So What? Dark Data Mining

Well, hey, happy Thursday, happy April 1st. We are not pulling any pranks today. Um, you know, quite honestly, we just don't have the energy. So welcome to So What The Marketing Analytics and Insights live show. Uh today we're talking about dark data.

What is it? Where do you find it within your own systems and what to do about it? And so, you know, I remember when we first launched Trust Insights, we were we talked a lot about dark data. And what we realized very quickly was that people didn't fully understand what it meant. And so I do remember I was uh either doing an interview or a podcast, and Chris, you commented afterwards that the example that the interviewer gave was incorrect.

And so a very it's um, you know, I certainly don't want to call anybody out, but basically, you know, the way that we think about dark data is it's data that's collected that is not then used. And I think one of the examples that we give a lot is, you know, fitness data. If you have a fitness tracker and you're constantly tracking your steps and your heartbeat and you know, all of these things, but you never look at the information, you never do anything with it, you never use it to make a decision to make a change. That then is by definition dark data, not because it's something you can't find, but because it's something that literally just sits there on a shelf in the dark, never getting used. So Chris, specifically dark data.

It's all a little bit. Um, this is something that pretty much everybody has. It is something that uh pretty much everybody does nothing about, and it impacts our ability to actually make decisions uh to know what's going on. So let's actually let's do this. Let's go into our friend Google Analytics here.

We have the uh trust insights account here. This is Google Analytics 3. And I'm gonna go into acquisition. I'm gonna go into all traffic and just do source medium. Now, source medium is pretty straightforward where the traffic comes from, so what general channel is it?

And one of the things that you'll notice in this is true of everyone's web analytics, is you'll have this direct none category here, right? This is literally dark web traffic. We don't know what it is. There, this is traffic that came in that has no attribution whatsoever, so we can't tell where it came from. We can't tell if it was uh from an ad we were running, maybe if it was somebody literally just typing in trust insights.ai, which if you were that person, thank you.

Um but we don't know. And what we need to know is given that this is the number three traffic source on our website. Is there a problem? Like, do we have a dark data problem? I'm gonna just flip over really quickly into uh bark graph mode here, and I'm gonna extend this out.

Let's look at the last month. Actually, let's look at the whole quarter, right? It's quarter end. Um, let's see what happened in in Q1. For us, 10% of our traffic is is missing, right?

We have no attribution on and I don't know, Katie. As the person making decisions, is that a problem for you? It is. Um, you know, 10%. When you think about it, you're like, oh, well, it's only 10%.

But, you know, especially where we have to be really strategic and tight about our budget, I want to know exactly where people are finding us so that I know exactly where I should be putting our time and resources and effort. And, you know, we've seen it upwards of 70% for some companies, you know, which is a bigger problem, but 10%, it's not insignificant when you look at the whole thing. You know, if you drop that 10% out, then what are you losing? And so I think that, you know, my first question to you, Chris, is you know, yes, I want to know what that actually is. You know, why does it come in that way?

Why can't Google Analytics figure out what it is? Aren't they Google? Aren't they super smart? It's their system, you know, and I think that these are all the questions that come to mind for me. I'm sure they come to mind for other marketers.

Um, you know, and I know that sort of the stock answer is, well, Google can't figure it out. And there's gotta be something we can do. But I absolutely want to know is there a way to reduce that number over time. Exactly. And so I think it's really worth digging into that point of what causes this.

So direct traffic comes from a few different things. One, it can come in from people typing in your URL. Two, it can come in from sources that have no attribution. So, for example, when you click on an email uh link in your email, if you've got no tracking codes in your email, it's gonna come in as direct none. It just simply will have no attribution because the email, your your like your Microsoft Outlook uh on your desktop is recognized, is not recognized as a source of traffic.

Um for systems that are encrypted, like Apple Safari browser on iOS, um, blocks an awful lot of tracking data. Um so if you're you've got a visitor coming from an encrypted system, it will strip off some of that attribution data unless it's explicitly encoded with uh UTM tracking codes. And something we discovered this week, actually, um, if even one of Google's uh stock tracking codes is malformed, you fat-fingered the the the word, all of it breaks. We tested this, Katie, on your website. We tested it on uh and and we said, okay, UTM source equals whatever, and then just you know, we we mangled the little ampersand and it came in as direct nuts.

