So What? Applying the data-driven customer journey

Well, hey, howdy everyone. Happy Thursday. Welcome to So What the Marketing Analytics and Insights Live Show. I'm Katie, joined by Chris and John. Hey guys.

Hello. This week we are talking about applying the data-driven customer journey. We'll cover what is the data-driven customer journey in 2024, how to apply the data-driven customer journey to your own data, and what actions to take with your analysis. This is a conversation that came about uh from our own reporting. So, like a lot of companies and teams, we have just sort of fallen into the routine of creating the same report and not really taking a step back to think through is this still the right report for us?

Is this still data that we can take action with? So I was saying to Chris that I wasn't sure that the data-driven customer journey analysis that we currently run for Trust Insights was the most useful to me. And he said, Well, what do you want? And I said, that's an excellent question. And so we decided to turn it into a live stream.

So you will be seeing us figure it out live. But I feel like it's a really good exercise because I'm not the only person who uses this data. You know, John uses it for further down the funnel. Chris uses it probably for more of the awareness, because that's a lot of the content he's creating, and I'm trying to look at it as a whole. And I think that we can all agree that the digital customer journey report that we create for ourselves, we've been doing it the same way for pretty much the entire time that Trust Insights has existed.

You know, the underlying technology may have changed, but the output itself is still the same. So I feel like we're due for a refresh. So why not now? Why not now? So for folks who are unfamiliar, this is the customer journey, right?

It's seven phases. Awareness, does anyone know they exist, right? What's the problem that people do people know that even though they have a problem? Consideration is okay, well, what should I be thinking about to solve this problem? Evaluation is who can do this for me?

Uh or how do I do it myself? Purchase or conversion, of course. Like, like I'm I've made a choice. Ownership is okay, what's my experience like uh now that I've I've bought the thing, is my problem solved? Uh loyalty is uh I am so impressed with the the what I've gotten out of this purchase or whatever that uh I'm happy to I'm happy with maybe even go buy more of of whatever it is.

And of course, evangelism is the the uh ultimate expression of that, saying, like, yeah, I am so thrilled with this, I'm gonna tell my friends that this is an awesome product or service or company uh and try and get other people to get involved with this as well. And you'll notice that it's broken down into two sections. So one is the buyer's journey and one is the owner's journey. Because once someone purchases something, they're now they have have ownership of the thing. And so the goal with the owner's journey is to keep them in.

The goal with the buyer's journey is to move them through the three phases to purchase. Um, you know, and I know John and I often talk about our sales funnel, and we try to map the stages of our sales funnel to the customer journey. So awareness, consideration, evaluation, purchase, and then uh at least loyalty. So we don't have it as deep as this customer journey is, but we definitely have the phases mapped uh to our funnel so that we kind of know like what stage of the funnel is working the best and working the least. Exactly.

And so the question uh we have now that we want to try and figure out is what do you need to know, Katie? And as befits the name of the show, so what? What does that data mean? What are we gonna do with that data? Yeah, no, I mean, and it's a it's a great place to start.

And so um the place that we always start is with the five Ps, the purpose, people, process, platform, and performance. Um, and with that creating user stories. And so I was thinking about this the other day, and the problem that we have right now is my user story. And so a user story, as Chris has on the screen, is a three-part sentence. Uh, as a role, that's your persona.

I need to, that's your task, your action, so that that's your outcome. Um, my user story has been as the CEO, I need to know what channels are performing so that I know where we should be spending, you know, our time and resources. And I realized as I was thinking about that, that's not really the question I'm trying to answer. The question I'm really trying to answer is which channels are converting, not performing, but converting to specific, you know, event conversions, so that I know, you know, that we need to do more of it or adjust our strategy for those. And so it's similar in nature, but it's actually answering a different question.

