah. You're, it's after the fact thing. They first send you the traffic, then they convert and then they have to trust you to tell them that that user converted right.

[00:17:56] Timo Dechau: I think this is, you just mentioned a very interesting topic and I think, I don't know. Do maybe you see this as well? It's with all, , let's say GDPR requirements. So of course  you have consent, so you ask for consent. So of course the data volume drops that you can basically get.

[00:18:13] What I see now in a lot of projects is  that. Of course I always, before were  tensions between classic online marketing agencies on let's say analytics departments or analytics teams. And these tensions definitely grew bigger because now,  you often get from, let's say agencies that they say, okay, we basically cannot optimize anymore because, we are bound so much to these algorithm of these platforms and these platforms need basically, I mean, their food are conversions.

[00:18:41] So let's say indicators that something is successful. Cause all the algorithms works on that. And so they say  when we reduce this, so we definitely will get increased costs, which I think is definitely possible. So I never really spent a lot of time to dive into. But it's interesting to see, because you just mentioned, imagine you would Google ads when you just switch off tracking, which  in the old days when you did Google or when it was still at AdWords, the tracking was not so big and sophisticated also,  there was sometimes really lag and time, so it didn't really have the systems that you had there.

[00:19:15] in that time, often they, I know that at least some companies really drove these kind of tests where they aggressively in specific kind of targetings in areas really, switch it off, measure the volume changes on all, ket's say metrics that are on operational level, I don't know, orders and so on.

[00:19:37] And by that making basically models to get an idea, what specific kind of channels are bringing in, which you still have to do with, let's say these with display influencers and, and TV and on which you of course cannot really track. And so what are your experience to work with these, how to say this conflict between marketing agencies to say, okay, when we cannot track, we cannot work and coming up with different ways.

[00:20:04] Rick Dronkers: I've seen this, especially with companies that move towards a more mixed media approach. Right? So, once they start doing out of home, not digital out of home, but just out of home advertising, TV advertising, then usually one of the more traditional attribution agencies, from the traditional advertising world comes in to help right.

[00:20:31] To, basically help them with, and they've been doing. these incrementality tests for years, for decades, for however, however long. And I think in digital, we are, we are spoiled with deterministic measurement and we are now due to privacy. Both privacy legislation and privacy technology is enforcing it and legislation is prescribing it.

[00:20:57] Are going more probabilistic measurement. And I think these incrementality tests are really valuable, but what I've seen in the past they do require quite some volume to give you a yeah, a good feel of what's going on. So I know eBay has done alot. And, that's some time ago, but I've, I used to work for the Dutch eBay when I worked at an agency before and there, also did a lot of incrementality testing by simply switching off, for instance, advertising via Google ads for a certain category.

[00:21:33] Right? So they had these pillars of categories on the, on the platform and then switching it off and on for a certain category for a specific amount of time. And that, yeah, that was really interesting, cuz it, you could clearly see that there is an impact, but the impact is never as big as what Google would  to believe.

[00:21:48] Right. Or what, what other advertising source would like you to believe and that's to be expected because if you only use Google ads and that's the only way people can discover your website then sure. Once you start using multiple sources and you start getting word of mouth and you start getting customers that refer each other and talk about your brand and you, yeah.

[00:22:07] Then of course, it becomes more complicated than the preciseness of our digital is actually fooling us it looks so precise. So you're inclined to believe it, but actually it is hiding the complexity of what is really going on.

[00:22:27] Timo Dechau: I think this is, this is kind of, let's say the side effects of all this GDPR things. I quite like. So because it, it basically takes away this mask of this is a precision engine, which it never was. I mean, maybe  we all who are in this industry, we can all remember this time when  attribution models was super hot and  everyone was basically calling, yeah, we want our attribution model.

[00:22:52] So we are definitely ready for that. and in that time it was already super complicated and also  complicated to showcase that. I mean, it's a model. Let's say for a reason, because it's a model. I think this was, this was strange times as well, where people were really assing.

[00:23:08] In the end, we will just have a machine that is basically, so we build a great product. We just plug in the marketing black box and then the marketing black box is just selling it out. And so we don't have to do anything. So  really  the developers paradise. So you just develop a great product stuck in the marketing machine and then everything works by itself.

