Attribution is the key to informing data-driven companies where to invest and expand, where to cut back, and how to increase engagement with their customers. But achieving accurate and timely attribution is no simple task. Join Molly, Chris, Lisa, and Alex for a discussion on attribution best practices that help you gain actionable data on what’s working and what’s not—and how to use those insights to create effective efforts that drive growth for your company.
Alex Lowe O'Connor (00:02):
Thank you guys so much for joining us. My name is Alex Lowe from Search Discovery. I'm a Media Manager here at Search Discovery, and we invited one of our clients, Ashton Woods and Starlight Homes to come in to talk with us today. I'm here with one of my colleagues, Lisa, and I'll let them go around and introduce themselves, but we'll go ahead and we'll introduce ourselves and then get started talking about attribution. So Lisa, I'll pass it over to you.
Lisa Altshul (02:07):
Sounds great. Hey, I'm Lisa Altshul. I'm a Senior Media Analyst here at Search Discovery and really excited to be here with our clients, Chris and Molly. So now I'll pass it on to you too.
Molly Pilgrim (02:21):
Hey everyone. My name's Molly Pilgrim. I'm the Brand Manager at Starlight Homes. We work with Search Discovery almost every day of the week and have a really long standing partnership with them. So super excited to be here today.
Chris Thornton (02:37):
Yep. I'm Chris Thorton. I'm the Vice President of Digital Marketing and Technology for Ashton Woods and Starlight. So thanks for having us.
Alex Lowe O'Connor (02:48):
Perfect, all right. So getting started on the topic of attribution and talking about data-driven decision making, I wanted to kind of kick off today's discussion with a first question. What's the first step you guys took towards getting your full data stack in order with end-to-end attribution? And Chris I'll pass it over to you to kind of get us started with that.
Chris Thornton (03:12):
It was actually a really big initiative for us. We kind of started working on it, gosh, I would say, you know, close to six years ago now. And we paid a lot of dumb tax along the way, I think trying to figure it out as we went, but you know, it was always kind of a big, important piece to us when we're in an industry like home building, you know, every lead is very important, right? And we're very focused on driving quality of leads and driving down, kind of, for sale on every house that we sell from a marketing perspective. So we really wanted to understand the marketing efforts that we were putting into it. And it was also a time, I feel like, when as an industry, you know, home building has historically been kind of farther behind other industries in how it's used different tactics and tools.
Chris Thornton (04:07):
And so we really wanted to invest in understanding how we were driving the most qualified traffic and the best, and most likely people to buy our homes. And so again, we spent a lot of time upfront, really kind of looking at exactly what were those touchpoints, how we were looking at potentially shifting our marketing mix from one channel to the next. And that was a big part where we started, was really defining, well, what are the touchpoints that we feel like are probably important? You know, we had to start somewhere and create that kind of baseline. And so this was at a time when we were still probably using things like billboards and newspapers in a lot of cases which, by the way, because of this, we don't really use anymore because we were able to sort of see, you know, the impact of those. While those are tried and true things in our industry that people continue to use, we were able to really use our attribution to measure and say, well, these things don't work.
Chris Thornton (05:01):
And that's something that we were able to walk away from. The other thing I feel like that was always very important for us, or that helped us kind of in the beginning, was really sort of defining that attribution model. That was one of the first things that we really wanted to come out of the gate with. And that was an area where we struggled a lot. In the beginning, we were looking at a lot of last-touch type of attribution. I think that's kind of a more standard way. Most companies still look at it. We were able to make an early shift to kind of a first-touch attribution.
Chris Thornton (05:41):
And that helped us sort of really see the bigger picture, which was, you know, it's really a combination, right? We really needed to understand the full breadth of attribution and all of the different touchpoints that our customers have on their home buying journey. You know, typically in our industry, it's a much longer trail, it's a much longer journey than most things because, you know, when you think about buying a home, typically it's the most expensive purchase that individual's gonna make in their lifetime. So they tend to take more time with it. And so that's been something that we've really had to fully kind of get our arms around and grasp. And so I think defining that model and understanding which touchpoints were the most important and how we can kind of weight the different moments of attribution to those sales through our marketing channels, that's what was key.
Lisa Altshul (06:33):
And for us from a media buying standpoint, being able to have that idea of, hey, we're not just driving a really efficient lead, but are we driving efficient paths down that entire journey and which channels are driving that, and it was really able to help us create a more holistic picture of how our media was working.