Because if we say emails performing, we know our email channel is performing, that's great. Okay, we'll do more email, but it doesn't tell me enough detail so that I know what pieces of our email are actually performing and what are they doing? Are they driving awareness or are they driving conversions? And I'm missing that level of detail, because I want to be able to make sure that we're putting the right content on the right channels at the right time. Okay.

So as the as the knee jerk uh practitioner who just wants the the manager office case, which is every um, I'm gonna go into Google Analytics. I'm gonna go into the advertising section here, choose conversion paths. Now, the one that we care about a lot is like contact us, right? People, we we want people to contact us. So I'm gonna uncheck everything else except contact us.

I'm gonna change our look back window. Let's go, let's go back a few months that we get like at least at least one quarter's worth of data, maybe one quarter and change. And today is the 15th, so we'll hit the 14th. So one of the things that's interesting about Google Analytics, the the GA4 version is that they actually have a funnel, uh sort of a buyer's journey um map, which is what you're seeing here in conversion paths. Now, this is not the full customer journey.

This is only essentially awareness consideration and evaluation and then conversion, whatever the online conversion is. It is not the rest of the journey. We don't see anything else besides that. So for the contact us form, this is what we see for the Trust Insights website. We see for at the sort of that awareness stage, organic search, which no surprise, right?

That's that makes logical sense. You have email, referral, organic social, organic search, referral, email, organic social for the middle of the funnel, your consideration phase, and then direct organic email, referral organic social for the contact us for the last five months. So from a perspective of what's converting for the contact form, it is organic search, is essentially the the driver to focus on here. Which is interesting because that's not what the digital customer journey report previously had been telling us. That report had been saying email is your best channel.

You gotta go with email. And in this particular analysis, email is in there, but it's not the strongest most effective channel for the specific conversion that I'm looking for. Exactly. Now, if we switch conversions and we switch to something like uh people filling out like our webinar pages and stuff like that. Now you don't imagine email makes sense.

It's it's all email all the time. Right. So for people to advance in a nurturing, which is what you know, John tries to do a lot of. Yeah, John should be sending email all day, every day to keep people moving down the funnel towards um engaging more with us. That is my life.

Email all day every day. I think that pretty much sums it up. Need a shirt. But it's interesting because John and I have different priorities. So for John, getting people to fill out like a webinar form or download a paper or something that falls into our any thank you conversion is a hundred percent his priority because that's where he starts to figure out who's paying attention to us.

Whereas I'm more concerned about who's raising their hand and saying, I need to buy something from you. I need help right now. And obviously the contact us is much farther down the funnel, so there's going to be less of them. But that tells me a very different story than what John's getting in terms of what his priorities are. Right.

So now you have an answer essentially for that contact form. Um organic search. And then it makes sense, direct would be you know, if if people are aware of us that by that point late in the funnel, yeah, they they they will probably come back to us by name. So there wouldn't be a lot of of other channels there. And I would imagine if you scroll down on this report, so right now it's primary channel channel group.

I could switch that to be source medium and get more granular uh data, which is something that I would be interested in knowing. Exactly. You can actually do that here instead of primary channel grouping. You can do your default. So it's Google.

Um interesting. That's Twitter's link shortener up front there. There's me, there's the almost time in the newsletter, there's the almost time of the newsletter separate version. Um there is a URL shortener, there's marketing over coffee in the middle there. And this is for the contact form.

There's Google, the almost time of the newsletter, me, set and so on and so forth. So from a a source perspective, this definitely shows a bit more detail. And I feel like that's what I've been missing in our previous customer journey reports is, you know, I look at this and I still and I think this is still a little too much information. I really and I know that this is going to come out wrong given everything we know about the way that we do attribution, but in some ways I really only care about the last touch, which isn't the only thing contributing to people filling out the form. I'm fully aware of that.

But at the same time, there's so many different ways that people use our assets that I have to focus really in on the majority of how people are getting to us, you know, between Google search and then our email newsletter, or Google search and then your email newsletter. Like I can't, I'm only one person. I can't focus on every single version of the path that people are taking, I guess is my point. Yeah. And you can see the percentages, how much of each stage of the customer journey each section fulfills.

