Episode Transcript
[00:00:09] Speaker A: Hello everyone, and welcome back to Couch Confidentials, the podcast where we dive into stories, insights, and sometimes even the myths behind data and marketing technology. I'm your host, Matthew, and today I'm thrilled to have a truly special guest joining us for our second episode. Our guest today is none other than Brian Clifton, a name that should resonate across the analytics and data privacy world. Brian has had an incredible journey that began in Manchester, took him through university in Bristol just as the web was going public, and eventually led him to Sweden, where he's been living since 2008. As a child, Brian has always been fascinated by data, something he's carried through a remarkable career that includes leading Google Analytics in Europe from 2005 to 2008, authoring books like Advanced Web Metrics and even Successful analytics, and now championing privacy and data quality through his latest venture, Verified Data.
By the way, remember those Google Analytics individual qualifications, the JAIQ exams? You've got Brian to thank for that. Brian boasts several more roles such as a guest lecturer at University College London, Copenhagen University, and the Stockholm School of Economics. And if you think I was done, you would be wrong. He's also the Director of Data Quality and Compliance at Search Integration and is an advisory board member for PWIC Pro Analytics. There has to be a clone of Brian walking around somewhere that's nonetheless, it's too long. Anyway, I'll admit I've been a fan of Brian's work for years. I've read all of his books and followed his contributions to the analytics field closely. I've been a user of verified data from day one, and I believe my face is floating around somewhere on the website with a review. But today, we're taking a deep dive into something that's been troubling Brian as much as it should trouble all of us. The myths and realities of enhanced conversions, or as Brian refers to it, cookies on steroids. They could sound tasty anyway, but these are just buzzwords for major ad platforms like Google and Meta, and social media platforms like LinkedIn, TikTok and such that might not be as privacy compliant or user friendly as they sound. We'll also explore Brian's thoughts on what's happening now with data collection and the fine line between actionable analytics and invasive surveillance. This will be a conversation about the crossroads of innovation, ethics and the future of analytics. So grab your coffee or maybe a Swedish Fika treat and settle down for what promises to be a thought provoking chat with the legendary Brian Clifton.
So, Brian, welcome to Couch Confidentials. You're number Two on the list right now. In terms of guests, I have to say I've been a fan of you for a long time. I also have one of your books here. I have read them. What do you call it? The back looks pretty, pretty uncreased. But I usually buy the digital versions as well, so I can read them on my Kindle. And it takes note, makes note making a whole lot easier. Everyone knows you in the analytics world, but one thing that I kind of drew my attention before we get into questions is everybody who seems to have worked with data analytics or analytics in general, has gone to the data privacy side. And we've seen Orly Poles. Who. Who' she ran Lost Boys, I believe it's called LBi. I think the I was for International was an analytics agency, I think around 15 years ago. And she's gone heavily into privacy as well. We see Sergio Maldonado, he ran the dashboard tooling company called Sweet Spot Intelligence. He's also gone into the privacy side. And for the last couple of years we've seen you transitioning into privacy as well. What is the, is it something that we know about analytics that is driving us to be a little bit more thoughtful about what we do with the dev that we collect?
[00:03:58] Speaker B: I think from a European perspective, yes, I don't see this happening in other parts of the world, this transition from data practitioner to privacy practitioner. But I think if you're European based and I think if you've been around a long time, from the early days in the web and you've seen what data can be collected, how benign it can be, how innocent it can be, but also how dangerous and invasive it can be in terms of identifying people, I think if you come from that kind of era and you have a European kind of mindset, the idea of human rights and privacy and all of the history that Europe has with its world wars and its Cold war, I think, yeah, it's understandable how people in data are the main drivers or certainly a big part of the driving force of where privacy is these days.
It's invasive. It can be, it can be incredibly liberating to have all of this data. Let's face it, where are governments going to build schools and hospitals these days? How many staff do you need? How much stock do you have to keep in your warehouse? Data is fantastic. You can't run a business these days without the huge amounts of data that we collect. But I think the big change is that in my career anyway, is that it's gone from being very benign and Very aggregate to being very personal and very invasive. And that's the switch I've seen. And I think that just makes people like me nervous.
[00:05:23] Speaker A: Yeah, I see what you mean. I mean, in terms of if we look at, I mean, you were the head of Google Analytics in Europe between 2005 and 2008.
[00:05:32] Speaker B: I'm responsible for some of this mess, you're right.
[00:05:34] Speaker A: Yeah. Well, no, I mean, it's interesting to see that you've now kind of become more of an advocate for PWIC Pro. Does that have something to do with what you have seen on Google's side or what you've been experiencing since then?
[00:05:52] Speaker B: Very much so. I mean, I was a big advocate for Google for, yeah, a lot more than a decade. I really bought into the Kool Aid when I was there, as you said, 2005. And I helped launched Google Analytics, or universal analytics as it was then.
And well, Google's a very different business now these days. It really is all about the money. I think what happened is, I don't know, 2008, 2010, that kind of period when I was leaving, I mean, no connection. But when I was leaving, I think, to be honest, I think Google looked at Facebook and the micro targeting that it could do, you know, individuals, you know, be able to deliver an ad at any time, any place to the right person.
I think they looked at that and they crapped their pants. I think they thought, oh no, this could overtake our business model. You know, the regular pay per click. And the Google in those days were the contextual marketing Kings with its AdSense that was contextual advertising.
They realized that it's a lot easier to sell personal targeted ads to marketers and it's a trillion dollar industry. So I think they took a path and they've been doing that for the best part of 10 years now. And it's just a path that I don't like. Yes. When it comes to peewee Pro, they have a different approach. They're not adtech, you know, they're not tied to selling ads. So it's kind of going back to traditional site analytics, website analytics as it used to be.
[00:07:22] Speaker A: Behavioral data.
[00:07:23] Speaker B: Yeah.