Alex Lowe O'Connor (06:57):
Nice. Yeah. I think that having the full view of not only are we driving an efficient lead, like you said, are we driving an efficient sale, or are we creating more obstacles for the sales team, or are we working hand in hand with them to drive exactly what they need at the right time? I think that's really important and a really critical piece when we're talking about attribution here. Awesome. So I guess the next step is what roadblocks did we kind of overcome when we were initially setting everything up and kind of along the way each step? And again, Chris, I kind of want to pass this over to you since you've been here for the majority of the setting up of attribution and getting everything end to end, but I know that we've all experienced roadblocks in our own way with this set of data, so you can get it kicked off there.
Chris Thornton (07:46):
Sure. Yeah. It's interesting. Again, I'm kind of leaning into the real estate industry as being somewhat unique. It's a lot of, you know, historically, if any one of you has ever purchased a new home, a lot of it is you walk into a sales community and they give you a registration card. And on that card it always says, well, how did you hear about us? And, you know, kind of before we got into this, that was really how we did attribution. And what we really discovered is basically people lie a lot in that situation, and will kind of answer whatever's at the top of the list because they don't really care. They're just trying to kind of check the box, right? So a lot of where we spend time was kind of looking at how we can really automate and capture those attribution moments and really have more visibility into actual consumer behavior versus, you know, more their opinion of how they saw you or how they first engaged with you.
Chris Thornton (08:45):
And and so obviously on the digital side, that was a lot easier to do. The biggest roadblocks, I would say that we had were kind of on the offline measures, right? When we wanted to track, you know, when the first time somebody meets us or registers, they walk into a community one, we needed to capture that moment of that registration, but then being able to take that offline registration where they're meeting face to face with with one of our sales associates, and then being able to kind of track that back to other touchpoints that they may have previously had prior to coming into one of our communities. That was a huge lift. And I will say we went through a lot of different iterations of that to really get to a point where we could truly capture that information. And it required both, you know, technical things like call attribution, you know, systems that we had to put in place to be able to capture it, as well as just kind of training our staff to be able to get better data upfront, to make sure that the integrity of the data that we got was accurate and on point to be able to kind of connect all those dots at the end of the day and take those measures.
Molly Pilgrim (10:03):
Yeah, I'll add one to that as well. One internal roadblock is showing the value to people that don't totally understand the value of data and why this is important and why it's going to give us a bigger bang for our buck. So I think that's having to explain that it's also extremely complicated and that comes into the internal training. A lot of times people don't understand, why can't I use the paper form versus the iPad? It's a lot easier to just do it on the paper. And you have to explain the value and providing that information and what it's going to give to us in terms of knowing what tactics are working and what aren't working.
Chris Thornton (10:46):
Yeah, I think, just to kind of iterate on what Molly just said. I do think it was that moment, right? In the beginning, there was a lot of internal sale selling that had to be done and it wasn't until we actually were able to show sales people, hey look, when you do this, it makes us smarter as your marketing team and we're going to drive you better leads, more qualified leads, you're going to see the sales. And so it's been great in that respect of them going out, you know, it kind of brings on the light bulb and gets people to kind of work with you on solving the problem. So from that, it's been super helpful, but it did take time to make that change.
Alex Lowe O'Connor (11:23):
Yeah. In my perspective, I think back to every conversation that I had with every marketing manager ever about why I should spend two times as long writing an email and tagging my email, versus just sending it out and getting the leads, and thinking back, like comparing data year over year and being able to say, well, we don't really know why that spike occurred two years ago because we don't have insight into it. It's all just direct traffic. But seeing that transformation over the years and really being able to—really brings light to our reporting and context to how we want to shape this summer season versus last, and what sorts of things worked and didn't work, and what we want to try now. I think it's that continuous pushing and continuous reeducation and making sure that people stay on track with putting in that tiny bit of extra effort.
Alex Lowe O'Connor (12:23):
Perfect. Any other roadblocks that we want to talk through? All right, then I guess the next piece is really just what actions are we able to immediately take once we saw these fully attributed sales and saw this full journey, and I kind of want to pose this over to our side of the rink and to talk with the agency. So, Lisa, I'll pass it over to you and see what your thoughts are there first.
Lisa Altshul (12:52):
Yeah. I mean, this was really exciting for us to be able to get, to see that whole picture. I kind of alluded to it earlier, but just working off of an efficient lead doesn't tell us the whole story. We might have a marketing channel that's driving extremely efficient leads, but buying a home is a long process and it's not necessarily an easy process and it's not something that everyone is able to fulfill in the end. So being able to get that whole picture, get it back to our channels and really granularly do media mix modeling, it made a huge difference for us to be able to actually, you know, properly allocate funds to the places that were driving us, the sales or the appointments that we needed, to be able to fulfill our goals and to drive the overall business objective. So really at a base level, just making sure that the right places are getting those media dollars is everything, and then taking it a step further. I mean, doing some testing, too, of audience testing, like, are these audiences, maybe they're more expensive to grab for leads, but are these people going to be more efficient sales in the end? And getting to test that on very granular levels and get the full picture has been really important from a marketing side.