So if you were to focus on last touch, that's still 50%, that's 55% of the credit in this customer journey. And so that's not an unreasonable thing to say. Like, okay, if we have limited resources, limited time, then let's make sure that we're making the most of the part where people are the most active. Um a long-term strategy be like, okay, can we get this to be a little bit more balanced to have more interaction early and mid? But for now, um, if I was in your shoes, I would be saying, okay, we need to to focus on our our Google search optimization, specifically for Google.

Um, and we need to to get this Christopher Penn guy to to put us in his newsletter more. Yeah, he's a slacker though. So I'm not gonna hold my breath for that. Well, what's interesting is that you know, again, sort of in our historical uh digital customer journey reports, we were over indexing on your assets, Chris. But when we break it down in this sense, your assets are driving awareness, which is great, which means we need to rethink the messaging in your newsletter to make sure that it is the introductory awareness drive to trust insights, but then the you know, the journey from your assets to our assets is clear so that we pick up where you leave off.

Yep. So I guess I probably shouldn't do that whole Pokemon issue. I know nothing about Pokemon, so I would be uh very unhelpful in that sense. You're lucky. Yeah.

All right, John, you can handle that one. And the other question that uh is worth asking is again going back to your conversion events when you have the different new uh conversion events that are set up in your GA4, are they the right ones? Uh because if you if you just say all of them, which a lot of people do just sort of by default, right? Yeah, you get a very different look. Or if you say, I want to look at a very specific act set of activities, um, that may not be reflected here.

Uh and there's some things there are some gaps as well here. Like for example, um webinar registration for the webinar we just did is not in here because we can't put our GA4 tracking code on our webinar landing page right now. I think we get have to go in and fix that for the for the next webinar, but right now it's not in here. Um so those that would be another type of conversion that's that's not here. Do we have our course completion as an event?

And is it in here? Uh I think we do. Because that's a very clear, but to your point, you know, we should be auditing our events more regularly because our courses, um, which you can find at TrustInsights.ai/slash academy, um, those are how we educate. And I want to know where people are coming from that they've decided, yes, I am ready to sign up for a course. And so that way we can do more of that.

Like that's the kind of data that I feel like as you know, the stand-in CMO. Uh, because I feel like we kind of share that responsibility. But from my perspective, I feel like we're missing that level of data. I agree. John, what are your thoughts?

What kind of data do you if you said like if I just had this, I could do something more effective? Yeah, one of the challenges because every time that I have to dig in, I usually start backwards with from HubSpot, you know, because I want to go with the conversions. A lot of times when I've gone through the Google data, there's just too many gaps as far as figuring out who's who, especially when we do something that's an enterprise sale where there's seven or eight people from a single company that are, you know, hitting stuff. Um HubSpot does a great job of bundling them up and keeping them under a single reporting structure. So that works well.

Um yeah, I don't know. And we have a bunch of challenges too, because we have, you know, there's a wide array of products, you know. You know, I mean, we courses are easy, and that's why this is great here, but then so much of what we do is custom consulting or other projects that um, you know, it's not like we can just put up a billboard in San Francisco and be done with it for the month. It's a lot more uh hands on and a lot more discussion. Well, and so you mentioned you bring up a good point.

So a lot of teams are looking to their CRM for their customer journey data. And so, you know, as you mentioned, we use HubSpot and we have our Google Analytics data theoretically integrated into HubSpot. Are you finding that to be enough? Does it give you the information you're looking for to figure out where someone came from, what pages drove them? Like, does it give enough customer journey information in that system?

Or are you trying to blend the two systems together? Yeah, we tried to blend, but you know, the good news, you know, unfortunately, yes, every customer gets a unique product, but the upside to that is there's a lot of discussion. So we just have that baked into our process. I mean, it's the classic marketer, you know. The first time we talk to them, I'm always always like, how did you hear from us?