[00:07:24] Speaker A: And not connecting that to an ad network. So you mentioned you wrote an article and that's kind of the basis of the conversation that I want to have with you today around enhanced conversions. And the first time that I got into touch with enhanced conversions was when I was working with a cdp. And the CDP at the time was segment and they said, hey, we can also do server side side Matching with an ad platform and it was Facebook at the time with their conversion API. You called the conversion or you called this kind of mechanisms cookies on steroids. I mean, can we dive a little bit into that? I mean, that's quite a label. You're hanging onto this especially in light of what I want to say. Duplicitous in the sense that Google's trying to get rid of the whole cookie world, but they're also trying to inject a new and much more aggressive version of it.
[00:08:12] Speaker B: No, Well, I mean Google were forced into this, into this path by the regulators.
So the GDPR and the UK as well were pushing hard to get rid of it and they got themselves into trouble because so many other industries that rely on Google technologies have built up. I mean there's literally hundreds of thousands of people working in ad tech industries. And of course when you dominate that market with your, your third party cookies, then you can't just take them away. So it was a real battle for Google to try and be more privacy focused and just get rid of the idea that you can follow a visitor anywhere around the web but without destroying an entire industry and people's jobs. So I think what's driving the deprecation of third party cookies now, it's not Google anymore because they've kind of given up on this. It's users. Users kind of realize third party data is sharing data around like confetti on the web and that's just not a good thing for privacy.
And so the idea is now to move away from that and use this kind of innocent sounding term of first party data, which is like, hey, it's your data. Why don't we just use that to stitch together all of these different.
[00:09:24] Speaker A: But you're still passing it on to the same endpoint.
[00:09:27] Speaker B: It's still passing on the fact that it comes from your website and rather than being picked up from somebody else's website makes no difference really. So the reason why I use the cookie on steroids metaphor is the fact that there's so much more invasive than a cookie could ever be. So enhanced conversions, as you know, it's based on hashing personal information, which is.
[00:09:49] Speaker A: No security at all.
[00:09:51] Speaker B: No, exactly. It's just a unique fingerprint now. So we've gone from this kind of just a little text file that's on a device to something that if I submit my telephone number on my email address, I mean, I don't change those very often, you know, every 10, 20 years maybe. So it means if I input that same Information on another website anywhere online. You hash it with the same algorithm. It's the same unique fingerprint. And so they've got me and they've got me for a long, long time. So that's where the analogy comes from.
[00:10:20] Speaker A: How confident are you then? With some of the CDPs that I have worked with, they do offer a feature where if you have a deletion request, you can pass that on to advertising networks. The CD very clearly state, hey, it's not our responsibility what happens after we pass it on. How confident are you that these advertising networks are actually deleting your data? I've seen methods in other companies where they will delete any non hash data that you might have, but keep all your, let's say all your orders and all your behavioral data so it doesn't skew the rest of the data that they've already processed.
[00:10:57] Speaker B: So well, you asked me that question three years ago or even less. I would say not very confident. I think the regulators now are getting tougher mostly because there are other organizations, privacy activist organizations.
[00:11:12] Speaker A: Yeah, none of your business wanted being one of the bigger ones.
[00:11:14] Speaker B: Yeah, noib none of your business by Matt Scrams. You know, really pushing the regulators to enforce the law. So I certainly think now it's better. You'd be in serious trouble as a business if you were not honoring deletion requests.
I think the challenges with CDPs is there's a real blurring of the line of what is a customer these days.
Obviously if someone's transacted with you, they're your customer.
[00:11:39] Speaker A: Yeah, absolutely. Good point.
[00:11:40] Speaker B: You can handle what that means directly with them because you have a relationship with them. In fact, you actually have a contract with them because you've got to deliver goods and they've got to pay for them. So a cdp, the traditional cdp, which is kind of like a souped up CRM, kind of no problem with that. I think users are kind of happy with, you know, customers are happy with the fact that you're going to use the data to give them better offers, better deals.
[00:12:03] Speaker A: Yeah, personalization is expected these days, but.
[00:12:06] Speaker B: It seems to be that CDB is kind of now being dragged into this kind of, you know, anyone on your website world just because they made some kind of contact and these kind of, you know, prospects rather than your customers. That's a bit of a shady area for me to kind of class that as a cdp, because ultimately then most of these people that contact you on the web aren't your customers. That's a very General statement, but generally most people that go to your website are anonymous to you. So to try and suck them up and say are there our customers and it's first party data and we can do what we want, that's a real blurring of the lines for me. So that's where I kind of hesitate on that. CDP is in principle great, but the way it's being used in the anonymous way to replace your web analytics tool I think is a step too far from a privacy point of view.
[00:12:52] Speaker A: Yeah, because so when we look at it from that perspective, you weren't happy at all that for instance, Google Analytics 4 now allows you, or now it's been some time now to build audiences that you can sync with Google Ads. So based on the events you track in Google Analytics 4, set up the audiences or at least build the audiences in GA4 and then have them synced to Google Ads so that you can remarket them. That must be kind of a nightmare scenario that you were not hoping that Google Analytics would move towards when you were at the reins.
[00:13:20] Speaker B: Well, I think, yeah. I mean going back that length of time now, I'm very much a champion of this kind of aggregate data. I think you can do a huge amount of optimization of your marketing and your website and the user experience without knowing anything about individuals. You know, you just look at where they come from, what ads they clicked on, how many came from Google AD1 versus Google AD2. So there's a ton of stuff you can do with aggregate data. So I'm not a big fan of that in general. But ultimately, in terms of building audiences and remarketing, it all comes down to whether you have the permission. Now in Europe, that's a fundamental law. Now, that's what kind of GDPR and the E Privacy Directive are about.
[00:13:59] Speaker A: Yeah. And those fines are only increasing as we've seen in some of the reports lately.
[00:14:03] Speaker B: So if you have the permission, yeah, you have the right to do what you said you're going to do when you ask for the permission. The problem is people often don't know what they're agreeing to. When you see these pop up banners, they just want to get rid of the banner. And so they often say, yes, I give you my permission without really fully understanding what that means. And that's a challenge because to explain what it means is an awful long document that nobody wants to read. So we're in this terrible dichotomy at the moment with these pop up banners which nobody likes. Privacy advocates like myself, don't I hear.
