Learning to Fly: Taking Your Campaign Tracking to New Heights

March 11, 2021

Marketers use multiple channels to increase traffic to their site including paid search, paid ads, email, etc. Managing just one of those channels can be a full-time job, managing all of them and getting the full insights and dimensionality needed to make better decisions and help increase conversions can feel nearly impossible.

Join Eric Hansen, Manager of Digital Analytics Product at WGU, and Cameron Cowan, Senior Director of Product Strategy at ObservePoint, for a webinar on how to:

  • Move from manual campaign tracking processes to automated solutions
  • Grow beyond limited capabilities of UTM parameters to full dimensionality
  • Manage and align multiple campaigns, teams, etc.

Cameron Cowan
Hello, everyone and welcome to today's webinar. My name is Cameron Cowan and I help run Product Strategy here at ObservePoint. The session today is going to be all about Taking Your Campaign Tracking to New Heights, talking about, not only the maturity model of progressively better measurement for the digital space, but also a little bit of a history lesson that goes along with that maturity model. I'm really excited today to have speaking with me, Eric Hansen. Eric is the manager of digital analytics product at Western Governors University. He's a long time friend, former colleague, and a customer of ours here at ObservePoint. As we were asked who would be the best person that to have as far as a panelist and discuss these types of topics with, I immediately jumped to Eric because of his background, and some of our shared history. We both started our careers early on working for Omniture, which of course became Adobe Analytics. So a ton of good foundation there in the analytics space. Eric has also built a number of programs, client side, as far as digital digital measurement programs. and then he currently serves as an officer for the local DAA extension. So, from a number of different ways, Eric, you were the right choice to bring in on this discussion. I want to give you just a quick moment to introduce your role and yourself and then go from there.
 
Eric Hansen
Thanks, I'm happy and honored to be here and be able to participate in this conversation. I think it's a really important one, and honestly often a little bit often overlooked when people are considering their digital measurement strategy. As Cameron said I've worked in a variety of capacities and roles over the years in various forms of digital analytics. It's been great learning, whether it's been consulting with various third-party clients or working at companies to help build up their digital measurement programs. Or just serving in the industry community as one of the Digital Analytics Associations, co-founding officers for the Salt Lake City chapter. I found there's a lot of great people in this industry, a lot of great opportunities to network. As I'd like to say working smarter, not necessarily harder and making the most of the resources that we have. Really, really excited to be here and thanks for the invitation.
 
Cameron Cowan
Absolutely. And I think that that theme of working smarter, not just harder is an excellent one to kind of tee up the discussion for us today. Now, for those of you that are familiar with ObservePoint, you may be asking yourself why does ObservePoint care about touchpoint management and campaign tracking? And a lot of that goes to our history as an organization. If you look at what ObservePoint historically has been very good at, the core technology has been all about validating and automating governance around data collection. In the early days started specifically around analytics, but branched out to really any marketing tech and tags that you have on your digital properties. And we realized that no matter how good your implementations are, how appropriately your tags are firing off on each and every page, if you don't have the right data and measurement coming into those pages to begin with, you're still gonna have gaps in your measurement strategy and the decisions you're going to be able to make based on that data.
 
So our focus today is going to be very much on the campaign performance side of things and how we look at digital measurement and digital tracking. Specific, I think a little bit to the analytics world but the applications are searching certainly much broader than that. And like I said at the beginning, this is going to be a little bit of history, and the history kind of follows the maturity curve as far as how we're hoping we can help you as listeners and certainly the customers of ours here at ObservePoint move up that maturity curve to get more better, smarter data and work smarter, not just harder.
 
So let's go back in time. Maybe not this far back for me, but at least when I think about the infancy of my career in the digital space, I mentioned that both Eric and I started off early on in the Omniture world. And so when I look at, specifically April, 2005, if you want to go that far back with me two very important things, at least in my life happened in April, 2005. The first is, as many of you know, that was when Google acquired a little company called Urchin and that company became Google Analytics that, that whole platform in technology. The other thing that happened in April, 2005 was I got hired by a little green company in Orem, Utah - that was Omniture and eventually was acquired by Adobe and became Adobe Analytics. Really those two organizations shaped what is the vast majority of all of our experience with the analytics world. Certainly there are other platforms that are, are fantastic out there on the market, but I think Google and Adobe are kind of the pillars that stick out to me.
 