Like, oh, that's interesting. So Google D didn't even attempt to preserve the working parts, it just threw the whole thing out. So that's where that's kind of what causes this. Now, our first step obviously is to figure out if we have a problem. In this case, number four, traffic source, I would agree is kind of a problem.

We we want to know it. So now we have to kind of figure out what is the nature of the problem. And the easiest way to do this is to think about where is this traffic going? Like if it was going to say one of our you know purchase pages, like uh that's really not good. On the other hand, it's just going like blog posts, like, okay, fine, I can live with it, you know, uh unattributed traffic going to the blog.

So how do we do this? Go ahead, Katie. I was gonna say, well, I slightly disagree with you because one of the things that we do as part of our own health check analysis is what we call the most valuable pages report, which is an attribution analysis based on the content from our own website. And so you're saying, you know, just as an example, it's not as important if we don't know the traffic going to the blog. However, if we see in our most valuable pages report that the blog is a big driver of conversion traffic, then yeah, I want to know what that is.

So I guess I disagree with you that it's okay, we can live with it because you know our blog, when you look at our most valuable pages report, is what number six, number seven, in terms of the top driver of conversions for us. So therefore, I still want to know what it is. You can't you can't get out of this one, Chris Penn. All right. So let's go and figure out where the traffic goes.

I'm gonna pull up Google Data Studio here. I'm gonna switch over to our edit mode. I'm gonna bring up a brand new scratch page, a place we can have just a little bit of fun. Let's go ahead and clear everything is here. I'm gonna add in a new table, and we'll put in some colors, some heat maps and colors for fun.

Okay. Now, the first thing we want to do is uh we have I want users, want to know how many people. So I'm gonna swap out new users with all users, and we have the page. I actually I personally prefer the page uh the page URL itself as opposed to the title, because sometimes if you're careless, you have overlapping titles, and so we see this is our our traffic. Now the last thing I want to do is I want to add in our source medium, and I'm gonna actually swap that here and put source medium first to page second.

Okay. Now there's a lot of different sources here, so I want to filter this by direct none. Yeah, direct only. I want to include a source medium, which equals direct. Now this is going to tell us when they come into our website from this mysterious spot.

This is where they go. Yes, this is where they go. Because again, if we're sending we're getting a lot of direct traffic to like major conversion pages, I would view that as kind of an emergency, right? That's something we need to fix sooner rather than later because if we don't fix it, then we could be, for example, spending money on ad campaigns that and our ads are broken. That would really suck.

Um so in this case, we have oh that's a little hard to straighten out our gifted data. There we go. Okay. So we have the majority of it goes to the home page by a substantial margin, right? The blog contact us, and then somebody came back, the bookmarked uh that's a newsletter subscription page.

Uh then we have uh a lead gen thing there. So there's not a ton of stuff on here that's like, oh my goodness, we're we're in a whole lot of trouble. Right? It's it's mostly just the home page. So I actually feel pretty good about this.

That's uh people typing in, you know, if it when the direct traffic goes to the homepage, it literally is people just typing in the URL. So that that's not a bad thing. It's not a bad thing. Um and so I guess then the next question is let's say you know it was going to a services page, or let's say it was going to the contact form, which actually you know is number three on this list for I'm guessing what the past 30 days or so, which is the default date range. Um year to date.

I would want to know where that was coming from. So how how do you start to resolve that in your own data? Because I know that we've seen again, we've worked with clients where the first time we get into their Google Analytics, it's upwards of 70% of their data is on attributable, but we've been able to reduce that number down through you know the work that we've been doing down to like 35%, which still sounds really, really high, but compared to 70%, it's now less than half. Right, exactly. So the next thing to do is to try and figure out does the direct traffic resemble something else, right?

Is there a a channel that uh it looks like that you could say, okay, I can kind of see like this this has a uh a correlation of sorts um to another channel because there are some channels like social media, for example, there's dark social where uh you don't know where the attribution is coming from. For example, when you use Slack, if you're in our analytics for marketers, Slack group at TrustInsights.ai/slash analytics for marketers. If I paste a link in there and I don't put tracking codes on it, it's gonna come in as direct because Slack is not a browser, right? And so it there is no attribution. So that would be an example of of that.