[00:14:37] Speaker A: My Wife complaining about them all the time.
[00:14:39] Speaker B: Everybody hates them. They're an awful user experience, and people just want to get rid of them as fast as possible. Now the way I do that is I just reject everything by default. There's a lot of people that worry that the website is just not going to work or they're not going to get the offers or the user experience that they're expecting. So a lot of people feel pressured that they have to opt in without understanding what's going to happen next. And that's really the problem. So in theory, with, you know, having consent, having the permission from your users solves all of these privacy problems. And I'm happy with that. But that's the problem. People don't realize that.
[00:15:13] Speaker A: No. And when I was still working a lot back with analytics and I was doing a lot of user research, I used to have a blog called Actual Insights, and I would try to do tests and publish the results online. And one of them was just around consent banners.
One of the discussions that popped up constantly is what is the responsibility of the business? Look, if I'm trying to offer you a service that you're interested, you've shown interest because you've visited the website, you're curious about something. It could just be company information or an actual product. I should try to convince you why you should opt in. But that's never the case with these consent banners. It's all in the small letters. I've never seen anything really enticing saying, hey, you see the popups now everywhere about newsletters? Get a 5 to 10% discount on your first order. That's a great move. But what is your advice? What have you seen? I mean, I've. I mentioned a long list of roles that you do. A lecturer you teach everywhere, you work for, search integration. But what should companies do to still remain within the confines of the law, but still get the most out of the visitors, anonymous and known users?
[00:16:26] Speaker B: Well, I think it's very simple advice, which is, you know, you have to build trust, you know, with your prospects and with your customers. I mean, if you lose that trust, it's a bit like love, you know, once you've lost trust, you're never going to.
[00:16:38] Speaker A: I think we can write a song about that, right?
[00:16:40] Speaker B: So I think that it's all about that. And some brands have spent decades, maybe hundreds of years building that trust. You think of companies like Volvo, for example. It's all about safety and brand trust and things like that. And it's very easy to destroy that in the Digital world, I mean, it just takes literally a few seconds, maybe even less. So I think keep it simple. That's the one thing. Don't overcomplicate it and focus very much on what your customers, what your users want. So a lot of people, when they're building these banners, they say, how do we categorize cookie X or how do we categorize pixel Y loading on our script? Oh, well, that's necessary. It's necessary for us to run our business. But actually that's their point of view. That's the business point of view. I totally understand that. But from a privacy law point of view, you have to think of this from your customers, from the users. So if it's not necessary for the users to have that pixel on the page when they're looking at their service, in other words, it's an extra feature, it doesn't really affect them, then it's simply not necessary. And that's the dichotomy that we have is what's necessary for the user that's visiting the site versus what's necessary for the business. From a law point of view, it has to be the user. You have to think in those terms as a user and keep it very transparent and very simple.
[00:17:54] Speaker A: That kind of brings me to, well, a story that I had when GDPR was first introduced. I remember the night before I was working for a company and they said, we need to have the, we need to be GDPR compliant from the next day onwards. So I was working at Google Tag Manager, making sure everything was set, but I believe it was within 48 hours. They saw the dip in marketing performance. So within a facet that you could probably jump up and run around your house and say, hey, I'm done. The big bosses were on the line saying, hey, we need to revert that because we're not have, we're not gaining any insights on our marketing. Is that just kind of the hit a company needs to take for the team or that we just need to accept as is, that we will not be able to report on 100% of the users? And if that's the case, what would your advice be to those kind of companies who do want to try and see the full picture, should they just fill out on averages, there is going.
[00:18:50] Speaker B: To be a hit? I mean, if you give people the option to opt out transparently to be in track, there's going to be a significant proportion of those that will opt out. Yeah, I mean, I work with some brands where that's as low as 20%. So 80% of the people that you know, of the total population side opt in and allow to be tracked. And that's a very significant proportion. And you can probably carry on as you did before. But these are brands that have spent, as I said, decades, even longer and an awful lot of budget on building that trust up so that it's kind of a given when the visitor comes out. They know who they are and they're happy to, to hand over the data. If you're not a very well known brand or. Yeah, you just don't have that kind of trust built in yet. If you're a new player to the market, for example, you're going to struggle because that 20% grows and I've seen as high as 50%. So losing half of your traffic, that's not half of your visitors. They're still the same. Visitors come, but they opt out. So you've really got to work on that. I have a slightly controversial view. A few years ago I wrote a white paper on this and I still get people, you know, slagging me off because of what I wrote. But the view is simply this is that, you know, I'm quite pragmatic. You know, I run a small business myself, so I come over from a business owner, even though I'm a data practitioner and also a privacy advocate. So I have multiple hats and they all kind of compete with each other.
[00:20:12] Speaker A: Yeah.
[00:20:13] Speaker B: But my view is, you know, one click for accept all and two clicks for reject all. So you come to a banner you want to accept. Okay, great. You know, we have all your data. Of course, that's what the business wants.
[00:20:25] Speaker A: Yeah.
[00:20:26] Speaker B: If you want to not accept everything, then I want you as a business owner to just think, just take a second to think and you know, to kind of enforce that. It's like, okay, here's another button. The customize or reject all or customize is kind of the what the button says. So they know if they want to reject. This is the button to go to this other button, they click on that and then it's just one more click to say, no, I don't want any tracking. So two clicks for reject and one for accept. That's quite controversial because a lot of people say the law says it must be equal. So one law, sorry, one click to accept and one click at the same level, first level to reject. I'm just a bit more pragmatic and say it's a balance between the business having a right to do business and the individual having the right to be private.
[00:21:12] Speaker A: Yeah.
[00:21:13] Speaker B: And of course you've got to say, where's the limit? Is it 2 clicks, 3 clicks, 4 clicks? I have a very strict rule, it's 2 clicks and nothing more.