[00:23:27] But yeah, I think it's quite interesting because, I think now it brings back all the, let's say at some point really ignored things of marketing, I don't know, positioning, branding and so on and all the different kind of things. Brings it back, which I guess they were never really gone. And they were always important, but , I mean, there could be easily ignored, especially when you really driving this hot selling eCommerce models.

[00:23:50] Where of course when you were really good, at least in the first place where AdWords was not so crowded. You could build amazing models. So really if you understand the machine and go for it, and you, you had some money to basically pump into the machine. there, there were companies just, just living out of that.

[00:24:08] Rick Dronkers: I would go a step further. I would say it digital marketing. So I think indeed, there's a lot of companies that got started on, for instance, Google ads, but also later on, there's a wave that got started on Facebook ads and there will now be a group of companies that get started on TikTok ads.

[00:24:25] And then after that, there will be mix smoke ads, right? Whatever. But I think what are showing is that if you are a first mover to an advertising platform there, where the gap between the amount of eyeballs that are on it and what you have to pay for getting in front of those eyeballs is still really wide.

[00:24:45] Then you are gonna profit off of that, right? Because you are basically, you are chasing attention and attention has shifted from platform A to platform B usually because platform A was over monetizing and people left because they were, I'm only looking at ads.  if you see, if you currently see on Instagram, about Instagram, I don't even see people from my own feed anymore.

[00:25:10] That's what people say. switch to TikTok because they like it better. I can a hundred percent guarantee that TikTok is gonna happen because they're gonna get that advertising money in, then they're gonna up the advertisements and then a new platform comes out and people move there.

[00:25:26] Right? So this, the, and I think the first mover will always benefit from that. And think that has maybe overemphasize the value of digital marketing and has also allowed for all this complexity to arise around it. Cause, cause I think a lot of the things that we have built around these models Give us, in some cases, a fake sense of control. Like with analytics, I think for some companies we check everything and we put it in a dashboard for them, and then they look at the dashboard, but they don't take any action on the data.

[00:26:03] Right. It's just to verify, it's a verification or gives them maybe some trust or some certainty, but it's not actually helping them take action and optimize for a lot of companies. It's more showing what is happening rather than informing decision making. And I think this will expose a lot of that because once that data starts to crumble, but then if the business stays the same and the data crbumbles, then basically you're gonna figure out, yeah, we didn't even use that data.

[00:26:35] Timo Dechau: This is maybe another side effect of, of GDPR that I guess now may at least, I think maybe some companies will start to ask themselves the question, what would happen if we just remove everything? So if we remove the dashboards, if we remove analytics, does it change something in our company?

[00:26:51] And I think. The honest ones might arrive at the point where they say, actually not. As you said, so often you have these dashboards. I would not say call it a trick, but of course, sometimes it happens, things break in these kind of things. And where I really get scared is, when something is breaking and it breaks four weeks ago.

[00:27:14] Because I don't have monitoring place, which of course then is my fault, but no one else recognize this definitely is a big point of that. Something's going wrong.

[00:27:25] Rick Dronkers: I think it's one of the side effects of Google being free. Right? So when something is free, it's really hard to say let's not do it. And then it creates fake sense of safety because having this data somehow gives managers the feeling seeing that the amount of visitors to the website is at a certain level, gives them some form of safety, but it's not actually being used to optimize.

[00:27:52] So I, I would argue indeed, if you, if you remove the time staring at numbers and invest that time into doing something valuable for your customers, probably net of for a lot of businesses at the end of the year, they're gonna be better off than staring at the numbers, which they don't use to inform any decisions.

[00:28:10] Timo Dechau: And this, I think the light version of this infinity. I think the heavy version is basically  people investing millions in data stack and doing the same thing. I mean, this is quite interesting. So I started out with analytics in 2006.

[00:28:26] So basically when the first real Google analytics version was coming out, I think it took four or five years to basically fight, to get some acceptance for analytics, because it wasn't  that everyone was erasing. It was more that people were highly skeptical about these things.