And when I think about what Google did with the Urchin platform, and really what Urchin had started to pioneer before the Google acquisition, we think very specifically about a UTM parameters. So it was really one of the first formalization of campaign tracking, that we knew in the digital world. and many of us are familiar with this, so I'm not going to spend a ton of time, but I do want to talk just very specifically to make sure we're all on the same page, about what UTM is. Eric, do you know what UTM stands for?
 
Eric Hansen
I think there are many interpretations of that acronym. I'm sure there's lots of popular ones. I like to make up some of my own, but I believe it's Urchin Tracking Module was the last one that I heard.
 
Cameron Cowan
Urchin Tracking Module is usually the most accepted form, or interpretation of that abbreviation. I know I've actually heard one of the co-founders of Urchin say it's something different, it's Urgent Traffic Monitors. So there's a few different ways to look at that. But no matter how you spell UTM, it means these tracking parameters. UTM has source and medium, which are essentially required. Every channel, every touch point post ad you have should have both a source and medium source, and in most cases is actually required for automatic processing by Google Analytics. So if you're missing that and you have medium and other things, a lot of times that won't even be picked up. Campaign, while though not necessarily required, some things don't always fall into campaigns, most things do and it's strongly recommended that you include that dimension. And then there's a couple of other optional parameters to round off the bunch as far as term, which is very specific to the paid search world. And content, which can get a little fuzzy at times, but has its uses.
 
And so when I think as far as, a maturity model and this kind of beginning state of just getting some tracking in place, I know that we here at ObservePoint has seen a number of the biggest brands in the world that simply aren't tracking at all, a lot of their external touch points. They've got good links. They go to a site, the site lands there's even analytics on that page, but there's no tracking to capture. And so there's no way to identify the, the attribution that should be going back to those channels. And so looking at the benefits of just getting to that UTM standard, I think there's a lot of very clear ones.
 
Simplicity is the first one with five parameters. Most people know, what needs to be entered. it's, it's a basic high level view of the performance of your data. It's configured by default, so there's no additional implementation or code deployment to get it to work in Google analytics. As long as you have those parameters in your URLs, and you have Google analytics on the page, it's going to pick them up. Because of those two things, because it's simple, and because it is part of the default Google Analytics world, it's a fairly well-known standard. Many of us that started our careers in digital started on a free version of Google and learned on that platform. So UTMs have become part of our history, as we've begun that journey.
 
And then finally, because it's very universal because Google Free is out there. And one of the most used platforms in the world, there's a lot of free link builders that help you build these UTMs fairly straightforward and gets people up and running quickly. Now, Eric, as one of the founding officers of the local chapter of the DAA, you see businesses of all levels of maturity. Those that are just starting out, those that are, paying millions of dollars a year for an analytics package. Talk to me about how you've seen those benefits of just getting started and maybe some of the basics of the UTM world.
 
Eric Hansen
What's really nice about the ubiquity aspect of UTM parameters is that even though it's configured by default and Google analytics, a lot of other analytic systems take advantage of those UTM parameters just because they're out there. Just because you might not be using Google Analytics Free, or you might be using something else, a lot of times people still use UTM parameters to pick up on it because it makes it easier to transition to a more sophisticated platform, and that sort of thing. So a lot of the people that I've worked with, even though they may have grown beyond Google Analytics Free, they still use UTM parameters just because they're already in place and they set up their new analytics system to start pulling in those parameters as well.
 
Even though this is Crawl, or this is kind of old history, as you might say, it's well-established and it gives people insights into how people are getting to your website. When we think about digital experiences, digital experiences don't start when someone comes to your app or to your website, they start when they first get introduced to your brand. So understanding how people are being introduced to your brand and making their way onto your site is a critical piece of information that if you're not capturing, you're missing out a huge part of the story.
 