Right. And, you know, and and even then it's still there like, oh yeah, well, I went to the webinar on X, but then you dig deeper and you're like, well, and they've been subscribed to Chris's newsletter for seven years, and you know, they went to podcamp back in 1934 when we had the first one. Um, you know, keep uh keep going back. But um, yeah, the good news is we're able to do a pretty decent job because we have quality discussions. It's not like we're some B2C company where we like have no idea where Billy, you know, bought that last Coca-Cola because it's one of the seven billion that were sold.

Mm-hmm. All right, Chris, you've been pretty busy. What are you working on? So we actually didn't have the GA4 installed in a ThinkIpic instance. Um, it turns out it's it was relatively easy to install.

So now it is. How about the thank you page though that comes back to our site at the end of that cycle? I thought there was a conversion. But this is good now. If we have this for ThinkIffic, we can actually see show me the course finishers and stuff like that.

Yeah, it's going to show all the the engagements and it's going to push a bunch of new events into GA4, which tomorrow we will have to come back and start assigning those events uh and and uh designating which ones are going to be conversions. So um even just something like that is you know, it's a basic governance thing, you know, uh as the expression goes, you know, cobbler's kids have no shoes. Uh so we did not set that up for ourselves. Right. Well, and I think you bring up a good point, Chris is that there's a misunderstanding that systems like Google Analytics and other pieces of software are set it and forget it.

Like you set it up once and it just runs. You need a constant maintenance plan. And that's you know, where I always go back to the five Ps. And specifically in this instance, it would be the process. So uh, you know, of the people who's setting it up, but also who's maintaining it and what is the process for maintaining this particular system.

How often are you doing it, who needs to be involved, what of it needs to be maintained? How do you know that your data quality is good? You know, so we were just casually having a conversation about course completions, and it brought you down this rabbit hole of, oh, and by the way, now this thing. And so, you know, I say a little bit of shame on us for not doing that planning ahead of time because you know, you're not gonna catch everything, but our process should have included once a quarter, you know, auditing our own system. And it even that gets you further to the finish line than not doing it at all.

Oh, yeah, for sure. Uh, I think that that's important. And then also, even in the planning process itself. So when we launched our course uh back in December for uh the for a generative AI course, uh we didn't actually say, let's put the conversions on Thank Giving in here. Which is, you know, and it's interesting because we did put together a whole course launch process, and it's not the first course that we'd launch.

You know, we have our Google Analytics 4 course, we have, as you mentioned, our generative AI course, we also have our Google Search Console course. And somehow that was just overlooked for all of them. 2023 was a rough year. Well, and that's the thing is, you know, I think that what we're showing is like it happens, but then you have to go back and do the work and correct it. Exactly.

So that's that is now what so that's now installed. Um the downside is we can't run any attribution data on what's already happened because nothing's ever retroactive. And that is, you know, and so as we think about like our goals for the company for 2024, that's disappointing to me. I'm disappointed in myself for not asking about it sooner because that's the kind of data when we talk about the digital customer journey and what we need to make decisions on. That is the exact data that I've been looking for, not realizing that it just wasn't being collected at all.

Yep. We're all fired. Get out. Yeah. Well, all right.

Trust Insights is no more. Bye, everyone. Start fresh tomorrow. There we go. Exactly.

Well, I say today is the best day to start with it with any of this stuff. Okay. So all that's set up now. Now what? Well, and that's, you know, when we look, do you have an example of our old digital customer journey report?

And when I say old, I mean literally like last month, you know, last couple of months. Um because I want to give everyone a little bit of context as to what we're talking about. If you've never seen our digital customer journey uh report, um, this is something we do for ourselves, it's something we do for our clients. Um, if you're interested, uh, we can also do it for you at trustinsights.ai slash contact, and you can talk with the uh illustrious John Wall about the options. Yep.

