Latest / The Same Side: A Digital Marketing Podcast Rooted in Human Connection / Episode 7: Turning Data Into Insights

Transcript

Episode 7: Turning Data Into Insights · The Same Side: A Digital Marketing Podcast Rooted in Human Connection

• 0:17
  SPEAKER_00: Hey there.
• 0:18
  SPEAKER_00: Welcome to the same side, a digital marketing podcast rooted in human connection.
• 0:23
  SPEAKER_00: I'm your host, Jonathan Dans, Director of AdOps and Analytics at Digital Relativity, a full service marketing agency.
• 0:31
  SPEAKER_00: Today we're going to do a high-level look at turning data into insights.
• 0:36
  SPEAKER_00: I'd like to revisit something I've touched on a few times already.
• 0:40
  SPEAKER_00: You know, how many meetings have you been in where someone reads through a campaign or website performance dashboard citing number after number?
• 0:48
  SPEAKER_00: Maybe some month over month increases or decreases, probably increases because everyone loves good news, including me.
• 0:56
  SPEAKER_00: Maybe they'll finish up with any questions, like David S.
• 1:01
  SPEAKER_00: Pumpkins, and then it's over.
• 1:03
  SPEAKER_00: No questions asked.
• 1:05
  SPEAKER_00: And I think that's because no one is sure what they're supposed to do with that information, with just these numbers.
• 1:11
  SPEAKER_00: I mean, maybe occasionally someone will ask if those numbers are good, and I'll take that over no questions, but ideally we've given them more context to go on, the kind of information that really gets the wheels turning, that causes the client's mind to zoom out to the bigger picture or home in on something that needs adjusting.
• 1:31
  SPEAKER_00: Because again, numbers without context, they're just that.
• 1:34
  SPEAKER_00: They're numbers, they're symbols.
• 1:36
  SPEAKER_00: They don't quote unquote tell us anything, they just are.
• 1:40
  SPEAKER_00: It's all very zen, just like the report or dashboard that is just there and just sinks into the digital wasteland of your download folder, or disappears somewhere in the cloud.
• 1:53
  SPEAKER_00: That's all well and good for the numbers and dashboards, but as someone in that meeting, you're probably left wanting more.
• 2:00
  SPEAKER_00: And if you're not left wanting more, it's probably a good indicator that you could use a GA4 or Google Analytics reframe.
• 2:08
  SPEAKER_00: So let's talk about this for a second.
• 2:12
  SPEAKER_00: You know, what is GA4 for?
• 2:15
  SPEAKER_00: A couple of episodes ago, I likened GA4 to a video camera.
• 2:18
  SPEAKER_00: It can only show patterns, what actions were taken, but it can't show why people decided to do what they did.
• 2:24
  SPEAKER_00: And that still holds true.
• 2:26
  SPEAKER_00: So GA4 collects data based on what happens on your website, right?
• 2:30
  SPEAKER_00: And much like I talked about AI in a previous episode, it's not the final destination.
• 2:36
  SPEAKER_00: It's it's just a tool.
• 2:38
  SPEAKER_00: Its value lies in how you use the information in the platform to guide your decision making.
• 2:45
  SPEAKER_00: And just like using AI, your experience and your understanding of your industry and your skill set, those are things that you layer on top of your website performance observations.
• 2:57
  SPEAKER_00: So again, it's context that brings meaning to all the numbers in GA4.
• 3:02
  SPEAKER_00: And as a little sidebar, there are even fancy AI insights that you can consult in GA4.
• 3:09
  SPEAKER_00: But there's of course a caveat, just like there's with the numbers in GA4.
• 3:15
  SPEAKER_00: These insights are no different than the descriptive analytics from that fictional yet all too real meeting that I described in the introduction.
• 3:23
  SPEAKER_00: So check it out.
• 3:25
  SPEAKER_00: I'm looking at one of our partner websites.
• 3:27
  SPEAKER_00: It's a higher education website, and there's this little symbol in the upper right-hand corner that looks sort of magic wandy.