Cameron Cowan
I think that's exactly right. That foundation, while simple and very ubiquitous, also obviously then has some limitations as well. And if we think about the flip side of the benefits, what are some of the drawbacks? What are some of the things that you can't get from using this fairly basic approach to measurement? I think the list is fairly long, which is why there's a progression in the maturity model. Some of the ones I'll call out, limited flexibility, obviously, there's just those five parameters. So if you want to measure more than that, if you want to measure 10, 15, several dozen different dimensions, you can't do that with the basic five UTM parameters. Also there's some limitations in the default hard coding. For example, with the medium, if you want it to automatically be bucketed into paid search, you have to use CPC (cost per click), or PPC or paid search all as one string. And if you use anything else, it doesn't automatically get bucketed and you have to rewrite the rules or you've got data going into a separate bucket. So there's some of those things where it's, it's nice to have defaults that you don't have to do anything to set up, but at the same point that can create some constraints in what you're doing.
 
Also because there are multiple parameters, it's not just a single ID, a lot of times those URL strings can get really long. I think the average UTM parameter link that I see externally has anywhere between 250 and 350 characters on it. And if you're adding any other tracking or anything that gets really long, really quick, and all of a sudden the average consumer starts looking at that URL saying, what are they actually tracking about me? I don't know how all of this stuff works, but it looks like there's a lot more than just a base website there. So there's a handful of things there.
 
I think the last thing I'll quickly call out is just the open to interpretation nature of it. Whereas it is fairly ubiquitous and understood as a good standard, I still have customers today that just mentally flip medium and source. Well source, my source is email. Okay. No, that's not my source, that's my medium. Or the source is paid search, that's the program that I'm running, so those things get flipped. Campaign versus content. A lot of times people will actually just put the call to action or the promotion, which is part of the content, but they'll also put that as their campaign name. So there's certainly a lot of variability that can happen there. Yet, even with five parameters there's a number of mistakes you can make or drawbacks. Any other observations that you've seen as far as limitations of UTM America?
 
Eric Hansen
It's not a huge one, but you know, with so much information in your URL that's open plain text, it can open yourself up to a little bit of competitive intelligence threats and risks. Where your campaign names are right there in the URL. And if someone wanted to do research on your company and get a sense for how you structure your campaigns and where they might find some opportunities to go up to bat against you, that's that's one place they might look.
 
Cameron Cowan
I think that's exactly right. And being able to mask that information is at least one of the reasons why we started seeing standards change in the mid to late 2000 era. Like I said, that's when I started working with Omniture back in 2005, I know Eric, you started, what about two years later?
 
Eric Hansen
Yep, 2007.
 
Cameron Cowan
And as we look then, as the progression from just basic Urchin based tracking, which dates back almost two decades, to custom CID (custom campaign identifiers). And how do we still get a lot of those same benefits that we saw in the early UTM world, how do we get that in our analytics measurement, but we get rid of a lot of those drawbacks we just mentioned? Eric, talk to me a little bit about as you came on board with Omniture and were helping a number of different customers across the globe get their campaign tracking in place, and also your early experience doing this in brands themselves.
 
Eric Hansen
I think that the whole concept of custom CID's really starting to open up pathways and doors of opportunities that were previously close to us with the UTM parameters and Urchin parameters. UTM parameters were great, it's better than nothing. And honestly, if you're just starting out in analytics, you're a business of one or in a small business, that might be where you need to go and that's okay. There's nothing wrong with starting at the basic building blocks. the thing that we learned, that applies regardless of where you're at in this process is being intentional about how you develop your strategy for tracking campaigns develop a plan and have some consistency. And that definitely started to show itself whenever we started talking about building custom CID's.
 
When I was at Omniture working with various clients, and when I've been in house at various organizations, we would create different types of resources that could be used by our marketers and other individuals that needed to create these codes to help provide some guidance and guide rails. For example, we would create a template kind of document, whether it be an Excel or a web form, that would help guide each of our marketers in creating the components of a custom CID that might have a concatenated string of words or different types of symbols or identifiers to help us understand things that mattered to our organization. Whether it was the channel that the specific tactic, or the campaign name, the date, or any of those different aspects. Then we would use various tools or systems, whether it was as simple as an Excel workbook, or use a regular expressions to help us classify and sort out, and kind of parse out the meanings behind each of these custom IDs and integrate that into our analytics systems. In Adobe analytics, it's really popular for us to use a what's called a classification system that would allow us to upload metadata about our custom CID's. So that way we're not having the super-duper long URLs anymore, and they're a bit more concise. So that's not just a big spaghetti mess of words and things on that URL and the person's browser, but we still get that valuable information in our analytics systems broken out by the dimensions of data that we started to care about.
 