So this is newsletter subscription for December 23. Uh where this is. And so in this report, the challenge, it you know, it's interesting because it's not dissimilar to the data we were just looking at in Google Analytics, but the data represented like this, I look at it and I just kind of sit there and go, I'm not sure what I need to do with this, because I feel like there's just not enough context. So this says almost timely, which is your personal newsletter, Chris, is the number one driver of newsletter subscriptions to the Trust Insights newsletter, followed by Google Organic, um, and so on so forth. And so, you know, I can look at that and go, okay, well, what about almost timely newsletter is driving newsletter subscriptions?

You know, does that hold water? And how can we connect the two a little bit more to make it a more seamless um transition for people? And I just don't know the answers to those questions because what this doesn't tell me is are there particular issues, are there particular topics? And so I feel like the challenge with this, especially a company like ours, where we have a lot of different content assets, is that it doesn't tell me in specifics what's working, it tells me in broad strokes. And that's where in these canned reports, you're not going to get as much detail as you're going to get by slicing and dicing all the stuff within uh the explorations.

So the explorations are really where this has to happen because that's where the data is going to live. So let's go ahead and start an exploration here. So we want to know you want to know what conversions by so the here's here's here's how we have campaign. Yeah, exactly. By campaign.

So let's again, let's go back to say October through February. Get a nice long window here. We're going to look at uh source medium and we're gonna do uh session source medium because we we care about where the traffic came from within a session. Um we want to check in because this is um marketing stuff, we want to use sessions as our metric. Um we also want to have uh events available, conversions available.

Okay, and let's let's let's start with that. Let's drag in uh session source medium as our rows, and let's look at converg conversions, but we want to we forgot event name let's bring in conversion event while we're at it. Um we want our event name here to contain contact us. Right. Because that's what we care about.

So for this time period, we see Google organic almost time. So now we need to bring in campaign. Yep. You know, and this is all with the caveat that we are following our own advice and tagging things correctly. So let's add in our camp session campaign.

And so now we have a bit more granite. Let's add some more rows here. Okay, so we have Google organic. There's no campaign for for those referrals. Um the January 14th issue of the newsletter brought it in a thank you conversion, the 2014, the February 11th, the October 22nd, um, then some inbox insights ones as well.

And so that to me is more useful because then I can go look at those particular issues and find out what the topic was, what drove people to finally raise their hand and say, hey, I need some help. You guys did the thing. Can you help me also do the thing? Right. So the 2024, so the January 14th issue.

What was in that issue? That issue was um disclosures, uh future of generative AI, doing model comparisons, and then of course, all of our usual stuff. And that one had I there's always the promotion for the course. Um the one link that gets a lot of traffic uh from the newsletter every week is the content authenticity statement. So this is a link at the top of the newsletter that bounces over to the site, and this page gets updated uh fairly regularly as laws change.

But this is one of the most frequently trafficked pages on our website. But that we don't know that that is then driving conversions to contact us. Right. So what you'd want to do then is you'd want to start bringing in page data. So let's put in page title and page path inquiry string.

Import that kill off session source medium campaign for right now, and let's just look at page title and see. Oh, that's not helpful. That's not helpful at all. What that I mean, that's that is uh technically true. Um there isn't see this is the challenge with GA4.

In the old version, you had a previous page. I remember, yeah. And you could pass that count along. You can't do that here. Um this is why we build attribution models.

But that's a really good point because we do know that there's limitations with Google Analytics 4. And I know that the questions I'm asking are not unreasonable, and it's the questions that a lot of people are asking because that's the kind of data they want to be able to make decisions on. They aren't able to build these reports in Looker Studio very cleanly. We clearly can't build them in Explore Hub very cleanly. So then what do we do?

How do we get that? How do we make really good data driven decisions with our customer journey if we can't get to the data? So where you would have to go for this at the person level in whatever your marketing automation system is, and extract the audit trail data from there and use that as your attribution data because then you could see you know, you know, on the way to conversion. So here's a list of all your prospects that are converted, and then here's the audit trail of all the pages they visited and the channels it came in from. So you would filter that data down until you were left with here's how many people had this page in their audit trail somewhere that they went then went on to convert.