Um, however, those interactions may look like say Twitter or Facebook or another social channel where there is attribution data. So we wanted to try and figure out is there a relationship that it looks like. Does our dark traffic look like search traffic, for example? Um if people are searching on their iPhones, uh we we're not gonna get that data. So one of the things the first things we're gonna want to do is we're actually going to want to go into our audience data here and we're going to want to look at our device or our our uh devices uh and technology here again we there are some brands of technology based on browser that we know um obscure data in this case our our from our mobile traffic which is 26% of our site uh of that half about 13% is Apple right so okay huh Apple devices would not be sending us uh tracking data unless we explicitly encoded our URLs everywhere with UTM codes um so somebody coming to us from a a Google search inside of the Safari browser on an Apple device would not come in would they come in as direct none right so that would be dark an example of dark search so we've already ascertained that they're we're we're gonna have some interference from mobile devices for sure.

Our next step is to figure out again let's do that just first version let's just eyeball it just take a look here let's do uh so Chris while you're pulling that up a question came up while you were putting together that simple dashboard um you know what about looking at the path for those direct to the home page ones might give you more of a clue if there's a pattern what do you think about that approach one of the things that um we do as a filtering mechanism to look at other direct traffic is actually screen out the home page um to because obviously if someone is just typing in the URL there we we we are pretty sure that's actual direct traffic. Um so one of the things that you can do is if you want to, you can say I instead of um page, you can say you can specify. Let's go back here and look at landing page. So let's do that. And landing page would be the first step in the journey.

Put that in here. And go back to view. So in this example, what is the difference between page and landing page? Page is that page was visited uh by that user. Landing page was that was the first page that they went to.

Um and what we're seeing here is for example, there's there's direct traffic to contact page, right? But the first page in that particular journey was uh was just the home page, the uh stuff. So um if you wanted to to like real let's get rid of page entirely. Yeah, I think in this example, landing page is more helpful. Yeah.

So now we see we have 201 uh visits went straight to the home page as their first very first stop. Second was unknown. Right. We literally have no idea what happened there. Um that by the way would be from someone using a very uh highly secure browser, like there's uh certain versions of Firefox that uh block everything.

Uh there's certain ad blockers, uh like Ghostry, for example, that can even prohibit the loading of most tracking codes. We have uh resources, blog, and so on, so on and so forth. Contact page drop center six. So this is this is looking pretty good. Okay.

So we have our all users. Uh we're actually gonna turn that off. We're gonna take our direct traffic. Now let's put in our uh organic search traffic. This is based on your uh hypothesis.

Yes, and we're gonna put in email traffic. Uh now the the reason that you're doing that, Chris, for us specifically, is because we know the majority of the channels that we're using are organic search, organic social, and email. We're not currently at this time running ads. Um, so we don't need to factor in paid search and paid social. That's correct.

Now, if you were a company that was doing all of those things, you would want to make sure that you have those segments created and then use them in uh this kind of an exercise. But for us, we you know, for lack of resources, aren't doing more than you know, the uh organic exactly. So now when we're looking at our direct traffic, which is sort of the orange line in there, again, it's difficult to see. Um, does it look like does is there a relationship uh with any of the other lines? It's difficult to say in this case.

Uh it's difficult to say because you really can't eyeball correlations, and you probably shouldn't. Um, but initially, uh initially doesn't appear to be anything like strongly obvious. So that brings us to a more complex version of this analysis where instead of something very simple, we'd actually want to look at every source medium, right? Because to your point, Katie, for a larger organization, they might have 20 or 30 or 40 different uh campaigns and sources going on at the same time. They might have paid email, paid social, uh SEM running display ads, and all that stuff means that we need to instead of you know doing a basic version of this, we need to move to an or more advanced version.

So I've gone ahead uh and run uh we actually have a piece of software we wrote that does this that says okay, take your direct traffic and show us the correlations um between direct traffic and existing known channels. And in this case, for us, this is for our website, uh, this does not apply to anything else. Um Google Organic has the strongest correlation to our our dark uh web traffic, which suggests that our initial analysis of Apple devices being about 13% of our overall uh uh traffic um and dark search being the thing, right? We also have in the headlights, uh, which is there's a 0.38 correlation, it's not as strong. And then after that, we start going into essentially no statistical relationship.