• 3:35
  SPEAKER_00: I'm not sure what it's supposed to be.
• 3:36
  SPEAKER_00: It doesn't matter, but if you hover over it, it'll say insights.
• 3:38
  SPEAKER_00: And if you click on that, it opens up a sidebar.
• 3:42
  SPEAKER_00: And so the insights for this one, for example, there's a section for traffic analysis.
• 3:49
  SPEAKER_00: And the insights are you can ask trend of monthly users last year or month over-month growth in users.
• 3:57
  SPEAKER_00: In the basic performance section, you can ask how many users did I have last week?
• 4:01
  SPEAKER_00: What are my top pages and screens by views?
• 4:04
  SPEAKER_00: On what days do I get the most users?
• 4:08
  SPEAKER_00: You get the idea.
• 4:09
  SPEAKER_00: So there are a set and finite number of questions that you can ask with these insights.
• 4:16
  SPEAKER_00: And if you were listening, you'll notice that the questions that we're asking are just asking about numbers and maybe some trends, and that's fine.
• 4:28
  SPEAKER_00: But it you don't get any other information with it.
• 4:30
  SPEAKER_00: So something to be aware of from AI Insights, it doesn't have context for the numbers it's looking at.
• 4:37
  SPEAKER_00: I would call them observations rather than insights, as insights usually include the why and lead to what comes next.
• 4:45
  SPEAKER_00: And that's why we're here today, to get from observations to insights.
• 4:50
  SPEAKER_00: To get there, let's take a look at one of these insights from that higher education website.
• 4:55
  SPEAKER_00: So I click trend of monthly users last year, and it's a line graph, which of course it is, right?
• 5:02
  SPEAKER_00: It's a trend, it's a line, and that's fine.
• 5:05
  SPEAKER_00: And I can see a trend, and it's somewhat interesting.
• 5:09
  SPEAKER_00: You know, again, this is a higher education site.
• 5:12
  SPEAKER_00: So the highest number of active users came in December.
• 5:17
  SPEAKER_00: Now I should say that I have a kid in college right now, and I can speculate based on my behavior over the past two and a half years so far, is that during Q4, for lack of a better term, but primarily in November and December, there's like a big uptick in the trend line.
• 5:37
  SPEAKER_00: And there are probably a few things going on, right?
• 5:41
  SPEAKER_00: Uh there's the FAFSA, the financial aid form that you fill out to see if you qualify for aid.
• 5:47
  SPEAKER_00: Uh, you have to submit that in that time period.
• 5:51
  SPEAKER_00: Uh, people are logging in to make payments, or they're the students are logging in to enroll for spring semester or checking final grades, and on and on and on.
• 6:00
  SPEAKER_00: So that's purely anecdotal evidence or empirical evidence for me, but it's really my own experience.
• 6:08
  SPEAKER_00: So I can't I can sort of extrapolate that out to a larger group of people, but I don't have any concrete evidence that that's the case.
• 6:17
  SPEAKER_00: After seeing this chart, I decided to look at the monthly trend of active users from this year to date compared to last year to date, just to see if there were any revealing trends year over year.
• 6:29
  SPEAKER_00: Direct traffic was about 38,000 visitors, which was down from 71,000 from last year.
• 6:36
  SPEAKER_00: If you've never been in GA4 much, a drop like that is pretty astounding.
• 6:43
  SPEAKER_00: And it gets me thinking.
• 6:45
  SPEAKER_00: It's like, so a light bulb went off my brain.
• 6:48
  SPEAKER_00: I decided to check the device category, the city, the operating system, and a couple of other dimensions.
• 6:55
  SPEAKER_00: The vast bulk of the traffic was coming primarily from users, over 40,000 of them, on Linux with zero seconds of average engagement time, which is highly suggestive of bot traffic.
• 7:09
  SPEAKER_00: And that's kind of how I knew to look at that.