[00:21:21] Speaker A: But let me change that around a bit because something else that I've never seen in my life is, let's say I go to a website and I decline to be tracked by them. Of course they're still going to. I believe it is still legal for them to track a potential return visit. Am I wrong on that one? Because they'll know that I've rejected the tracking. So there is, there is a cookie on my system saying that I've rejected.
[00:21:47] Speaker B: Well, usually it works by the fact that if there's a cookie that means you've given some kind of acceptance. Yeah, there's no cookie. Then show the banner, basically.
[00:21:55] Speaker A: Okay.
[00:21:56] Speaker B: That way rather than.
[00:21:57] Speaker A: So if I, if I was to reject it would show the cookie banner the next time again. Okay, yeah.
[00:22:04] Speaker B: I mean if you've actively clicked on the banner. Yeah. Lots of different banners work in different ways. But if you've actively made a selection then it's likely to cook you. But if you haven't made a selection, if you haven't given any check.
[00:22:15] Speaker A: Okay, yeah. If you haven't, haven't clicked on it at all. So. But let's take the scenario where I've hit reject.
Would it be feasible, would it be allowed to, for instance, after you've gained some trust that you seem to have been doing business with the user coming back to your website to show the pop up again or to ask for permission to track them again? Because that's something I've never seen in the real world now. And there are plenty of sites where I've rejected, but I visit frequently.
[00:22:45] Speaker B: Yeah, I have seen that. Yeah. Where unless you, unless you accept the banner keeps annoyingly popping up.
I think that's just bad user experience. I think the only. Putting the law to one side, I don't think the law is interested in that level of detail. I think all you're doing is really messing your users around, your customers around and they will speak for themselves by just going somewhere else. That's how I would focus that one rather than a legal point of view.
[00:23:07] Speaker A: But yeah, absolutely.
[00:23:09] Speaker B: Well, the banner stays in place. You know, even if you hit reject, the banner just stays in place. And so you have this kind of letterbox view of the website and of course if you, if you do accept all that letterbox expands and you get the full view. But yeah, I've seen that as well. Yeah.
[00:23:22] Speaker A: And that letterbox irritation can go both ways. It can be an accept or full on reject there. So when I was doing implementations, I. Have you, have you done a Google Analytics implementation? I know you were head of Google Analytics, but have you ever done an implementation yourself?
[00:23:38] Speaker B: I would say hundreds, if not, well, maybe not thousands. I'm a consultant rather than a, than a factory worker, but I would say hundreds. Yeah. GA4 universal and a few other tools. Yeah.
[00:23:49] Speaker A: Well, so have you ever been in a situation where a customer says, hey, just, just track everything? Where I kind of want to lead to is that before you just mentioned, hey, you know, every piece of data that you want to collect in some way needs to serve a purpose. I was working with a customer in Saudi Arabia and within pdpl, it's, it's one of them, it's one of the main rules. You cannot use any type of data unless it serves a business purpose in terms of processing. I know that in the Wild west days of Google Analytics, I personally have been one to say, hey, let's track everything. I would go through a website, just track all the interactions that, or at least document all the interactions that we could potentially measure, put that into an implementation plan and get that sorted. But I mean, is that something businesses kind of need to roll back now? You know, if you look at Shopify, when we do an integration with Shopify and Google Analytics, they'll, they'll track every E commerce feature that Google will offer, add to cart, view list, view item, et cetera. So is that, is that going too far?
[00:24:46] Speaker B: Well, I mean, it used to be kind of in the early days, it's, you know, this idea of the signal versus the noise. I think it's one of the chapters in my book is, you know, in the very early days, the early 2000s, it was collect everything and then figure it out later whether it's important. Then it became, well, hold on a minute. A lot of this is noise and you have to often pay for more noise because it's more processing, more storage. So, you know, and I've literally worked on setups where the noise is 90% of the data. So not only does that mean you're paying, you know, often 90% of your bill is just wasted. So that's one incentive for the client. But also you can't see the signal because you're buried. It's a needle in a haystack. It's buried under all of this noise. So that was in the early days. I'm talking 10, 15, maybe more years ago was the kind of the rationale of collecting what you need and understand what you're collecting so that when you analyze it, you're making good, solid decisions.
Now it's kind of obviously much more privacy first it's like, okay, what is the purpose of this data? What are you using it for? And of course all of that impacts on banners and things like that and trust. So I mean, I work with pretty much exclusively European clients and there has to be ultimately a sign off process with the implementation plan by the compliance team. So they have to feel comfortable that all the checks and balances have been put in place. So it's certainly different. But I've always had this mentality of data minimization, as I say, come from the noise. But now it's very much about privacy and purpose.
[00:26:14] Speaker A: And do you see that with all your clients now that compliance gets involved? Because personally, in most of the projects that I've worked with, I probably can count on one hand the amount of times that compliance has actually been involved in a sign off. It's Saudi Arabia, definitely, but you know, in Europe, not yet is that different per company size, but maybe. Is there a relationship there?
[00:26:33] Speaker B: Yeah, I mean I work very much at enterprise scale these days. Yeah, every client is talking about compliance.
There's been some big decisions in the part of the world where I work, some big fines for a bank and for online pharmacies and we're talking multi euro fines and plus the awful PR that comes with it because it destroys trust. And there's plenty of competition for banks and online pharmacies. So I'm pretty sure there's a big business cost to that decision that went against them as well as the actual fine cost.
[00:27:09] Speaker A: Yeah.
[00:27:09] Speaker B: So in Europe, I would say pretty much every client has compliance as part of their analytics thinking at some level. Not some of its details, some of it's superficial and I help them with that, but it's there somewhere.
[00:27:22] Speaker A: So coming back to your signal versus Noise and this whole compliance part, this is not a question that I had planned, but it does plug in really nice into Verified Data. Is that partially part of the motivation to start Verified Data as a service? Because you focus a lot on the quality of data tracked within Google Analytics as a tool. And I believe you also highlight the signal versus noise level there on the differences between typical events, but also on the compliance side where you do the page inspector.