[00:28:41] I know that I had to prove a lot of things before people were really  taking into account and I was always envisioning this kind of future where everyone's using it. and yeah, so then I saw  how everything goes to the other end, where it became totally hip and kind of a requirement to use analytics.

[00:28:58] And I think this is also maybe some of the reasons that we have today is that people think you have to do analytics and so you do it. So of course you, when you want to do analytics, so you Google or you, I think it's not even that you Google, I mean, of course,  then you might come across Google analytics.

[00:29:15] I think it's almost that. Of course, you ask colleagues or you ask friends in other companies, what are you doing? and then it's basically spreading and…

[00:29:23] Rick Dronkers: I think every CMS has in the settings field, there's fill out your Google analytics ID. Right?  that already, how many times I've had a client, okay, can you, we want to implement Google analytics. So I'm like, okay, let's walk through the steps. They're like no, we just need the ID.

[00:29:40]  I'm like, okay, wait a second. What are you gonna do? And then they're like, yeah, but we need to fill it out here and then it's done. Right. but that, yeah, that also, of course from the web developer point of view, it also became a default thing.  You deliver a website that has analytics, right.

[00:29:57] I don't think any web developer would disagree. At least clients want it. But I think in most cases now, actually the web development agencies themselves will offer some basic form of, okay. We will have Google analytics page shoes on there.

[00:30:11] Timo Dechau: Yeah, but I think this is now I think this is  what you already said. So it's kind of the time now. When we were talking about tag manager. So, I mean, this is the interesting time as well. So, it's a good time to revisit everything. I mean, everyone was jumping on this GA train for years.

[00:30:26] And I did the same. I mean, I started out with Google analytics because it was basically, it was the best way to start out with analytics and we still have to say, and I still say it Google analytics is a good product. It's definitely. So it's not that.

[00:30:40] Rick Dronkers: It's hard to be free.

[00:30:41] Timo Dechau: Yeah. It doesn't have to be free.

[00:30:43] So it would be even there when people had to play it, because it did a lot, especially when, when you follow up the whole journey. So I started to use GA classic, I think 2006. So it was really the first thing. It was the time when you couldn't create segments or segments, for example, was not introduced yet.

[00:30:59] So when you wanted to segment your traffic, you had to create a new view and the view took two days to populate. And so when you did the filtering wrong, you had to wait another two days. So sometimes it took three weeks to come up with a new segmentation And to see how Google was really  evolving over this time and  to speed up the whole machine.

[00:31:17] That was impressive. So when you work with the early systems and I was working in some projects still with, let's say some competitors, which I will not name out. And when you work with them, it was a pain because these systems were super slow, super clunky, and it still is. And so, it's natural that you really understand that.

[00:31:35] why Google of course, got this cut of attention. Not just because it was free, not just because Google is a great lot to tell people. Yeah. If you want to work use AdWords should also use analytics. I think they they're really  different reason, but the interesting thing is now is really, I guess it's now really a first opportunity to revisit this kind of strategy.

[00:31:56] And I think this is the interesting time to really see actually what is our data strategy and how does it look ? And. I don't say it's a super simple talk. So it's highly complicated. I mean, it's deciding on your software stack. I mean, if you build an application, it's the same thing.

[00:32:11] So you have to decide at some point, how would you do this? Do we go with buying something and put this as a core? Do we build everything ourself? Do we do microservices? I don't know. So it's the same with data stacks. It's the same. So you can make yeah, really, really different decisions based on what you really need.

[00:32:32] Rick Dronkers: Yeah, a hundred percent agreed. I think, privacy regulation is maybe the trigger that we all needed, but I think we could that even if there was nothing with regards to privacy regulation still. There is now way more to choose from and better solutions for certain use cases right. Than to just go blankly with Google analytics.

[00:32:55] Timo Dechau: Maybe I think, I think privacy definitely helped to speed the whole thing up, but I think the tendency was, was already there because. What I often see is, in  the different kind of projects that I do is, the gaps sometimes are really huge when you look at, at the different kind of companies, how they embrace basically data and how they would put data into their product development. This is sometimes really scary when you see how big the differences are. of course everyone is pointing out then on companies Netflix and Uber and Airbnb and say, yeah, they have 200 people. And so on. Yes, they have these kind of people, but they have a totally different approach and philosophy where it comes to data.