• 7:12
  SPEAKER_00: I I've seen this before, where there's a huge change in direct traffic and there's been no significant website change.
• 7:20
  SPEAKER_00: Sometimes, if you roll over to a new website, some things break, but that was not the case here.
• 7:25
  SPEAKER_00: And once I sorted the data by these various dimensions, I noticed in the direct traffic in general that based on the line chart that was showing, the traffic was averaging 144 users a day, almost to the exact number of 144 or a gross, if you will, a dozen dozen.
• 7:47
  SPEAKER_00: But anyway, this was happening from April of 2022, 2022 up until late September 2025, where it drops to zero.
• 8:00
  SPEAKER_00: So here, not only do we have these dimensions that help us understand where most of this traffic is coming from, but we can also see the pattern in the line graph where it's like, okay, something is up because this is not how normal traffic behaves.
• 8:12
  SPEAKER_00: And then again, GA4 doesn't offer insights on filtered views with custom date ranges either.
• 8:19
  SPEAKER_00: It has that finite group of questions that you can ask it, and that's pretty much it.
• 8:23
  SPEAKER_00: So you're on your own anyway.
• 8:26
  SPEAKER_00: Now, I should note that for this account, we've done paid media on and off for the past six years, but beyond basic campaign reporting, we've never had a need or a directive to dig into their overall web traffic.
• 8:39
  SPEAKER_00: And so here's an example of an insight that didn't come from AI, but from slicing and dicing the data by meaningful dimensions that AI won't do for you.
• 8:49
  SPEAKER_00: And we were able to apply our skill of parsing web traffic alongside our experience of actually understanding what we were looking at to realize that it was bot traffic.
• 9:02
  SPEAKER_00: And so that's just an example of digging in to try to understand what those numbers actually mean.
• 9:08
  SPEAKER_00: And it's easy to get lost in a sea of data.
• 9:11
  SPEAKER_00: It can be difficult to know what numbers are important for your business or organization and which numbers you can safely ignore or at least keep in the background, because there's a lot of noise in GA4 and a lot of features that, frankly, they're not critical to the success of your website.
• 9:28
  SPEAKER_00: And ideally, you're working with an agency or you work with other people who can help provide some of that context as well.
• 9:36
  SPEAKER_00: So there it is, right?
• 9:38
  SPEAKER_00: You shouldn't take numbers at face value, just like you shouldn't take answers from AI chatbots at face value.
• 9:44
  SPEAKER_00: You need to have, say it with me, loud and proud, context.
• 9:49
  SPEAKER_00: And context consists of not only the things that are happening with your business, but it's also your career experience, your understanding of your industry, your institutional knowledge of your company, and your company's marketing efforts and benchmarks.
• 10:04
  SPEAKER_00: So there's all this stuff that goes in there.
• 10:06
  SPEAKER_00: And if you're new to adopts and analytics, or if you're new to your particular industry sector, it's going to take a little bit to build up that experience layer.
• 10:15
  SPEAKER_00: And that's okay.
• 10:16
  SPEAKER_00: That's just the way it goes.
• 10:18
  SPEAKER_00: But you do have to be willing to get in there to explore and have curiosity and make the effort to understand.
• 10:27
  SPEAKER_00: So earlier I mentioned benchmarks, and ideally, you have historical GA4 data that we can work with to establish benchmarks, looking at metrics like engaged sessions, engagement rate, average engagement time, active users, and of course conversions to determine what qualifies as quote unquote normal, just so you have an idea of what to expect.
• 10:51
  SPEAKER_00: Sometimes, however, there's no data to work with.
• 10:54
  SPEAKER_00: We've encountered this on completely new website builds, where there's no previous website, or on sites that didn't have the GA4 pixel installed and so never collected data.
• 11:05
  SPEAKER_00: And we've also worked with partners who'd never run paid media before, and so they didn't have any paid media data to create benchmarks from.
• 11:13
  SPEAKER_00: In the case of the not having any website data, that's just a matter of time.