[00:27:53] Speaker B: Well, I mean, you asked me a few minutes ago about how many setups have I done? Literally hundreds.
[00:27:59] Speaker A: Oh, it's thousands. Just a Few minutes ago. But no, I'm just kidding.
[00:28:02] Speaker B: Well, spoiling the timing of the podcast, but.
So I've been involved in a lot of setups and the reason that I got involved with setups is I just couldn't do any analysis. The data quality was so poor. I was looking at things and saying, well, that when I look at it this way, the chart goes up into the right and when I look at it that way, it goes down into the left. It's like, yeah.
[00:28:21] Speaker A: And if I can interrupt just for one second, that's one of the reasons I got into tag management, because the quality was always so poor.
[00:28:27] Speaker B: So I was, I got into that a lot. And in order to show this, in order to figure out, okay, what do I need to clean up? Do we need to start again? Is it just one thing systematically wrong. I got into auditing. So I was doing manual audits for many, many years. And these are hard work. They make your eyes go bad and you lose your hair and things like that, as you can see. So that's how I got into auditing. And then I wrote my third book and I thought, okay, I need to describe auditing because it's so important to the data quality to not only to know what the data quality is, but to keep on top of it because it changes very, very rapidly. That's what I saw. You do a great implementation six months later.
So I got into auditing that way. And these are, yeah, time consuming, difficult audits. I could never run more than two or three in any one month. And I thought, this must be able to scale. And of course it can. You know, that's what machines are for. So that's how I got into verified data automating, what I could do, you know, looking at browser tools and dev console and things like that and just saying, is there a better way we can do this? Crawl a website, listen for the signals, look at, you know, if you have a consent banner, click on one of the consent choices and figure out, okay, what are the cookies being set? What are the pixels being set?
Does that match the consent that I've just chosen? So that's how you figure out compliance. And of course, if you have Access to a GA4 account, we can use the API and log in there and look at other signals about data quality and pii. Is there any email addresses or passwords being sucked into Google Analytics, stuff like that?
[00:29:59] Speaker A: I was biting my lip. I wanted to mention PII so badly because I, I, yeah, as I tried to, trying to see how I can frame this. I, I don't think I was one of the first users of Verified Data, but I thought I got on board pretty quickly. And when I did start using it, I was so surprised about the number of hits I would always get on PII data. It'd be an email address, postcodes, full addresses, just plugged into the URL. And it's, it's always been fascinating and it's. I don't know if you do any kind of aggregate data collection across all the accounts that you have with Verified Data, maybe this falls under compliance. But are you able to see the amount of PII data that's collected in J4 accounts in general?
[00:30:39] Speaker B: Well, I see the failure signals because I'm always interested in what fails. I don't see the data that's okay protected, but. But I see there's a lot of signals. I mean, certainly when it was universal analytics, it was awful. I mean, it was everywhere. And I think that's one of the drivers why Google switched to GA4 and a different data collection model. Because the quality of data was so easily bad and there was so much PII in there, I think they had to close the system pretty much because of that. So it's definitely better in GA4 that I do know. But I came across it recently for a client that I was working with literally this week.
Email addresses and passwords in plain text and plain text.
And in 2024, you know, here we are six years after GDPR. It's very surprising, but it does happen. It still happens.
[00:31:28] Speaker A: Yeah. And it's so easy to make a business case for audit work just using Verified data. Not again to plug it even. My personal experience is that it's just made everything so easy. But when you look at the clients that you work with with Verified Data, do you kind of see a trend in that they're adopting audit more on a regular basis? I usually use it as an introduction. Like you said with Page Inspector, you can use it even if you don't. You don't need access to any account. And I remember you telling me saying, hey, if you want to use Verified data, you just do the Page Inspector and go talk to these customers whose websites you're collecting data on. But I've never been able to see personally with the accounts that I have is that they're doing this on a regular basis. Do you see a difference in companies who do do it?
[00:32:12] Speaker B: Yeah, I think that's really buy in. You know, if you have somebody at the client side who understands the importance of this rather than do a one off audit, which is maybe what gets you through the door. The key really is to do monitoring. So you want to regularly do this. You can do it daily with our tool to see what's going on and flag you if there's a problem. So auditing is great, like first time, what's happening, why is it happening, where is it happening, you know, if there's any problems but you want to be flagged if there's any changes over time. So yeah, if you get that buy in from the customer, I think the monitoring is kind of sells itself.
[00:32:48] Speaker A: Yeah. Okay, quite interesting. I want to get back to PWIC, bro, because PWIC I think is for the last two years is. Well, pretty much ever since Google Analytics 4 more or less launched, Piwic has been gaining a lot of ground in the industry in terms of implementations or being becoming the de facto analytics solutions for a lot of companies. And I've also noticed you've been doing a lot of blogs with them. What drew you to PWIC Pro as a company and yeah, what do you think that they are doing better than Google at the moment? You already mentioned that they're not involved in ad tech.
[00:33:20] Speaker B: Well, I think that there's, I mean going back a few years, Peewee Pro looks very similar and works in a very similar way to Universal Analytics. And I love Universal Analytics. I think it's a beautiful interface. It's limited of course, in terms of its flexibility, you know, this kind of hierarchy of event, category, name, action and value. It is limited, but I think it's a brilliant tool and still is, I think groundbreaking in terms of its user interface of getting to information quickly and intuitively in a way that GA4, let's face it, just is not so a lot more flexibility in that product GA4, but a terrible user experience in my view.
In fact, some people just don't even use the user experience, sorry, the user interface. They just use, you know, what comes into BigQuery.
[00:34:07] Speaker A: BigQuery, yeah. And that's.