[00:33:38] So data is an essential part. It's technology, it's their software. It's thought in different ways. So because of that, Go different routes and go different kind of approaches because they know when they change, improve something on their data, set up at some point, I mean maybe increase data quality by 5%.

[00:33:58] It in the end has an immediate effect on their business. I think when you come to the stage where, where data is contributing, let's say a part of your revenue that you can attribute, because, for example, you, you have a specific recommendation engine and in your product and this needs to feed with data.

[00:34:17] Then data plays a totally different role and then for example, then to betting on, on introducing an external system is definitely a business risk. Because in the end, you basically let's assess you basically make 10 millions out of this recommendation engine because this from the data and then the place of be on a new system or a different system that you don't even own.

[00:34:41] And that you don't, it's a black box and you couldn't really look into it. And they say, yeah, I put my 10 millions on this kind of thing. Interesting choice. No one would basically go through this. And I think this is where the minds are starting to shift and the decisions are starting to shift. It's quite, quite interesting to see these kind of companies that are in this stage.

[00:34:59] Rick Dronkers: If data is a competitive advantage to your company, right? So if it's a structural element of what you provide, then the decision should by definition already top down, be okay, we want to own this, right. We, we want to be on top of this. What I also see is I work a lot on the marketing side of things.

[00:35:19] I think you also do. I think, Google analytics has for obvious reasons, always been a lot on the marketing side because it came along with Google ads. Right? So and I see it, a lot of companies now I see the realization of, okay, data is a, is, or should be a competitive advantage us in our business model.

[00:35:38] And then they have, they start building a data warehouse based product, data recommendation engine. So let's say it's a travel website. Right. And they wanna serve you the best search result based on whatever kind of inputs. Then you start to see this segregation where, okay, we now have.

[00:35:53] Internal data, product data, and then we have Google analytics or alternative, right? Most times still Google analytics, which is only being used by marketers, just for, just for conversion. attribution. What I saw a couple years ago was a lot of companies trying to stuff Google analytics with a lot of data, also this product data.

[00:36:16] And now I'm seeing separation where, okay, let's just do the conversion. The landing page hit and the conversion in Google analytics, that's enough for marketeers. We're not gonna make it more complex than that. Use it for attribution, maybe, you know, add to cart and then conversion. So you have the beginning of funnel, endo funnel and you have landing page hit.

[00:36:34] And then all the other complex stuff, we have our data engineers and they don't even talk to the marketing people. Right. They don't want to, they're just building their own stuff for the recommendation engine or whatever their business And I feel that that also. that compliments what you just said.

[00:36:49] A lot of companies are realizing how important data is for them and they are, and they took steps to yeah. To, to build that. And I think for marketing, for marketing now, the stick is privacy regulation where we're realizing, okay, if data is also really important to marketing. So if you really, if you spend a lot, that's usually the first thing to look at.

[00:37:13] If your ad spend is really high and you spend across multiple channels, then probably then your conversion data is important enough for you to consider an alternative to Google analytics, right? Because the chances of you being able to use Google analytics in the future will become less and less if you operate within the European Union.

[00:37:32] Yeah, you will have to evaluate, okay, let's say you spend 20 million a month. Right. And you, and you do that across, I don't know, 10, 15 channels. If you don't have conversion tracking anymore, will you still be able to spend efficiently? If yes, then good. [Laughs] You don't need analytics.

[00:37:50] Right. So you're just keep on spending. But otherwise you have to build some solution. I wanna explore that with you. Let's say we have this client, right. This client,it's a client in GDPR region case for a second that well, you're in Denmark, right?

[00:38:06] The Danish, just said Google analytics, also, allowed, at least not in the implementation.

[00:38:16] Timo Dechau: The interesting thing, what they did is they even pointed out that it was a European decision, which no one else said before. So it was really, they're already spoiled what will happen in other countries as well.

[00:38:27] Rick Dronkers: So let's take that as a given, right? Let's say Google analytics, they have this, okay. You can use it in a proxy form, but basically then you're using, you're losing all the value. So let's just assume Google analytics out the window. Right. We are company spending 20 million a month across 10 online sources. What are we gonna do? What are all the tools, where are we gonna switch to?