Cameron Cowan
I think you hit on a number of these things, the potential to expand, the dimensionality beyond just the basic UTM parameters. So, as we started deploying custom CID's across the world, people saw that I don't just have to look at the world and those five buckets, I can look at 10 different things, or 15 different things. I think even back to the earliest days of site catalyst, and now today in Adobe Analytics, the general limitation is about 30 different custom dimensions, or like you say, classifications on top of that metadata. So certainly expanding what we can do and the way we think about our measurement programs. A much more flexible to adapt to what specific to my business and my need. And maybe if you're in a different vertical, you have different parameters or different dimensions for your content, for example, that just don't matter to someone else, but it's really important for you to measure.
 
The other thing I like about it is it's not fixed, that you can grow over time. Maybe I'm just going to start with 10 different dimensions today based on my CID, but as my program gets more sophisticated, more advanced, there's more things that I'm being asked to dig into from an analysis perspective, I can grow that over time as well. And you don't really have that in a more rigid world of UTM parameters. But even with custom CIDs. And I don't want to make this sound like it's Google versus Adobe, because that's not the case at all. Google also allows you now to use custom CID's and try and track those in custom dimensions. It's more just moving beyond just the basic UTM parameters, but there's still drawbacks that exist today with custom CIDs. And certainly what we were seeing, probably 10 plus years ago.
 
You mentioned the spreadsheets, and we all had them. Just these manual spreadsheets that existed on somebody's machine, or, I mean, at that point, we weren't even really sharing Google Docs or anything like that. And so it was, do you have the right version, and is it up to date, and are people overriding each other? So there wasn't a ton of governance in place there. And even if you could get that part of it cleaned up and you got the IDs deployed, there was still a very manual classification process. How do I then apply the right metadata on top of those key codes? So I can get the analysis I need. I know that that's improved in most analytics systems, just over the last decade, where it now, some of it is more rules-based you can write rejects. There's a few other things that make it easier.
 
There's still the maintenance of those rules or those processes, and making sure your IDs are at least somewhat, either human or machine readable, so they do follow regular expressions that can be parsed out by someone or something. So there are some natural limits there. One of the things that as I followed your career, as you've gone and done a number of things for a couple of different brands, client side, is you've taken this core foundation of custom CID's, looked at all these drawbacks and said we could do better. And so as we move on to the running phase of our maturity model, rather than actually give it a name, for the technology, I wanted to give you a moment to talk about, WGU specifically, and how you built out that measurement program for maybe a more slow walk, to a brisk walk, to a run and how you've developed that over time.
 
Eric Hansen
Yeah, sure. I would just want to call out really quickly as we transitioned away from the crawling of UTM parameters to the walking of CID's, one of the other benefits that I forgot to call out was the benefit of being able to fix mistakes. With UTM parameters, if you messed up in your spelling, too bad, you're stuck with whatever you sent in. But thankfully with whenever you're separating your metadata away from your code or your tracking ID, you're able to change that metadata after the fact, especially if you accidentally messed up or needed to make corrections. And we all make errors, especially when we're manually managing all these IDs in spread sheets and such.
 
Cameron Cowan
That's right. I like to think about it as using a Sharpie on a whiteboard, you can do it and it'll send a message, but you can't change it after the fact. And it becomes really hard to fix it when you do make those mistakes.
 
Eric Hansen:
I can empathize with that all too much, unfortunately. But as far as WGU story, it's been an interesting run, no pun intended. When I started at WGU about four and a half years ago, we were in a less mature state, we'll just say that. We were definitely working off of a more rudimentary system of tracking codes that we cared about. It had been established back in 2012, so you can see what we were working with there. It was a little bit dated, but it was still useful. It gave us some really good insights. One of the major challenges that we started to run into with these CIDs that we had created and were managing is that we had all of this great metadata, about our campaigns and the things that we were doing. But we were quickly running into roadblocks and understanding how do we optimize against the, this information that we have? Because oftentimes the data we were collecting was at such a high level, it was hard for us to understand where we could pull a lever here, or make a tweak there and really drive that campaign further, or fix a campaign that was faltering or flagging behind.
 