[00:34:10] Speaker B: And there's a need for that. Of course. You know, there are some people doing some very advanced, you know, analysis and slicing and dicing but, but I think there's a huge need for businesses to use a pretty straightforward reporting tool. And Universal did that very well unfortunately, you know, rest in peace. But kind of Peewee Pro takes over that mantle. It's very, very similar. If you like Universal, you'll pick up Peewee Pro as you know, very quickly. So that's part of the thing I like about it just the kind of simplicity. It's kind of built on the principle of aggregate data. You can take it to another level. They do have a cdp.
[00:34:43] Speaker A: Yeah, I was just about touching that.
[00:34:44] Speaker B: Yeah, it has the principle of aggregate data and privacy by design. So built in in terms of how it stitches data together. So there's a post on that. But anyway, so there's that part, there's the other part that they're very much about building a community. This is what I liked about them, is that they kind of just have a different approach. Not being ad tech, not being Martech, just going back to basics of what analytics and optimizing your website and marketing is about. So all of those things attracted me to them and just at the same time when Google just seemed to be going the other way. So it wasn't that peewic shone out as a great beacon. I think it was a combination of they have some good stuff that they're doing and Google was really kicking themselves in the head in terms of what they were doing with the user interface, the way they managed all of that. The privacy debacle that's been going on for the last few years. The community that they really dropped the ball on, I mean, I helped build the community in the very early days recruiting partners like Arelli Polls. I first met her trying to recruit her as a partner for GA in the very, very early days, 2005, something like that. So you know, all of those factors kind of combined and I just thought, yeah, I'd like to talk to these guys. And they were also interested in talking to me. So that's how I got involved.
[00:36:01] Speaker A: And how do you see that they're going to be pushing the bar even higher now because I mean recently implemented it on martechtherapy.com and what I liked about them, I had discussions before with them around their cdp, which I find fascinating, but I still think it's very early days in terms of development. But what I like about their system is that it's integrated. You have your analytics, they have tag management solution, but they have consent integrated in the entire process. But how are they going to push the bar even further, especially once they develop the cdp? Yeah, sorry for focusing on cdp, but I'm curious to see how are they going to push that with consent and compliance into that CDP being a non ad tech focused company, I mean, are they going to add extra features for the enhanced conversion parts?
[00:36:47] Speaker B: Well, if you think of, I mean CDP obviously has a great attraction for Martech. You know, building your audiences and pushing them out for for different ad tech providers to display the ads. But CDP can do a lot more than that so it can optimize the experience. When the visitor's on the website there can be pop up messages. Did you see this offer? Did you know if you bought two of these you get one for free? So there's a lot of real time optimization that you can do with the CDP once you.
[00:37:15] Speaker A: Absolutely.
[00:37:16] Speaker B: So there's that kind of angle to it. I think something like peeweek Pro, they can be a lot more proactive in detecting ad fraud. So visitors that click on an ad that aren't real visitors basically and you as an advertiser are paying for those. How big of an issue is that? I don't actually know. There are some people that work in this area. I'll give a shout out to a guy called Augustine Fu who built a product called Foo Analytics. Fou smart guy. Known him for a few years, I've been following him for a few years and just seeing he writes a lot about ad fraud. He thinks it's huge. It's much bigger than anyone's admitting, but you never really know until you measure it on your own website and your own campaigns. And I think that's something that I know peeweek Pro are experimenting with and you can imagine Google's never going to expose that. I just can't see that happening.
It goes too much against what ultimately they make their money on, which is a whole, you know, media landscape and advertising.
[00:38:20] Speaker A: Yeah, and even meta. The number of discussions that I've had with clients trying to explain to them why the data within meta does not match the data that they would have in their analytic systems. The number of visitors, the number of the clicks and conversions, even though the setup has been tested through and through. So that's an ongoing discussion. But when we look at on a recent project that I was on the topic of on premise came up and I have to say the more that I've been reading about on premise solutions and they're again not necessarily CDP only but being able to privately cloud host or physically on premise a solution, would that tackle privacy and compliancy a lot better in your opinion? I mean the argument that they, some of these companies had about not going on to a SaaS solution is that they don't know where their data is being duplicated to. So you might have a location in Frankfurt, but who says that your location is actually staying in Frankfurt?
[00:39:23] Speaker B: I mean if you want to stay in a cloud environment, the way for that is private clouds where you literally own or lease and have complete autonomy over the hardware. So there are private clouds, you clearly pay more money for that, but it's just you on the hardware and it's literally, you know where in the data center that hardware sits. But it's still cloud based. I think for large organization those private clouds are the way to go on premise. Puts a lot of overhead on your IT team just definitely managing server log files. So it's not something Peewee Pro do. For example, I know they, I think they had a solution for that many years ago, but the overhead was just too high. I think.
[00:40:00] Speaker A: Matomo, yeah, they split into two, I believe two different entities. One was P Wake and the other one was Matomo. If I'm saying that correctly, I might get hit on that.
[00:40:09] Speaker B: How many years ago? I think 10 years ago there was a close connection where Peewic was developing a lot of the kind of add on features and then they just said, well, we want to go down an enterprise route. And I think I'm right in saying that Matomo wanted to go much down a kind of small business route. Analytics for everyone. So it probably is, yeah, I guess analytics for everyone as well. But it's much more about enterprise scale or certainly people that are prepared to pay for analytics. Whereas Matomo can be $20 a month, something like that. So a different scale but still very good tools, both of them.
[00:40:42] Speaker A: And I don't want to use the term because it's everywhere these days. AI, what do you, you've got 20 plus years experience in analytics and now you see the rise of AI everywhere. What do you think? Do you see any good possibilities of using AI on top of analytics? And what are some potential pitfalls that you see looming on the horizon?
[00:41:02] Speaker B: Well, I mean I flip flop on AI every single day. I mean some days I think, wow, it's the next big thing. This is a new industrial revolution if you like. And we, I guess until the web came along, industrial revolutions happened every once every 200 years, something like that. Now we seem to be having one every decade and even less than that. So sometimes I'm very optimistic about AI and I use in my day to day stuff, I use a lot of chat, GTP just to get to the answers quickly. But there's always this caveat in my head, okay, what's the source of this? I mean I don't look at controversial subjects, I just look at data. Tell me the top 10, I don't know, finance websites in Europe and plot this by Revenue and show me which country they are registered in and chat. GTP can generate that in seconds. Whereas if I was to do that research online would take me hours.