[00:38:49] Timo Dechau: I think there are different ways. So the first important part you already said, data stacks are splitting up. Even Google analytics and most of the setup that I do, Google analytics is just a satellite anymore. It's not  a core thing. So we basically move it out. We use different things.

[00:39:03] There's a core thing where we basically collect all the data and then it might be passed on to Google analytics as well. For, as you said, for landing patients, so on. what kind of tool we are using? I think this depends on the setup of the company. So first of all, I think one criteria, for example, which I always ask is how do conversions work?

[00:39:21] Are you built in a way that you convert wherever you want to convert? So we can talk about this as well. It's, if you, for example, can make sure that a conversion usually happens pretty quickly. So let's say in the easy case you sell shampoo. So of course, people come on your website, buy a shampoo, go away.

[00:39:38] So, usually conversion timeframe within a day. So definitely quick selling and so on, but it could be also, if you are a, let's say B2B software service startup, then of course,  you could sell, try to sell people into your product, but this takes time. So usually in this kind of case, what the people try to do is, they want to get the people as quickly as possible in their CRM systems.

[00:40:02] And so they work with sleep magnets and so on. So  there's the same scenario. So you usually get people on, let's say a landing page and they can download the white paper. So usually it's, you immediately convert them in, in a way. And so if you have the scenario that you convert more quickly, then of course you could go with the, let's say the interesting breed of what, how they call themselves privacy first analytics provider, or cookieless tracking provider, so that they don't set a cookie where at the moment we asse this is totally legal.

[00:40:32] Which when at some point it might land in front of a court and maybe a court makes a different decision. But what these kind of tools are doing is they reduced identify cable. I don't know. So the data that you use to identify a user, they reduce it to this window of maximum 24 hours, because what they do is they take the IP, they shorten the IP, they take maybe sometimes the user agent as well.

[00:40:58] And then they basically create a daily salt based on the date. And then they create a hash and this hash basically rotates because the date is included. It rotates after 20 hours. So by that they reduce basically the possibility to track users over a long span of time.

[00:41:16] Rick Dronkers: What is also interesting about it is that the identifier that they place, let's say that another source, so Google or Facebook would read out that identifier, then also they would only be able to use that identifier for 24 hours. Right? That's the thought behind it.

[00:41:33] Timo Dechau: Yeah, and, and they don't even persist it. So they usually just use it in their kind of system. So even  other systems could not basically grab these kind of ideas and match it up with their ideas and then build these. I mean, these were all the fun things that people were doing in the past. So it's definitely a difference to how all the other analytics tools are working.

[00:41:52] And so, and this, for example, could be a way that you use this kind of data to at least get all your marketing information. What you could also do is then even if you want to go a step further and you want to prove a data privacy agency that you really take care of these kind of problems.

[00:42:11] So for example, you could make sure that you don't even care about this unique identifier. So you basically sent the data in some kind of analytics tool just by identifying them from the marketing campaign or marketing source. So the most granular level that you are, want to track marketing on. So on campaign level at level or whatever you want to do.

[00:42:30] And so you use this as an identifier. So I built, for example, one system was a company where they said, okay, we just want to have a marketing reporting. So we don't really care about what users are doing. So we just want to know our campaigns performing. And so we were just sending data into the system by using the campaign as an identifier.

[00:42:46] So a campaign could be something that they provide with a parameter, or it could be  a shortened referrer, or it could be just if we don't have information that has just, and so this was the things that we were sending into the analytics systems, because in. They just wanted to have a campaign reporting.

[00:43:03] And so we built them campaign reporting. So we were reducing the things that we were tracking beforehand, but still they had full campaign reporting across the full funnel. I would approach it first from this point.

[00:43:15] Rick Dronkers: I think that's also what WebKit is promoting with their link tracking solution, which is still a little bit rudimentary, I think. But basically they say, okay, we only wanna, instead of you sending all these unique click identifiers that could in theory identify a single user.

[00:43:34] Right. And which is of course true a Google click ID or Facebook click ID, Google, and Facebook can tie that back to a specific single user on their end. Right. So that they know that unique click ID is unique to that person. And then WebKit is saying, yeah, we will offer this API that will only give you, I think it's  a source medium campaign kind of combination or something.