Once you're below 0.25, everything to the left of here is is pretty much invalid. So you have you know some stuff from uh talk walker email, some Facebook uh stuff, there's a slightly stronger relationship, slightly stronger relationship with my email newsletter, uh with the company email newsletter and Google Organic. So we now have a sense of okay, this is probably what is our our our dark web traffic. We've decomposed it to this point. The next question is can we solve it?

Right? Can we solve any of this? Um the first thing before we go too far, Chris. I think one of the things that I just want to sort of like go back to is so this is software that we created. This is our code.

Yes. No other marketing agency has this. And so if you're looking at your data and saying, I have a bunch of direct none, this is something that you know we uniquely do. Is there a way, is there a version of this that other marketers can do that doesn't require us handing over our proprietary code? Yes.

Um, if you go into Google Analytics and you export uh this version that we were doing here, hit export into Google Sheets, it will spit this out. Uh you're sort of a nice um spreadsheet. Um you then have to break out column C into individual columns for each of those four channels, and then you can um once you've done that, um, then you can run a correlation right inside of the spreadsheet software of your choice to be able to do column by column uh uh correlation. Okay. I think that's helpful because again, not everyone has access to something like R or has the skill set to run something like R and write their own code.

So as long as there are other alternatives to do that, and remember, John, I saw you write this down. Do not eyeball correlations. That's all you just you're like, oh crap, I gotta remember not to do that. Yeah, that was the uh the economist in me twitching uncontrollable when I heard that. Um we've ascertained through basic correlation analysis that organic search is probably the number one thing, and then number two and number three being uh our email newsletter and then Facebook having the strongest correlations.

So in the spirit of the the name of the show, So What, uh, what we have to do is we have to start trying to get rid of as many things on this list as possible. The first thing, and probably the I would say the easiest thing to do would be to look at these different sources and say, do you have control over it? Right. So we have control over our email newsletter. So the first thing I would do if you know in this particular instance is fire up um our email newsletter software and just go in and double check, like, hey, um, for our newsletter, are there any uh are there anything in here where we have links to our website that don't have UTM tracking codes, right?

So here's uh one of our calls to action, it's got the UTM tracking codes, they're spelled correctly, uh things like that. Uh what else we got here? We've got the the home page. Now, if I were to go back earlier in the quarter, uh-oh. Earlier in the quarter, I had the headline banner did not have UTM tracking codes on it, right?

So you are so fired, Chris. So fired. All right, I'll sit back in the but I think that this is this is actually a really useful pro tip. This, I'm guessing, a big assumption, is a very, very common piece of the puzzle that's overlooked because you know, one of the things we say is you don't need to UTM your own website. Well, this isn't actually the website, this is an asset outside of the website that redirects to your website.

So you do want to UTM tag it, even though it is your own URL. It's a it's a subtle difference, but it's an important one. Exactly right. So, yes, this is your email. And here we have a case where that was not tracked, right?

So even though we use a piece of software called MODIC, that it um auto-appends UTM tracking codes, it's not guaranteed. It's you don't know for sure whether it correctly did it or whether it was correctly interpreted. Um if you put them in yourself manually and you you double check your work, you're oh okay, yep, I made sure that I've I've got those tracking codes in place. The same is true for social media posts, the same is true for any link that you share anywhere outside of your website. Stick those tracking codes on and make sure that they're using the conventions that are approved by Google, which brings us to the next point.

Um Google has a list of source mediums that they strongly recommend. Um let's put this in. So it's funny that you say strongly recommend. It reminds me of um one of my friends has a teenage daughter that she's been, you know, now I guess you just text with your kids, you don't actually talk to them. And so she's like, you know, I strongly recommend that you come downstairs and do the dishes, and her daughter will respond, I know that's not a question, you're actually telling me to do it.

And so it reminds me of the statement that you just made where you're saying Google strongly recommends these source mediums, but I'm guessing it's less of a question and recommendation, more of a do-it-this way. If you want your stuff to work, then yes. Um Google has this nice article and support that says these are the default channel definitions. So when they look at how um to allocate uh tracking and stuff and give credit without a whole lot of extra work, they actually specify exactly email medium exactly matches email. And by the way, that's all lowercase.