[00:41:59] Speaker A: Yeah.
[00:42:00] Speaker B: So it's a fantastic tool to speed things up. But if you start thinking about controversial, you know, anything to do with race or politics or war, then, then AI scares the shit out of me.
[00:42:10] Speaker A: And in combination with geolocation data is getting even scarier then.
[00:42:15] Speaker B: So I think there's a lot of potential you think about anonymously.
I knew I, you know, odd data, you know, looking at odd spikes that are perhaps buried again, you know, under the noise. How can you pick those up? And you can only really do that with, with some kind of machine learning. And the same with. We talked about verified data earlier. How do you detect PI? Pii, your personal information.
Again, that can be buried, you know, it might be one hit in a thousand email address. And is that number really a phone number or is it a product number? How do you. Or is it a credit card number? How do you differentiate a string of numbers from being something innocent to being something that's pii? So you have these learning technologies and AI technologies that can really help you just bubble up the mess that I think is often hiding under the carpet. So great opportunities. But also I do worry about the health of my kids, the mental health of my kids.
[00:43:11] Speaker A: Is it safe to say that you're considering AI for verified data then? To some extent to.
[00:43:15] Speaker B: We use, we use some of the Google technology to do that. The one I just described about hunting down personal information, that's a Google product that we use. So they're just very, very good at processing large scale data. So we do use it. So yeah, it's kind of on my mind. I'm trying to think how we can better do it. A lot of it is really decision tree logic. It's going through and asking a question. Yes, no, and then going to the next level.
So some of it is really what you would do if you were looking at a website yourself and just looking at dev tools. But anything that makes things faster, of course I'm interested in.
[00:43:51] Speaker A: Yeah, I can imagine that. Yeah. So you mentioned your kids and I think that would be a nice hope kind of heartwarming subject to end the podcast on kids and data. I think your experience and the experience of a lot of people in the data world or should enable us to act as educators around data. And I know from personal experience that I've had to teach my data or my, my data, I've had to Teach. My daughter begins with the same two letters around the the apps that she's installing on her phone and I've gotten her so far. It began with her wanting to install TikTok. And I sat her down, I showed her some short documentaries that Dutch television had made around data collection of TikTok, showed her where she could find the information when installing the app. And I've gotten her to the point now where she's actually coming up to me saying, hey, dad, can I install? I've looked at the data that they're collecting and there's no personal data and I don't need to share my terms of service.
[00:44:46] Speaker B: Fantastic.
[00:44:47] Speaker A: It feels like it.
All right. I'll be proud that I've gotten her so far, but she's only one in billions of kids. How is this, do you think there will be a trend with this awakeness around data and privacy in the next generation?
[00:45:04] Speaker B: Oh, definitely. I mean, I've done some kind of guest lecturing at various universities, so. And in Sweden, in Denmark and in the uk. So I'm talking to people that are, you know, at least a generation younger than me, you know, the 20 somethings. And I've really noticed a big difference. When I started doing that, let's say 2010, that, you know, I had a. Most of my presentation was supposed to be about data collection, but I spent a lot of time focusing on privacy and the implications of what this means. And I remember the guy that invited me to give the lecture just saying, well, why do you spend so much time talking about privacy? Surely no one's interested in that. Which I thought was interesting in itself. Yeah, well, you turn that around, you fast forward that to like, you know, 10 years later to say 2020. And you know, those students are much, much more privacy savvy. You know, they're having lessons about awareness of social media and tracking and stalking and all of these things in school. So when they get to university, it's actually, they're very, very interested. Oh, you can do that with a day. And this concept of, okay, it's anonymous, but if, you know what, if you have enough pieces, anonymous pieces of data, you can stitch them together like a jigsaw and reveal the identity of someone. And there's plenty of academic studies out there. So I actually find, now I haven't done one for a couple of years, but the last set of guest lectures I did, people were asking me much more to explain about privacy and its implications than ever before. So I think that's just in a space of 10 years, I think people are. We've had these TV programs, you know, the start of Snowden and then the Cambridge Analytica mess. People have watched these and are interested in them and, you know, they comment on them and they understand what being online is now.
So I think, I think we're in a good place. And that's.
I'm not saying the whole world is like that and more of it, you know, the more that type of stuff we can do at an earlier age, just to educate kids what trust means.
[00:46:58] Speaker A: And again, it'll surely inspire some innovation and it'll be interesting to see which way that goes. Are they going to find new ways to be invasive around our privacy, or are they going to find out ways that will make it more easier or, I don't want to say fun, but a little bit less difficult to make a decision about who you're going to share your data with, if it's going to be on the phone level or within a browser level or. Companies need to start taking more responsibility. Yeah, I think this could be, could be an interesting path that's going to show itself.
[00:47:27] Speaker B: But you think in, you know, the parts of the world that we live in, you know, kids, you know, sort of older, teenagers, they have more disposable income than I have, you know, so.
[00:47:37] Speaker A: They are, they make it look that.
[00:47:38] Speaker B: Way for people to market to. So, yeah, I understand there's a driver there. And that's why you have to have tough regulation, not to be party poopers. And they often hear this argument from the Americans that we in America are a lot more innovative and that we have Google and we have Facebook and you have nothing in Europe. And it's like, yeah, but that's because, you know, regulation and controlling a level playing field and stopping these huge monopolies even forming in the first place, that's an important part of our culture and how we work. I don't say they're insurmountable. I think, you know, we can work very closely with our American friends. But, you know, there are differences in terms of culture and how we, we treat data and individuals, you know, as in Europe, as humans and in America's citizens. So literally, if you're not an American, if you're not a US citizen, you are literally an alien. You know, you're not just a foreigner, you're called an alien. That's how they define it in law. Whereas in, in Europe, if you're a US citizen living in Europe, then you're still a human being and you come under fundamental Human rights and privacy rights and things like that.