And that goes back to that's the process that it sounds like what John is following as he's looking at our CRM data at individual people and the Google Analytics data that's included in that. Which, you know, based on our level of subscription to our CRM, is difficult for us to pull out. Yes, and that's I think a lot of people are running into the same is we were so used to being able to get this data from Google Analytics, Universal Analytics, and now it's sort of like the new world order of how do I how do I get this? Do I just forget about Google? Can I just not do my data-driven customer journey?

And I feel like that's really the core question is we just can't do it the way we used to. So, what do we do instead? The way you could get at this if you have the BigQuery data backup install is you can pull the anonymous user ID and look at the pages per session that way. That's a pain in the ass, though. And I think that that's you know worth noting is that you know, people are trying to figure it out and it's requiring more technical skill than it used to.

Yeah, let me see if I can do that. I may not be able to do that. At least I mean, John, what do you make of all of this? Yeah, well, this we've could talked about this a lot. You know, should clients just be dumping everything to BigQuery so they have all this data, and you're not trying to do any of this in GA.

Um it's tough. I don't I have not seen anybody that has like been able to have all their questions answered just going through you know the GA interface, right? You have to do additional stuff and uh find another way to get there, really. And so I feel like this is where the conversation of the single view of the customer then you know enters into the narrative, and where people are trying to figure out is the CDP, which is a customer data platform, the right move for me. What does it look like?

So we have a client who's working on putting together a CDP for that reason. They're such a big enterprise-sized company that they're trying to figure out at a very large scale with a lot of data, what their data-driven customer journey looks like, but they have so much data in so many different places that it feels like an impossible task. Yeah. So what you can see now on screen is the raw Google Analytics data from GA4 stored in BigQuery. And what we're specifically looking for is this here, which is the user pseudo ID.

That's a person ID that Google has assigned. So the way to extract information like the what you're after, which pages led to a conversion, uh, would be to write a big query query against this table asking it to group your traffic by user ID, filter for uh user IDs where the conversion event name was, you know, in our case, uh contact us conversion by page visited in chronological order. And if that seems intimidating, um, you can actually export us you know a small chunk of this data uh as is and ask a tool like Chat GPT to help you write the BigQuery query that you can then paste into BigQuery and have it run. And then could you take that information and push it into some sort of a dashboard? Yes, you could uh you could connect BigQuery to Looker Studio for that.

Um I'm trying to remember though if Looker Studio supports um transactions, transact SQL statements or not. I don't think it does. And so that kind of puts us back to the beginning where I can get at the data, but for me to have to go through and read the results of a big query query is probably not the best use of my time because I'm gonna have a lot of questions and then I'm gonna be, you know, borrowing your time to say, what does this mean? What's this? How do I read this?

And so it's yes, we're starting to get to the data but it's still not really accessible. Right. And and that's one of the reasons why people don't really use Google Analytics for nearly as much as they did universal analytics because the answers are not easy to get out of it. The the data is there. The data exists it is just not in a form that the average non-technical user can get at so I guess the next question is you know what do we do about this?

So if I were a client and I came to you and I had all of these questions what advice would you give to me if I said I need to be making decisions with my data driven customer journey number one how do I build my customer journey in 2024 and number two what decisions can I realistically make if not all of them so you can't make that page level um out of what's in the box. But what you do have access to is that top level stuff right so if we go back if we go back to where we were in the beginning which is uh let's uncheck all this stuff and pay attention to the contact us forms and look at the last you know 90 days or so here we go like this what we arrive at real in a relatively fast uh procedure is Google organic search and we do this because we looked at the source medium Google organic search and and my newsletter the two drivers of the conversions you care about. So the so what is how do you how do you do more with that information? So one of the things that I would start with that's real simple is go into um where do I put Search Console? Go into Google Search Console, right?