So if you had a capital E for email, it would not associate that with that channel grouping and stuff. So our recommendation definitely is uh make sure that you are adhering to Google's definitions in any of the UTM tracking codes you're using that's that's pretty straightforward. So the next thing that we recommend doing is when you are doing all this tracking uh make sure that you have the that you're keeping it someplace you know sensible um so that it's I guess governed I mean you know Katie you can talk about some of the clients we've had that don't have that level of governance. Well and that's just it so you know we jokeslash not joke that governance is this awful scary word but really what it just means is having you know at least in this context some kind of a repeatable process and so the easiest way to make sure that there is compliance um with correct UTM tracking is to put it in a spreadsheet and what we've done um and what we make available to our clients is an automated spreadsheet I put this sort of lightly but basically it's just a spreadsheet that auto creates your UTM URL with the information that you put in um you know and so you can restrict it down to only the source mediums that your agency that your company should be using in a drop down so you can see in column F, um it's a dropdown menu. And so we don't give people ourselves our clients the option to put in things that aren't on Google's approved list if they're using Google Analytics.

And that's one way to really keep better control over what those UTM uh URLs start to look like. Um you know the source, there's a little more flexibility there because you're going to have, you know, we have like marketing profs and so what? And so things that are unique to us versus just your straight Facebook, Twitter, Instagram, email, newsletter, like you want to have a little bit more flexibility there. But again, really make sure that you're thinking about it in a structured way. Exactly.

So in terms of getting bad data out of Google Analytics, having those ETM tracking codes really is the easiest way to start slimming down what that big pile of direct traffic. And again, if your site has you know 15, 20, 25, 30, 50, 70, 80% direct traffic, at that point you start running into real statistical problems like can you even rely on the decisions that are being made? So if you look at the steps we've taken today, which is number one, figure out is there a problem, right? Number two, uh look at the magnitude of the impact of the problem. If it's going to key pages on your website the this traffic then you need to do something about it.

Number three, do the analysis to figure out where the source of the problem could be coming from. And then number four implement at least what you have control over to to slim down the problem. You'll make inroads to it like Katie was saying one of our clients uh had 70% of the traffic was direct. Like at that point there's absolutely no way to to even think their analytics were reliable as we start cutting the app down. Even just this week, we identified yet another set of problems that they had um that will probably chop down an additional 10 or 11% of that of that 30 ish percent remaining.

Oh, good question here from our listeners. How would you ensure encourage tracking links outside of markers or typical content creators? Again, to what Katie was saying, um a lot of tools do help enforce this already. Uh if you're using like you know, a Salesforce CRM, you can actually build templates and stuff in in Salesforce and HubSpot, etc. That have uh pre-approved content.

Um one of the systems that I particularly like is having your own URL shortener. And having your own URL shortener means that you can't um change the tracking codes. So we use a system uh that's an open source system called uh URLs, um, because your UR your own URL shortener. And all of our stuff goes into this. As it goes in, it's auto UTM track tagged.

So no matter what the link is, it goes in here. And then what you get out of it is you get a short link. Uh so if you were to go into say uh the Trust Insights newsletter, I'm gonna scroll down here, and you look at the news, right? These are the all the links we share, every single one of them is a shortened URL. This shortened URL, you can't rewrite it.

You can't break the UTM tracking codes that are embedded in it. Uh, in fact, um, not only are these uh tracking codes have you Google Analytics codes in them, they also have you can see in our shortener stack adapt codes. So as you click on these URLs, we can track who clicks on them and then show them retargeting ads later, which is kind of a fun trick for another show. Um well, and I think that this is a good point, Chris, because we haven't even covered you take all of this painstaking time to put together the UTM tracking codes, and then some websites, some systems, some platforms strip them out all together and change them. You know, calling you out Facebook.

We know that you do that, so knock it off. And so I think that that's one of the things that, you know, even as you're building in, you need to sort of make sure that you're protecting all of that data that you're so carefully building into your URLs. Um, you know, to go back to this question about how do you ensure tracking links outside of typical content creators, you know, that's where the governance structure comes in. And so, you know, what we often see is that the larger the company, the more siloed it is, and the more disconnected, even if you have like one central digital marketing team, you might have a bunch of other business lines doing their own version of digital marketing. And so it's really needs to be a collaborative effort and have like one or two people sort of overseeing and QAing.