[00:48:42] Speaker A: So that's interesting that you say that. Yeah, I've, I've become an alien since March 8th. I believe I gave up my US citizenship. But that's, it's, it's more tax related than anything. But I fully understand what you're saying. And it's interesting that I never heard that comparison before between citizen and alien and within Europe as human. But I wanted to say those last one, just, the one just popped into my head. You move a lot between Sweden and other European countries. Do you see a difference even within the European Union or the European culture around privacy and in different countries?
[00:49:16] Speaker B: I think, well, certainly in terms of, you think about, you know, data collection, the problems, the challenges, the worries, the concerns, the resources that people have or don't have, I would say it's pretty universal in all parts of the world. It's a challenge for everybody to have the resource and to have the knowledge.
You know, data collection, where does it go to? I think, you know, all of these. I haven't seen any kind of differentiation, you know, even size of business. I find, you know, big businesses has the same problems pretty much as small business and vice versa. So I found very little differentiation in terms of culture or language.
Same challenges, you know, surprised. Even some of the biggest brands in the world, you know, struggle with resource. They obviously have bigger problems, but they still struggle with resource. And small businesses, by definition, you know, they struggle with resource as well. So.
[00:50:09] Speaker A: Yeah, comparatively.
[00:50:10] Speaker B: Consultant.
[00:50:13] Speaker A: Exactly.
[00:50:13] Speaker B: Same challenges.
[00:50:15] Speaker A: No. Hey, so 2025 is around the corner.
My interests are, I'm curious to hear about what do you see happening on maybe data collection or privacy in 2025 or would you like to see happening?
That list might be longer than you can answer the question with. But. But yeah, I'm curious to hear what your thoughts are about 2025.
[00:50:39] Speaker B: Always scary to do predictions. I'm sure this will come back and haunt me.
Just as I was leaving Google 2008, Jim Stern, who's kind of, you know, old master of web analytics industry, he asked me the question as I think conference in London. He said, brian, what's the next big thing in 2009 and beyond? And I said, privacy.
And there's a few sniggers in the room and it was very quiet. He goes, oh, is that it? Is. Is that what Brian's got to say? And I talked about the privacy and the importance of it, and of course I was wrong. It took another 10 years before privacy really became, I would say, mainstream.
[00:51:13] Speaker A: Yeah.
[00:51:14] Speaker B: So I think that the privacy snowball is still rolling.
As I said, in terms of consent rates, you know, we're somewhere between 50 and 80% of people opting in. So that means an awful lot of people not opting in. I think that's going to grow.
I think there's real challenges in terms of using enhanced conversions, which is Google's technology, and automatic Advanced Match, which is Facebook's first party technology to get around using cookies.
Other providers do that as well. So I think there's still.
Privacy is not dead. I think it's very much a living, breathing issue that we have to kind of address. And I think, I don't know whether it be next year, but I think a big thing that's going to come back is contextual advertising.
[00:52:04] Speaker A: So yeah, you mentioned that.
[00:52:05] Speaker B: Yeah, moving away from this personalized, you know, profiling the hell out of everybody, trying to figure out my gender, my income bracket, you know, my age, and just going to, you know, regardless of my age or, you know, what is my interest. So if you go back to context, contextual marketing, which ironically Google was king of, you know, 2005 or that time of period, I think we're going to make a move back to that. Driven by regulation, privacy regulation, but also by, I think advertisers realizing that that's really the only way around this. Otherwise they're just going to lose more and more people who just don't want to give the data.
[00:52:41] Speaker A: And you see a new company jumping in on this or do you think Google's going to jump back in the saddle?
[00:52:47] Speaker B: That would be interesting, wouldn't it?
I was kind of expecting Google to jump back into the saddle a few years ago just to hedge their bets and maybe they are, but I haven't seen any evidence of that. I do see lots of interesting companies out there experimenting. I don't know whether how big the market is just now and clearly the results are probably not going to be as good as highly personalized micro targeting. Just by definition, how can it be? But if it's not as good, what's the difference? I suspect, and I've seen some studies and proper academic studies that the difference in terms of, you know, cost per click and that type of thing is actually quite small. But there will be a difference. And that's the arguments that are happening now is are advertisers prepared to pay a little bit more, Are publishers prepared to take a little bit less for displaying ads in order for this ecosystem to continue to work? Yeah, if everybody sticks in their heads in the sand and just goes, la, la, la. Let's just continue. I think we're in real. I really think we're in real difficulty.
[00:53:44] Speaker A: Yeah.
[00:53:44] Speaker B: Especially with the next generation 2008. I think we're kind of heading in that direction. I think the difference is we all know in the industry that's the direction we're heading so we can do something about it. It was a bit unknown with the financial crash or what was going on, but I think we know the problems that we have.
[00:54:01] Speaker A: Interesting. Well, thank you very much, Brian. It's been great picking your brain for the last 50 minutes and yeah, thanks again for joining. And any plans for a new book, by the way, on the last seconds here.
[00:54:12] Speaker B: Sorry, say again, Any, any chance on.
[00:54:14] Speaker A: A new book that you're working on?
[00:54:16] Speaker B: My wife would kill me. I think she. This is not. I'm focusing very much. I mean, I'm the product owner and the product manager for verified data that really does absorb all my spare time. So I don't think there's going to be. But never say never. But I don't think there's going to be a new book.
[00:54:31] Speaker A: And there's. And there's a Dutch saying that your work speaks book chapters. There's a Dutch saying.
So I think your work is. Speaks for you, for everything you stand for and everything that you're trying to achieve in the industry. Thanks again for everything you've done to the community as well. And thank you for being on the show.
[00:54:48] Speaker B: Thanks, Matthew. Take care.
[00:54:50] Speaker A: You too.