It also didn't tell the whole story. Metadata is great to understand how people are coming to our site, but we could only evaluate efficiency metrics at a very high level, at the account or campaign level. Because costs were kind of aggregated, it was very hard to manage. We had many, many ads across many different channels, and so we ended up just saying, okay, so here are the costs, and this specific channel here was the conversions and we had a rough efficiency measure. But again, how do we know which campaigns inside of that specific channel or account were especially valuable or not? We realized there were some limitations there.
 
Cameron Cowan
We actually did some calculations and the time that was spent by our teams managing the metadata, managing the cost and doing these calculations, took up the equivalent of at least two full-time headcount. So, that's a pretty significant investment of time and money that was a manual process that was not really going away. In response to this, we decided we needed to expand our homegrown solution and make sure something that we could actually make more actionable. So that involved creating a taxonomy and stop thinking about just individual channels, and letting each individual channel make their own reports and presentations. Because honestly, if I was in their shoes, I would cherry pick whatever data I could to make myself look really good. So it involved a lot of effort and a lot of time and planning about six months worth of meeting with various stakeholders to create a cross-channel taxonomy that was unified. That everyone in marketing, no matter what your function was, could understand, okay, this is what defining as these channels. These are the attributes that we care about.
 
Eric Hansen
For example, WGU we're non-profit, so maybe the, the, the, the revenue aspects weren't as important, but the messaging that we're trying to share about the values or the missions of the nonprofit, that was really important to us. So one of the dimensions we had was, okay, these are our messaging pillars. What do we care about, how are we going to message this information? And we came up with about 18 to 20 different dimensions and had a very common taxonomy across channels, not just digital channels, ALL the marketing channels. To make sure that we were able to track things as granularly as possible. That way we could start getting into detailed levels of metadata about whatever channels we could and start to optimize our campaigns and tactics and strategies within each of these groups.
 
Cameron Cowan
That's great. And let's see if I captured some of those things as you were going through the progression of the brand. First, creating a cross channel taxonomy that, that becomes more possible if you're thinking about it not just for individual channels or tactics, but taking that broader level approach. and it also gives you a higher level of dimensionality. You're not just dealing at the channel or the publisher, or even the campaign level. you were getting really, really granular with how you're tracking these things. And because you were doing it at a much more granular level, individual ad elements, individual post components and messaging pillars, you're able to actually then get into a true ROI. I think that's one of the things we just, we immediately start using the word ROI when we really mean, well, what's the attribution, what's the conversion volume and just the return as far as total success? But ROI, it's a calculation, it's got two components and you mentioned cost very specifically, that in order to get a true ROI and down to the right level of granularity, first, you have to be tracking down to that granularity. And then you have to have processes in place to be able to bring in that cost.
 
As far as costs, you mentioned that being important part of what you were doing, bringing in the costing level at the right granularity. For something like a paid search or paid social or display, those costs are pretty straightforward. That's advertising, a lot of times, is by unit; whether it's by impression or click or whatever that is. But talk to me a little bit about how you incorporate it in cost for things that aren't as easily interpreted. Things like your investment into your SEO practice, your email marketing program, in-house, potentially even some offline advertising. How do you kind of line those things up?
 
Eric Hansen
It's a really good point because definitely it is a lot easier to track some of these digital channels, relative to some of the offline channels or other more ambiguous channels, things like our SEO investment. But yeah, so where you can get granular data, we go for it. Whether you're integrating with an API scheduling regular exports from your advertising partners, whether it's Google Ads, or whatever it might be, getting the individual costs per ad at the granularity, that makes sense for you really goes a long way to helping that individual group organization make optimizations. Now for the non traditional channels, there are still costs associated with it. And whether it's working with your organization's finance department to kind of figure out what those investment costs are, or getting information from your agency, that's doing your TV buys. There's lots of different data points that you can pull in.
 