It's not policing, it's really just QAing, making sure that everybody is getting the right credit for all of their really hard work that they're doing by ensuring that you have the right tracking codes on all of the URLs that are going out. You know, if you have an email team, if you have a social team, if you have outside, you know, partners, agencies that are doing things on your behalf, making sure that that is part of the onboarding conversation and that it's something that's revisited probably quarterly to make sure that these things are coming in correctly because they want to retain their contract with you. So they want to probably make sure that their stuff is getting the right credit in your data as well. That's really is the key is attribution modeling. That is shared publicly is the easiest way to get people to self-enforce, right?

People want to say, yeah, my thing did awesome this quarter. Like I should get I, you know, my budget should get more money because my thing did awesome. Uh, you know, if you pull up one of our our standard attribution models, you know, the my my newsletter did 37% of uh of the conversions for for Q1. So Katie, you can fire me, but I'm taking my newsletter with me. Dang it.

No, wait. I know what the contract says. You owe me 50% of your newsletter. Um But when you think about like ad partners, um, co-marketing agreements and stuff like that, uh you show this model to people and you say, okay, this is what we're using to make decisions on for the common quarter. What's gonna get budget?

What's gonna get resources? Um, who's gonna who who got credit? What should we do more of? And as soon as all the parties involved see this, they go, Oh heck, I need to improve my tracking because I want more budget. I want you know, I want a bigger seat at the table, uh, kind of thing.

Um, you know, we were doing one of these attribution models for one of our clients, there was a practically a knife fight um you know between three different teams because they were saying, No, no, no, we deserve credit. You know, we and and there was a whole big thing about like, oh, we didn't have tracking codes on our stuff. Uh-huh. Wouldn't you know it? Next month's meeting, they had tracking codes on everything, and and the contribution that they brought to marketing went from like two percent to 37%.

And suddenly they got religion. Like, yeah, my tracking code's going, I look better in front of the boss. So, you know, to that question, how do you encourage tracking links? Like it's pretty easy. You have us almost like it's a scoreboard of sorts.

Yeah, that's where your attribution model kind of is the score. Like, who's gonna get the biggest score this quarter? Now, John, you do a newsletter for marketing over coffee, and you have you know, links that go outside of the, you know, outside of your website, you have sponsors. Um, how often are people you're like providing you that information up front, you know, when you have sponsors, like they want to know where their stuff goes and how well you know sponsoring with marketing over coffee converts. What does that look like for you?

Yeah, we spend a lot of time just encouraging sponsors to you know give us a trackable link because it's just like you said, it's positive reinforcement. We want them to actually at the end of the month go back and say, oh my god, this newsletter drove traffic for us and was for real. Um and then the only other thing we see is you know, there are brands that once they get big enough, they already have the religion and they just say, Yeah, no, we're just gonna earmark, you know, X thousand dollars a month, and that'll just go straight to the home page and we don't care. But you know, every other business that's not to the point where you just always set aside you know 200,000 for general branding wants to know where their money's going. And so yeah, you really need to do that.

And is it like you can't undersell what you're just talking about, the positive reinforcement, you know, because with our spreadsheet where if you don't use a UTM code, somebody comes by to break your finger, like that's not really motivation. You know, you you want to have uh get positive results and get people on board with that because yeah, going around every month trying to beat UTM codes into people is just horrible all around. It is. And, and you know, it's something that, you know, we've seen again, sort of, you know, big companies, small companies, it's that extra step that, oh, I don't have to worry about that. It comes in to my Google Analytics correctly.

So, Chris, one of one of the other things we haven't really talked about is the actual infrastructure setup of Google Analytics, making sure that if you don't have a UTM tracking code, that at least your channel groupings are set up correctly in Google Analytics so that you kind of give yourself a little bit more of that advantage. I think I want to save that for another show because some things are really rapidly changing on the Google Analytics infrastructure front to the point where what will be best practice going forward in the next few months is a big departure. It's a big, big departure from the way we've traditionally done it. Uh, we were talking this week on uh this week's Marketing Over Coffee episode about some of those infrastructure changes, but the big one is server-side tagging. As systems like Apple crank up their privacy uh to the point where some things like your Google Analytics tracking code may not necessarily work correctly.