So for the same period of time, what pages are getting traffic, right? So let's go to search results. Let's go from October 1. Let's see custom date period time. October 1 apply look at pages.

We have sort by click. So these are the pages that are getting traffic from search. Now the question to ask yourself here is there are some there are pages that are getting clicks but there's they're getting a huge number of impressions and then you got to wonder well why aren't we getting more clicks out of those pages? Like if you do if you look at it on a percentage basis some of these pages are going to give us a lot of traffic some of them are going to give us not a whole lot yeah and I feel like now we're getting you know much deeper into the analysis so you have Search Console data. You could also then go back into Google Analytics, find these same sets of pages, get the session data for those pages and try to fit try to make a more complete story of what's going on.

Yep. But even just this I would say okay let's take that uh thing let's take the oh I have to unzip it thanks Google let's uh download open up file we have our chart now, which should be our largest pages. That it? Yeah, that's it. Okay, let's see.

Let's talk about SEO today. I have some Google search console data here about the top performing pages on my website. In this document are columns for clicks, columns for impressions, plus the page URL. What I'm looking to do is improve the amount of search traffic I get from Google to pages on my site. Pages that get a lot of impressions are pages that Google thinks are relevant for a given search term.

Pages that get a lot of clicks are pages the user agrees are a good fit for their search. What I think we should do is look at click through rate and find pages that get a lot of impressions, but not a lot of clicks, and prioritize them by impressions for a short list of pages that we should try to tune up to better fit user intent. What questions do you have? So that's our our query. That sounds like a solid strategy for improving your website search traffic.

Way to go. I feel like Chat GPT is very supportive. It it it is. When it's not mansplaining. Well.

As a man, do you ever feel mansplained too? Absolutely. It's awkward. It's very metal. It's like, I know this.

You don't have to tell me this. Um do we have any goals or any particular aspects of style tags, meta description, or content update? What is in this that it thinks that it can do that? That I like the train of thought that you're going down. And I know that we did this with a client the other day.

We actually pulled a screenshot of a particular page, put it in here, and asked for recommendations on what we should be doing. So I would imagine once we get our short list of pages that need to be improved, we could also then use this system for very tactical recommendations on what about those pages should be improved. So I'm gonna make you replicate this exactly. I screenshot this right here. All of that is changing.

And so we really need to be thinking, what should I be doing differently? Because the way we've been doing it just may not be working anymore. Mm-hmm. Yep. All right, so it's thinking about its thing.

Um one of the things that I would do with this is let's go ahead and take. So this is arranged by clicks. We've got uh AI copyright law. We've got an old 2019 post here. We've got our a bunch of our updates and things.

So one of the things, if this comes up, let's actually let's see what it comes up with. So what understanding YouTube analytics? Um three clicks, 7,000 impressions, position position 83. That page, that that's not great. Um, so let's take a look.

Will you actually let me see this page? Okay, so this is our page in question that is not great. That it says it's an opportunity. Is it? So that's understanding YouTube analytics, and the other page that you're on is algorithm.

But I mean, regardless, for the sake of the example, it's fine. Let's find out. Hang on for a second here. This is 7184. I think that's not the page title.

So that the data for that is I look at the actual file, it is the tools for competitive intelligence page. So it hallucinated the page name, but it got the URL correct. Huh. So the metrics are right. So this is the page in question that it thinks is is not great.

So okay. So let's look at this page and see what we can do to improve it. For so that's a that page is competitive intelligence for the concept and topic of competitive intelligence. Here's the raw page copy. So we're gonna go to that page.

I'm just gonna literally hit all select all copy. Based on this page copy and the term and focus of competitive intelligence, what is not on this page that should be. Well, of course, it's a transcript. Um tool examples and descriptions. Case studies and examples.