• 11:18
  SPEAKER_00: In case of paid media, there are broad industry benchmarks online, if you're so inclined to look for those.
• 11:25
  SPEAKER_00: But if you don't have any data, what do you do?
• 11:28
  SPEAKER_00: You get your house in order, making sure your pixels are firing correctly, and your events and key events or conversions are being tracked correctly.
• 11:38
  SPEAKER_00: From there, it's a good idea to keep an eye on things for about a month or so, because if your site is new, it takes search engines a bit to index your site and understand where it fits in the overall picture.
• 11:50
  SPEAKER_00: So there might be some wild fluctuations early on, or it might take a little bit to get some traction.
• 11:55
  SPEAKER_00: If your site isn't new, a month is a solid amount of data to start with.
• 12:00
  SPEAKER_00: And while all this is going on, it's a great time to evaluate your business trends, like sales or visitations, enrollments, whatever the main key performance indicators or KPIs are for your organization, and also to look at the factors that have historically affected those KPIs.
• 12:21
  SPEAKER_00: Maybe there's some seasonality, maybe it's weather related.
• 12:24
  SPEAKER_00: In time, you'll have baseline data to compare the results of future optimizations or paid media efforts.
• 12:33
  SPEAKER_00: I should note here that when doing SEO optimizations for your website, it's best to do one or two things at a time, because if you do all the things and then compulsively check to see if they move the needle, you won't know which lever you pulled that caused which result.
• 12:53
  SPEAKER_00: So take your time, do it incrementally, and know that it's more of a marathon than a sprint.
• 13:01
  SPEAKER_00: When we launch a website, our post-launch QA consists of monitoring mostly and minor fixes for basic functionality, like the stuff that breaks when the website goes live.
• 13:13
  SPEAKER_00: Rather than on improvements.
• 13:16
  SPEAKER_00: Again, this gives the site time to bet in with the search engines.
• 13:20
  SPEAKER_00: Improvements can come later.
• 13:22
  SPEAKER_00: After all, we've already gone through a rigorous design and dev process with our partners riding shotgun the entire way.
• 13:29
  SPEAKER_00: So consensus has brought us to this point.
• 13:31
  SPEAKER_00: So there shouldn't be any surprises.
• 13:34
  SPEAKER_00: Unless, of course, key stakeholders were somehow left out of the process, and I've seen it happen, but it absolutely should not happen.
• 13:43
  SPEAKER_00: So at any rate, once you have this baseline data, you have something to compare future organic results to, and you can fire up your paid media to start generating benchmarks for that.
• 13:55
  SPEAKER_00: Okay, so let's go back to our hypothetical meeting.
• 13:59
  SPEAKER_00: So instead of the meeting ending and everyone going their separate ways, only to do this again next month, like we're trapped in Groundhog Day or Edge of Tomorrow.
• 14:07
  SPEAKER_00: How about we figure out how to make decisions based on the data we're seeing?
• 14:12
  SPEAKER_00: And when I say we, I mean adopts and analytics, SPMs or strategic partnership managers, and our partners, right?
• 14:20
  SPEAKER_00: We want everyone involved here so that everyone's on the same page.
• 14:24
  SPEAKER_00: Now that we know who's involved, how do we actually make the decisions?
• 14:27
  SPEAKER_00: What are the criteria?
• 14:29
  SPEAKER_00: There are two key elements in play here.
• 14:31
  SPEAKER_00: One is the benchmarking we've done, so we have a baseline to compare our results against.
• 14:36
  SPEAKER_00: And the other is agreeing beforehand what the thresholds for action are.
• 14:41
  SPEAKER_00: And usually we're talking about percentages.
• 14:44
  SPEAKER_00: For example, if we decided that a drop of 15% or more in conversions is the trigger, and then that trigger is tripped, AdOps would put together the comparison data from the previous year along with the sources of conversions.
• 14:58
  SPEAKER_00: We'd look at heat maps of the conversion pages, we'd check tracking if it's a steep drop-off just to make sure something's not broken.