Um as browsers end their support for third-party cookies, uh, server-side tagging is going to become more important where no matter what the device is, no matter how private it is, at this at the end of the day, there's a person consuming resources on your web server, and those server logs can then be used as a base for you know uh improving tracking uh to the point where you may even be able to recover more of that direct traffic because you may be able to see just from the raw data on on the hardware that you own or rent um what that is. But it's more complicated than then we've been seeing in the past. In the old days, you copy your tracking code out of Google Analytics paste it on your website and you are done. Now it's you have to install a a new server on Google cloud and you know build a separate new container and tag manager there's a lot a lot to it so I think that would be a good more advanced show down the road once uh you know I think once we have a way to explain it that isn't literally four hours of techno babble I would agree with that because even you were just describing it now I'm like wait did I just black out where am I what's going on I mean I was totally paying attention I swear um no but it is it's very it's very technical and it can be overwhelming and so I think that we still want to sort it out a bit more before we try to re-explain it to people so you know we're focusing on the things that are knowable right now and that's within Google Analytics 3 making sure that you are setting up uh UTM tracking codes correctly now if you don't have a spreadsheet if you don't have one set up you can go to the free UTM builder as well so Google provides one of those they give you the same examples that they give on their support page um and you can start to build it through that free UTM builder. I believe Chris is going to pull it up now and it will give you the opportunity to also use um a short code.

So you put in your website URL you put in your campaign source it doesn't have to be a paid campaign it can be organic. So they give you the examples the one thing to note is they use lowercase for everything. If Google's giving you an example follow their example and do it the way that they're doing it. And so you can have up to those five categories. So source, medium, campaign name, term, and content.

Source and medium are the two things that we're really strict about. The other things are really sort of up to you. And then you can either copy the URL as is.ly. You can use the one supported now. I'm sorry, say that again.

Goo bitly's the only one supported now. Google has end of life to their URL shortener. Okay. Oh, it says authorization required. Okay.

You know, you can build your own URL shorter. There's a lot of different options. We do recommend shortening the link because again, it preserves those UTM tracking codes and doesn't give other services like Facebook to strip out the UTM tracking codes. So you can um build your you can build them one by one. It's not the most efficient.

We do recommend going the more automated spreadsheet route. Um, you know, and if you have other ways of doing it, if it's built into your uh systems like your social sharing system, I know that um, you know, you have like buffer and Agora Pulse and Sprout Social, a lot of those systems have that kind of thing built in. A lot of the email marketing tools have some level of UTM tracking built in. So definitely check those settings. Um, we were actually just talking with one of our partners about that.

Um so there's a lot of different ways to ensure that you get that tracking set up uh correctly. Yep. So in short, that's what dark traffic is, which is a very tiny example of dark data. Um it's stuff that you don't know what's in the box, so you can't use it to make decisions. And there are ways to decode it to some degree.

Uh, there are ways to mitigate it to some degree, and it will never be perfect. Uh, no matter what, there will always be some amount of data that you you can't make use of. But at least you're going from 70% to 7%, hopefully, and making making good use of it. So that's good. Well, before we wrap up, I do actually just want to acknowledge um this comment.

So, Chip, uh, you were saying all these tools are great, but they're adding friction to the process. I disagree with you. I say that the process needs to be revised to add these tools in because if you're not using them, then you're not collecting your data correctly. So I don't think they add friction to the process. I think your process is broken if you're not doing it.

And on that note, uh if you have comments or questions, uh let us up on our Slack group things. But until next week, we'll talk to you. Uh well, just enjoy the replay. Uh let's go ahead and and head on out of here. Thanks for watching today.

Be sure to subscribe to our show wherever you're watching it. For more resources and to learn more, check out the Trust Insights Podcast at TrustInsights.ai slash TI podcast and a weekly email newsletter at trustinsights.ai slash newsletter. Got questions about what you saw in today's episode? Join our free analytics for marketers Slack group at TrustInsights.ai slash analytics for marketers. See you next time.