[00:43:56] Right. I think for marketing organizations, if you were really grill them on it. Right. And really make them understand the issues with privacy coming up and figure out what do you really need then for a lot of them, you're gonna figure out, yeah, this marketing. optimizing on campaigns would already be amazing. [Laughs] Right. If we could just continue to do that and do that well that would already be a great solution.

[00:44:26] Timo Dechau: Yeah, because marketing is shifting too. I mean, in the end, we are talking about this, I think already for three or four years, but as I already pointed out in the, for example, in the B2B example that I just said, most of the stuff happens afterwards. So most of the marketing activity then happens when you basically have this kind of person who downloaded your white paper and you have it in the CRM system.

[00:44:48] And then you basically run CRM, marketing CRM marketing is a totally different thing also from data perspective and you shouldn't go into this. I think this is another episode to go onto the impact of CRM data. I think marketing is shifting too. And so this is, I think it's interesting because, maybe for this first part of, for this acquisition part, which was so powerful, I think five years ago was dominating every marketing team.

[00:45:09] I would not say definitely becomes not irrelevant because of course it's still important to get people from somewhere. But it's definitely getting a new role. You can find systems that basically then better fit for these kind of roles, which are not. So let's say not so privacy problematic that you use right now.

[00:45:29] Rick Dronkers: Yeah, a hundred percent agreed. So let's take that example one step further. So let's say we want to go with the source medium campaign system, right? So we only, the deepest level of zooming in our data would be a campaign level. So we have source medium campaign.

[00:45:46] That's how we group users. And then we have, I don't know, for page use and some conversion points that those source medium campaigns go through and dates. Of course. What would be your tool of choice? If you go down that route.

[00:46:00] Timo Dechau: In the project that I do, I usually use a tool called AppWalker. They're out of Hamburg, a small startup. They have one benefit. They basically work as a headless system because in this kind of systems, I don't care about any kind of dashboards and so on because I don't need them.

[00:46:14] So what they do is  they pipe the data directly into bit query. And so I can immediately use that. So I don't have to build all the pipelines myself. If I'm bigger company and I would do a risk assessment that I wouldn't bet on a solution from a small startup, cause it's always a risk.

[00:46:29] And let's say bigger setups. What you could do is, I mean, of course, you could go for the big solution, the big guns, which could be SnowPlow. I mean, SnowPlow don't have to discuss it too much, but snowplower, I would say has the industry best collectors, STKs out there, I think has the best way to handle this switch between I identify users and I don't identify users.

[00:46:52] So the snowblower collectors is pretty flexible that I can say when someone comes on the website, I handle them as anonymous. And the point when they give consent, I basically switch and say, I identify you now. And at the point where they revoke it, I just switch back and clear everything. there are others, I think pivot has kind something similar, but I think not on this kind of degree snowplower doing.

[00:47:15] And so. Of course, Nolo is a great alternative for this, but you definitely have to have the organization for it. So you have to have the team that can, if you want to run it yourself, so you have to run it yourself. It's an application and tracking application have one problem. You cannot just restart them and get everything back.

[00:47:30] Once they are down there, you don't get data. And so you really have to make sure that you can maintain it on a, let's say really good SLA, or you can just go to and say, yeah, please do this for me. So this would be the big solution, which, I mean, if I'm a company, for example, who say revenue depends on data.

[00:47:49] It would definitely very high on my consideration list because, this is something, this could be a criteria that I can guarantee my revenue comes from this kind of data some systems in, I mean, for example, pivot can definitely be solutions. So if, for example, I just need these kind of, it's say aggregated campaign data then I could run with them.

[00:48:11] So, because I think it's still, I think it's possible to get the raw data as well from, at least I know that they provided as well. but I never really experimented with it. I guess this would be maybe Matomo as well, so, I already burnt a little bit my finger, sometimes on Matomo. I would revisit it again.

[00:48:28] So I still, I love the product. The approach. The philosophy but I haven't invested a lot of time in the last two years into that.