What we found was even if you can't get costs down to the granular individual ad level, for your non traditional campaigns or your offline campaigns, getting as much granularity as you can, both from a tracking standpoint for metadata tracking codes and people coming into your experiences through those non-traditional channels, as well as understanding as well as getting as much cost detail as you can provided a great foundation for us. Musing that data as an input into our marketing mix model, so that when you develop a marketing mix model, you are able to then use those signals and depending on whatever approach you use, structural equation modeling or other more sophisticated machine learning models, you're able to apply those tools to help you get a sense for what were the contributive effects of those non-traditional channels or less directly attributable channels. At the end of the day, this whole taxonomy and system really enabled us to have a more robust attribution system in place that would be possible for us to consider things like multitouch attribution or things other than simplistic, last touch attribution.
 
Cameron Cowan
I think that's the right way to be thinking about a lot of people think, well, I do multi-touch attribution for all my digital channels, but I'm just going to do media mix modeling or marketing, mixed modeling for all my traditional channels there. They never touch each other. They never talk to each other. And I think you're right, you get all the data you can and inform your attribution models as best you can. Know there's going to be some fuzziness or gaps for certain channels or media type, but still then use the signals from those digital channels to inform your media mix model as well. You can't separate the two and if you do, at best, you're going to be skewed or worst you're going to be directionally inaccurate, which nobody wants.
 
Now, these are some of the benefits of the program you built out, but still even, where you guys are at today or where you at least were a couple of years ago, there's a couple of drawbacks to this homegrown approach. I think the first and most obvious one that you called out, was the intentional thought component to it. You have to actually plan, you have to bring people together. You said it took about six months. You have to line up different stakeholders. I think part of it is you have to understand what are the questions I want to be able to answer at the end of this journey, and then back your way into it, not just saying, okay, here's some dimensions that may be useful at some point in time. Did you find that tricky? Was it just time-consuming or tell me a little bit about that intentional thought.
 
Eric Hansen
Honestly, you have to be a champion of this. Every individual department, the paid search team or the social media team, they have their interest and as long as their interests are being met, they're not going to really care if there's a unifying taxonomy across the board, so that involves getting buy-in. And buy-in at the right levels in order to implement something that is more holistic and universal. So that involved us going to the CMO, presenting that proposal, and making that a top-down directive so that there was incentive for each of the different groups inside of marketing to work together towards building those types of things. But then there's still a lot of negotiation, back and forth, coming to a consensus of, okay, this is what it should look like. And getting that buy-in and enabling the teams to use these systems and start to see the value that could be possible by investing a little bit more time and training to track things at such a granular level.
 
Cameron Cowan
And I think one of the things that I've seen some brands make the mistake on is thinking, okay, I just spent six months, we did a great job, we're done. And that idea of we're done is a huge misstep and understanding that marketing taxonomies, especially when done right across channels, across organizations, they're living and breathing thing, and they need to evolve over time in order to evolve with your business. Your business needs don't just stay stagnant over long periods of time, so your measurement strategy needs to reflect that. I think the other thing that you sort of alluded to, and I know we've talked about in the past is helping all of those people. It's one thing to say, okay, we've agreed upon a standard. It's another thing to actually get them to follow the standard and then to use it going forward. So being able to enable teams across the board, as far as how are we doing this, and also why are we doing this? Is that what you found as well?
 
Eric Hansen
Yeah, absolutely. And putting whatever processes you can in place to make sure that you're able to support them, whether you're creating documentations, holding those in person trainings, and nice plug for ObservePoint, we actually use ObservePoint to help validate the tracking codes that people are creating to make sure that the data is coming into the systems on their new landing pages that they created. And we do that on a daily basis, that's great.
 
Cameron Cowan
As you're pulling in that data from all of these different sources, and I know another thing you've alluded to is, it's no longer just about analytics, that isn't always going to be your single source of truth. You're pulling in data from a lot of different places, ad platforms, you've talked to me a fair bit about your CRM is kind of being one of those sources of truth, if not the source of truth. Talk to the group a little bit about that.
 
Eric Hansen
Yeah. As a university we are essentially a lead gen type of front end marketing because we have to create leads that our enrollment counselors can work with and see if they're a good fit for WGU or not. So a lot of the really interesting things that happen with people that interact with our brand happen offline, on phone conversations or in other conversations outside of a marketing web website experience. So integrating with our CRM was a huge component, and that's kind of where things started off originally with our initial system. But that needed to be maintained and stable throughout this transition, as we expanded the capabilities of what we were trying to track there. Then talking about integrating with all of the different ad platforms, that's a lot of different pieces of data that we needed to scale with. And we quickly realized we didn't want to become a software development shop. We wanted to bring in some outside help.
 