Your point is well taken, Chris, because you stopped and said, of course, it's a transcript. And one of the pieces of advice that we give is if you put your transcript up, it's a good proxy for content. But what this is telling me is our transcript is great, but then we need to pull out even more information and supplement it with more detail. That's really gonna make that page shine. Yep, exactly.

So I would I would say with this, we have an outline essentially, because that's what this thing built is an outline. And so I would take the transcript, feed it to a language model and say, okay, I want you to take this initial data from this transcript, and I want you to reorganize the transcript to adhere to this outline. So that we're still keeping our words and our content, right? So um from copyright perspectives, that makes it a derivative work. Um so we still own our data, but it would it would slot this in into it into out um categories that would make this page probably perform better.

But then we would still have to write more because we don't have necessarily tool examples and case studies and so on and so forth. You know, so we would we could reorder the transcript, but we would still need to add additional detail to it. And that's not what we want generative AI to do for us. Right. Maybe.

That could then be patched into this. Gotcha. That makes sense. And so, you know, we've spent the past 45 minutes trying to figure out how to get to basically a single decision. And so it's useful, it's detailed, but at the same time, it still feels very overwhelming to try to get this page by page by page by page.

Um, you know, so I can look at a high level and see what's performing, what's driving conversions. I can go into Search Console and figure out what pages at that time. They may not be the pages that were driving conversions, but they're in the cluster of them, you know, and so I my next step is I have to decide what's the most important thing to do. Because there's a lot of things, there's no shortage of things to do, but I have to figure out what is the most important thing. And that's really uh the daunting decision that's on my shoulders now.

Exactly. Yeah, appreciate it. Well, uh, so that though I think is an excellent sort of conclusion on the use of both the data-driven customer journey and the use of generative AI. So tools can do a lot of the work. Like we did a lot of data processing in a relatively short amount of time.

Right. Human judgment is still needed to so like you're you still have a job, right? You're still employable. Trust insights is back in business. Because the machine can't do it all.

I mean, it could, but it wouldn't be optimal because it doesn't have the full context that it needs to understand why it's making the decisions it's making. We can give very tactical things to do, like hey, rewrite this transcript, reorder this transcript to fit these parameters, but it doesn't replace you saying, yeah, that's an important page. Uh there is stuff like this one here, the uh our old press release page from overuse of words from 2019. From a mathematical perspective, that's the fifth most important page on our site. From a are we do we have products and services as trust insights to sell for this?

No. Right. Exactly. Your judgment would be no, let's not focus on that page, because even though mathematically it's important, it's conceptually not. Well, and we saw in that very quick example that generative AI hallucinated the title of the page.

And so you still need that human judgment to make sure that the data it's giving you back is actually correct. Yes, exactly. So those are your th that's that's that's the value that you still provide as a human being is going, okay. And I'm going to tell you whether or not that's actually a priority. John, what are you going to do with all this?

Well, yeah, even when you get to the bottom of the barrel of it, it's still the question of like, okay, does that mean we need to do more of this or we need to do something else? You know, you you never know at what stage of growth you're at with any of those campaigns. So it, you know, you kind of take all the data in front of you and then pick your best shot. I mean, that's it. It doesn't really matter.

You know, and it's funny to sometimes you find out later on that, oh yeah, actually that data was all screwed up, but at least we did something. That's really all you can do. Yeah. For in John's case, for example, with marketing over coffee, I look at the interviews on you know, pages on the site that get the highest amount of traffic for those interviews. And if there's not a transcript there, the bare minimum is get a transcript on that page and then use generative AI to create a summary that's more appealing that maybe covers the highlights that supplements the show notes.

Because the show notes are just uh a short bullet point list, but they're they don't contain the content of the show. So for those super high performing pages like what we just did, I would repeat that exercise with generative AI. Makes sense. More to do. Hooray.

More to do. Always more to do. Yeah. All right, folks, that's gonna do it for this week's. I guess we will see you next time.

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 T I podcast at our 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.