[00:48:36] Rick Dronkers: Yeah. I, I think, I think the question you asked at the beginning,  what type of company are you right? That, that's of course the there are some assessments, do we make, okay , how much money is at stake here? How many people do we have to maintain it? All these kind of logical things that, from the outside in, we tend to skip over.

[00:48:54] But then from the inside out, you have, you wanna build something that is, so we, if we speak about this 20 million a month, Advertising spent company then probably , it justifies building a team that supports SnowPlow or, or hiring SnowPlow to support it and having, well, you still need somebody on your end, but then SnowPlow takes a bit off of your hands, but that's easy on a 20 million a month advertising budget, you could figure that out. and yeah, if that's not the case, and if it's way less or if it's maybe almost nothing, then there are a lot of solutions out there that, yeah, that could, that could help you go this route.

[00:49:32] Timo Dechau: Maybe for example, just one in which I forgot, which I, for example, tested one year ago extensively and also was running on my own projects. Example, plausible. Plausible is also pretty flexible when it comes to that. So you can get data out, you can run it on your own systems if you want, because in the end it's also open source.

[00:49:49] there are definitely really good tools out there. If what you want to achieve. I think then it makes it a lot easier to select the kind of right solution because you start to ask the right questions when you talk to the, because of course, when you talk to a vendor, it's, everything is great.

[00:50:05] You could, I mean, the sky is the limit. Sometimes there are vendors and I really appreciate them. They tell you where's the limit. And so for me, when I'm looking into new tools, this is the essential part I want to learn. I want to learn the limitation of a tool because no tool is perfect. Every tool has limitations because every tool has a good fit for a specific kind of use case.

[00:50:24] And so SnowPlow obvious limitation. It has a complex setup. If you run it yourself, Also limitation, it just pumps the data in a very raw format into your data warehouse. So you have to take care of what happens after that. But these are  the limitations I want to know. And if  a vendor is not open to tell you about their limitations, definitely. I would not say a red flag, but definitely an orange flag.

[00:50:50] Rick Dronkers: I always insist whenever I talk to vendors, I always insist, I want a technical person in the call and I'm gonna mostly talk to that technical person. And if they, if they cannot provide that then I'm not gonna do the call. Cause I've been in too many sales calls where the vendor is basically promising that everything is possible. And I'm, yeah, but you don't even understand how the tool works. That's not useful.

[00:51:12] Timo Dechau: The nice thing is people who know me from LinkedIn already know this and they already bring someone on the call who can do this. I even had people pointing this already out so they know. Yeah, I know that you do want the, so I already brought our solution architect and so, yeah. Appreciate.

[00:51:26] Rick Dronkers: Okay. Let's skip that.

[00:51:28] Timo Dechau: Yes. In the second minute, usually, so I'm very impatient. So I don't want to see slides.

[00:51:36] Rick Dronkers: Yeah, it's, it's also not worth it, right? It's not worth both of your time.

[00:51:39] Timo Dechau: Therefore what? You can see it on the website.

[00:51:42] Rick Dronkers: [00:51:42] I've been throwing this idea out on LinkedIn, but I want, I wanna have your take on this before we close off. One of the issues of moving everything server-side, which it seems we are going to do for a variety of reasons, I'm not specifically Google tag managed server-side, but a server-side solution. One of the downsides of that is that we remove the, I don't know if it's the word, but the auditability, right? For people to do an audit from the outside in, and I think it this new issue of. Us repeating the same mistake that got us here. Right. And by us, I mean, marketers throw in a lot of stuff that we shouldn't, and then in the end, more privacy regulation and ending up in an even worse place. What do you think we can do improve upon that and to make sure that we don't make that same mistake?

[00:52:34] Timo Dechau: One thing is to include data and engineers, data and engineers by for some reasons, which I never really figured out is, have a very good conscious about these things. So usually when I work with an engineering team, they usually call out things they think is ethically or morally how to say, on thin ice.

[00:52:54] This is not really a matter, this is just an idea. I think one thing that we can, that could work, so of course, we could enforce it. So for example, this is ridiculous, I think was just from bring it out. So of course, DPAs could go in, could set up all the teams would say, Hey, your company has been selected for great yearly audit.