Cameron Cowan
You've done that in your past year, you've moved on. And I think that's the last major drawback. The other points are less drawbacks, but more definitely considerations and things you have to incorporate into the process. Scalability truly is a drawback when you're starting to talk about, oh, well, I could build that, I could integrate with the Google Ads API, for example. There's a lot of businesses that have the technical chops in house that have development teams that could build that out. And now you say, well, okay, we're not just advertising on Google, but it's also Bing, and it's Facebook and it's Twitter and it's LinkedIn. And then we want to be able to send all that data to our analytics platform, but not just analytics. We also need to be able to send it to the CRM and to maybe an S3 buckets. So we have it for internal data warehousing, all of a sudden that's okay, well, that's a much bigger process than I thought. And am I really, is that my core competency is building out that software for some people it's still, yeah, I could do that. And I have the resources. We have budget to throw out that this next couple quarters we're going to do that.
 
The other thing that I think people often overlook is who's going to maintain it over time. There's always going to be a new ad platform, a new DSP, a new source of data that you need to pull in. Things are going to break APIs for external parties change all the time. So who manage that? Going back to our maturity model and kind of talking about this crawl, walk, run approach, all of those things are still us doing things crawling, walking, running, is us physically going in and doing the work.
 
When we start going to that last stage of learning to fly. I haven't yet met an actual human being that can fly on their own. Maybe there's some out there that have special talents that I don't know about. That's when we start relying on machines to do a lot of that work for us, but it takes us to heights we've never been able to go before. That's where a lot of that scalability comes into play. And when we think about what we're building here, ObservePoint, but just in general, how do we get people to better processes? I go by a few specific mantras first where possible, fully automate. There's a lot of great API out there that you don't have to do anything, and you can actually get far more data than you ever have before. most of these third party ad platforms, for example, that we've talked about, they have their own ideas just piggyback off of that.
 
Don't worry about creating your own tracking codes and these unique concatenations or adding UTMs, just use the ideas that exist in Google and Facebook and other places. If you do that, we also know that all of their metadata is tied to their own ID. So you can pull through a ton more data than you ever have before. Actually just jumping ahead. One slide, a good example of this is if you look at just UTM in the world of paid search. You've got medium, source, campaign, term, and content. Five dimensions. Great. If instead of doing that, adding those parameters, typing them in yourself, or you're using a URL builder and then pasting them into Google, you simply use the Google dynamic macros. You can get a couple of dozen different dimensions automatically without having to do any additional work. And that's really what we're trying to get customers to the point of doing and marketers thinking about I can fully automate a lot of this work. Maybe it's 50%, 60%, 70% of my total volume.
 
And then there's always going to be those channels where the full automation isn't possible. Either API's don't exist or B unstable or immature, or there's too much nuance and exactly what you want to track that those systems don't support it. So templatize that I still need governance, I still need a good clean way to say it needs to be done this way every time. but I'm not going to be able to fully automate that, so I put things in places that get me 70, 80% of the way. I think the other thing that I've noticed is getting data out is just as important as getting data in. We've talked about these sources of truth, whether it's your analytics platform, whether it's your enterprise customer data warehouse, whether it's your CRM platform. I know a lot of people are adopting CDPs. You need to be able to get data to and from the standard, and so being able to automate that as well is really important. I know as we've worked with WGU, getting all of your metadata from your ad platforms, through our system and onto Adobe analytics, for example, has been, been a big part of that. Talk to me just really quickly about what's the value there as far as automating and how much time is that saving you and your team?
 
Eric Hansen
Like I alluded to earlier, at least two full head count, right? So, I mean it really helps with the justification of investing in a platform. We did not want to maintain all of those APIs ourselves. Like you said, they change all the time. They're finicky. They can be hard to deal with. Let's pay someone who's focused on doing that, to do that work and let us focus on the real work of making decisions about the marketing data. It's been hugely beneficial. We were able to drive down our cost per conversions by double digit percentage year over year because of all of the new information we were getting about efficiency metrics.
 