• 15:05
  SPEAKER_00: And we'd take that information to the SPM.
• 15:08
  SPEAKER_00: Again, that's the strategic partnership manager or the account executive, you may call them, and have a conversation to make sure we're really seeing whatever it is we think we're seeing.
• 15:19
  SPEAKER_00: If we're sure there's an issue, we'd also come up with some solutions to counteract whatever the issue was seasonality, bad press, high gas prices, whatever.
• 15:30
  SPEAKER_00: And then we take those to the partner.
• 15:32
  SPEAKER_00: But it all starts with deciding as a team what the thresholds are that would trigger that approach.
• 15:38
  SPEAKER_00: Now keep in mind that these triggers are they're just a starting point.
• 15:42
  SPEAKER_00: There's wiggle room here.
• 15:43
  SPEAKER_00: If something doesn't cross the threshold, but there's something you didn't expect to see, a peak, maybe it's bot traffic, a drop, maybe someone paused a tag in Google Tag Manager.
• 15:55
  SPEAKER_00: You can still investigate.
• 15:57
  SPEAKER_00: Using website analytics for decision making isn't binary, yes or no, do or do not do.
• 16:04
  SPEAKER_00: It's a tool, just like we talked about before.
• 16:06
  SPEAKER_00: And like most of marketing, or probably all of marketing, website analytics and taking those insights, understanding what they mean, and making decisions based on them, it's an ongoing process.
• 16:18
  SPEAKER_00: So once you have things humming along, you've got your benchmarks, maybe you're running some paid media, you've been tweaking things as you go.
• 16:26
  SPEAKER_00: It's important to follow up after you introduce new variables.
• 16:30
  SPEAKER_00: So you make those tweaks.
• 16:32
  SPEAKER_00: Make sure you go back and follow up to see if they did what you thought they would do or if they fixed the problem or any of those things.
• 16:38
  SPEAKER_00: You always want to be gathering information, asking questions, trying to understand what meaning can be gleaned from the data in context.
• 16:47
  SPEAKER_00: If nothing seems to be changing, even though you're running paid media, that's information.
• 16:52
  SPEAKER_00: If you A B tested a landing page and updated it based on the results of the A B test, are you seeing an uptick in conversions as a result?
• 17:01
  SPEAKER_00: Are you not?
• 17:02
  SPEAKER_00: Either way, that's information.
• 17:05
  SPEAKER_00: You'd be surprised how much more effective your website can be by doing incremental optimizations.
• 17:11
  SPEAKER_00: Yeah, sure, it's not sexy to say, we changed the CTA link into a button, or we removed the giant header image, or we shortened the copy and moved the form higher up the page.
• 17:23
  SPEAKER_00: And all these things resulted in 4% more conversions.
• 17:26
  SPEAKER_00: 4% might not seem like a lot, but when you extrapolate that kind of change over the course of a year, those incremental results add up.
• 17:35
  SPEAKER_00: And more importantly, those kinds of results are longer lasting than a flash in the pan where maybe you tried something and got lucky and then you can't replicate it because you're not really sure why it worked the way it did.
• 17:48
  SPEAKER_00: So website analysis, again, taking those insights and turning them into decisions, it's not about chasing numbers or the latest and greatest fad on LinkedIn without intent, throwing everything at the wall and hoping that something sticks so you can win the internet.
• 18:04
  SPEAKER_00: Website analysis is about getting better and understanding why.
• 18:08
  SPEAKER_00: It's about building a knowledge base you can rely on instead of hoping you stumble on the answer.
• 18:14
  SPEAKER_00: So that's probably enough for today.
• 18:16
  SPEAKER_00: I appreciate you coming along on this journey.
• 18:18
  SPEAKER_00: I hope some of these insights, and I hopefully we can call them insights, were helpful to you.
• 18:24
  SPEAKER_00: And yeah, look forward to seeing you next time on the same side.