[00:53:13] And so we want to see all the documentation of your setup. And maybe we also do some test runs. I mean, usually what, what we see doing in due diligence, going in and checking how it looks. This will never work because these agencies are not, not in the position and I'm not really happy with these kind of things.

[00:53:30] So, but an alternative could be, for example, something that you see in production and so on. So where you have some specific kind of labors and standards. So the 9,009,000 and something like this, where, where you can basically, a company can prove we did this kind of measures and these to get these certifications, they're pretty hard.

[00:53:46] So you really have to do a lot to really showcase basically what I just described in, in the other example as well. It's so that you can use these kind of labels to gain trust and can say, yeah, so our company is data ISO, blah, blah, blah, approve, use it as a marketing and, and customer trusting.

[00:54:05] And then I think this could be, so I'm more  a friend of, let's say kind of voluntary measures. What is the benefit? And so I would go  this and the good thing is also, it can introduce standards and standards because I think some companies might want to do this, but they don't know how.

[00:54:22] This is also a problem. If you have some kind of, let's say standard approach, then at least you have some measure you have to implement it. So for example, password rotation within your company on these kind of things, I mean, simple measures, but still you increase security.

[00:54:36] Rick Dronkers: I really like that idea. I think, if let's say all tag manager solution providers, if they would all this and then they would all have a hashing method for a field for a variable field. And that is certified by that ISO standard, we hash that standard and we can prove it in, if you export our GTM, Jason, right. You can see it's hashed that way, those kind of things.

[00:55:06] Timo Dechau: Or their tags are certified. So, I mean, this is, that would be also cool. when you use, let's say a Facebook tag and okay.  Everyone is skeptic about Facebook, but it's still, for example, they have the certification that at least to which kind of standards they apply. They might can do other things beyond these kind of standards, but you still know, okay, these kind of things are already.

[00:55:24] Rick Dronkers: We, and I say this broadly, we should take it upon us to, you know, with the move to server site. I think Simo Ahava said with great power comes great responsibility. You can do whatever you want with server-side tech manager, right? Sky is the limit. It's from a technical point of view.

[00:55:41] It's really cool. Right. But it also opens up Pandora's box when it comes to ruining privacy even more. And, in the end it will come out. That's what you see right now with all the, all the lawsuits against, against Facebook and all these other companies, you see, they, keep digging and they keep, eventually it will come out. Let's assume that up until this point, a lot of people did what they did not out of wanting to do bad, but simply because of not understanding that what they implemented might result in this, but now I think the time has come to realize, okay, we can get away with that excuse anymore.

[00:56:16] We should probably try to do better. And should proactively show that we are doing better because otherwise our nice tools will be taken away from us. [Laughs]

[00:56:25] Timo Dechau: I think this is a really good smart. And especially with great power comes great responsibility. I think this is definitely true. the interesting thing that I think you just said, and maybe this is something I can all tell other people as well. It's things usually come out.

[00:56:42] Sometimes it takes use when they come out, things fall, if you do. And I often, these are things where you do it, not by accident, but really on a conscious decision. And sometimes some people will speak. So because they don't think that is a good idea. And so, and things come out. I think this is  maybe also  a good warning for companies embracing the dark side of service-side tag manager.

[00:57:09] Rick Dronkers: Good point. Good point. Good point to close this talk off. It was great talk. Where can people find you online? Is there anything else you wanna share with people before we before we close up, I think the easiest way is really to find me on LinkedIn because I write there regularly. You can just put in my name and then you can find me. And if you want to learn how to create really good tracking plan. So not the bloated ones, but really  the precise ones. So I'm currently running a free course. If you check on LinkedIn or if you on my not linked yet.

[00:57:41] Rick Dronkers: We'll put the link in the show notes below.

[00:57:43] Timo Dechau: That's nice I will post about it regularly on LinkedIn. So follow me on LinkedIn and you will find it. I hope that it helps..

[00:57:51] Rick Dronkers: Yeah. Highly recommend you share a lot of, a lot of good content and thoughts on there. So I can recommend people to follow. Cool then, I think we're definitely gonna have another podcast in the future, but thanks, thanks for today. And let's talk soon.

[00:58:06] Timo Dechau: Thank you very much.