Cameron Cowan
I think the other thing that you mentioned, as a final note, is as far as automating things, or at least putting in greater processes that are letting machines do what they do well, is the ongoing validation. Everybody's got hundreds or thousands or tens of thousands of different links that exist in all these external places on paid search, in display, on owned social media, in your email program. And to be able to have a way in which you can say, okay, I don't want to have to check every one of these one on one, go in and copy paste them into URL, make sure it's still working. It did a week ago. Does it still now, but actually having automated processes in place to keep that governance, flag where there are challenges, not only challenges of is the tracking code there and in the right structure, but once it gets to a website, for example, is the analytics code there and is it functioning the way it should so it can capture that data and we have that full closed loop? Once again, where possible, fully automate. Where not possible, build processes and templates and hopefully a good SAS based environment where you have the governance you need in order to execute.
 
The last thing I want to talk about in our final minute or two here, is we've talked about this progression from crawl, walk, run, fly... everybody's at a different stage of maturity. As you move, even from one organization to another, you may find yourself going forward or back significantly to adjust for where that organization is at. But let's assume everyone is flying today. And they've gotten to this level where they're measuring everything exactly the way they need to. Most of the works off their plate because they've automated a lot away, where do you see us going over the next five years? What are the opportunities? Is there something beyond fly or we're just gonna be flying in a more efficient manner? What does that look like to you?
 
Eric Hansen
Oh goodness. There are so many things to consider in the next five years. I've got to say a big part is with the changes that we're seeing in the privacy landscape and the technology landscape, cookie degradation, making sure that you have an appropriate strategy going into these, I guess, air turbulence, as we fly along, it's making sure that we handle that appropriately. There's actually a lot of opportunities there because the old paradigm of building a wide base of prospecting, retargeting on them, and then building a first party relationship with the customer, that's going to be turned on its head. Marketing is going to be all about that digital experience and relationship that you build with people that interact with your brand. And so that means, you need to start thinking about touch points more holistically, even beyond just your external marketing campaigns.
 
What about the high value assets that exist on your site? What are those videos or those app experiences that are really driving your customers forward and making that relationship more solid? Making sure you have that as part of your tracking, tracking the things that aren't necessarily directly part of your marketing. Word of mouth campaigns or referrals that happen organically, making sure that you have a plan to track those is super important. APIs are becoming more and more standardized, thankfully, and I think the expectations around that are increasing that we can reasonably expect in the next five years, that a lot of this thing, a lot of these integrations and partnerships that we have with various vendors or channels, we will integrate with them and need to be able to integrate with them more holistically at a very granular level.
 
Cameron Cowan
I think you're exactly right on a few of those fronts. Privacy is playing an ever more important role on everything we're doing when it comes to measurement. And as we see the cookie landscape change, I hear a lot of doom and gloom about cookies are dying. Maybe, but measurement isn't dying and we need to find good ways to re-engage with our customers. You mentioned the mobile app environments, most of those are authenticated experiences. So being able to have those direct one-to-one first party data relationships with our customers, I think is going to continue. Then you mentioned online and offline, the smartest brands in the world right now are finding ways to not just think about digital measurement, and think about a holistic measurement strategy for the business. How do I unify my taxonomy? How do I unify my data, and how do I get insights to help me understand no matter where it is online, offline? I know where I'd put my next dollars and where the best investments are. That's what this is all about is getting smarter no matter where we're out of that maturity curve.
 
Eric Hansen
Right. And post-conversion touchpoints too. This is a relationship, right? It's not going to stop after they buy your product or invest in you continuing that relationship on. There's a lot of data there.
 
Cameron Cowan
Absolutely. We oftentimes just think in terms of marketing channels and we need to realize it's about customer experiences and, and those go far beyond any single transaction. Well, Eric, I want to thank you for joining us. This has been great chatting with you. I've had awesome experiences working with you as a colleague, working for you as a technology vendor, but just really appreciate the opportunity to have a discussion today and let some other folks listen in. As we talk through how measurement has changed over the last couple decades.
 
Eric Hansen
Thanks for the invitation Cameron. I really appreciate it and hope everyone has a great day.

No Previous Videos

Next Video
Privacy: The New Normal
Privacy: The New Normal

Join Chris O’Neill, Solutions Architect at ObservePoint, for a discussion on privacy best practices to make...