Alex Lowe O'Connor (14:13):
Definitely. I think the one thing I would add to that potentially is just the process of budget forecasting and thinking through, at an annual level, what is the budget going to look like? How are we going to spend it? How do we want to forecast off of previous years' data? And when we have to pivot really quickly, like we've had to do during a global pandemic, where do we pull from? How do we do that? What's the safest, the least risky way to do things? But then also, where can we start to build back up? And I think that's something that a lot of companies right now, not just in the home building space, but a lot of companies are struggling with, is we didn't know where to pull from, to begin with. And now we don't know where to put dollars back into to see the biggest impact. And I feel like attribution and making those data-driven decisions and having that data at that granular of a level is really, really critical to being able to make those snap decisions and know that your data supports the move that you're making. So I think that's probably the biggest impact for me, at least most topical right now.
Chris Thornton (15:14):
Yeah. And just to kind of iterate on that, to give Search Discovery a lot of props on the work that they do with us, because we basically put them through hell every year when we're planning the next year and we're doing, you know, a lot of spreadsheet work and really kind of analyzing a lot of the data, but because we have that attribution, we're able to hit that pretty tight on our forecast for each year. And so when something does go awry, we know it a lot sooner. If something gets off, we're able to make those kinds of adjustments, and again, I give those guys a lot of credit for helping us solve that problem.
Molly Pilgrim (15:56):
Yeah, I would add onto that the seasonality that we're able to plan for is amazing. Especially with how our sales cycle works in terms of what we have built on the ground and what's going to be built in six months. I think we're able to plan around that and spend on the tactics that get us the leads that we need for those specific types of sales.
Lisa Altshul (16:17):
Something else, too, that I always think about is also like the expectation setting internally, too. When we do have this really great view and we know there's going to be a seasonal change, or something happens in the market, we can really plan for it internally. And, you know, from the very beginning, we know this is going to happen. We have all of this data to let the marketing teams know or the people on the sales force, they can be more aware of what's going to happen, too, which I think just helps the health of the business and the relationships there.
Alex Lowe O'Connor (16:52):
Absolutely. Perfect. Alright. So then the next one is just around, were there any big surprises in terms of what this data looked like? Did we learn anything about our audience based on having that full end-to-end attribution and making sure that—I think that at the end of the day, our audience is always, like, we have this idea of, we know who we're going after. Every business has that ideal buyer. Right. But I think that there's some surprises here. So I'd love to hear from you all about what those might have been for you guys.
Molly Pilgrim (17:22):
Yeah. I'm going to talk a little bit more about what we're surprised that we saw in data attribution first and then more about the audience. But, first off was something that Lisa mentioned in the question before was, a low CPL does not mean a low cost per lead. And I think it's really getting to the ground of, okay, what do we want to pay more for?
Molly Pilgrim (17:46):
We will pay more for a lead that's going to convert at a higher rate. And I think it's taking the understanding of what traffic's coming to your site to the next level and understanding what lead's going to be more valuable to you in the long run. So I thought that was super surprising. I think we are all like, let's get the lowest cost per lead. Let's get the lowest cost. Really, that's not what matters all the time. So I think that came to me as a shock because that's something that we try to drive to more efficient leads month over month. But you know, there's a lot of things that also go into a cost per sale that aren't in our control. The sales team is a huge part of that.
Molly Pilgrim (18:30):
And I think it's important for us to drill that down with them as well, because it's only going to be as good as their performance. Also, really, it's a team effort. In terms of our audience, it's different on the Ashton Woods and Starlight side, but there are a lot less touchpoints on the Starlight side than the Ashton side, because we have a quicker sales process, which is interesting. And I think, yes, we're all in homebuilding, but they're very different businesses. So I think it's always a good reminder, and we see that in our data.
Chris Thornton (19:07):
Yeah, that was probably the biggest surprise to me was how much the actual product—you know, because we obviously build a lot of different types of plans, a lot of different types of communities, and we've got two different brands, Ashton and Starlight. Seeing the variance that occurred just from product type, it is pretty crazy, but again, knowing that, we're now armed with all this knowledge that lets us, when we go into and we launch new product or a new neighborhood or a new community, we kind of, based on the product type, we know based on the geography, we know a lot of things that, you know, which channels, are more likely to be more effective. The other thing that I think was a big surprise to me honestly was the velocity of sale.
Chris Thornton (19:58):
And when I say velocity of sale, I'm really looking more at not just you know, from when we actually meet the customer, but from when we first engage that customer. So when we started seeing sales come in, but we were seeing when that first marketing touchpoint hit them, like how long have they actually been looking and doing their research and shopping prior to actually making a purchase was, you know, incredibly stunning, right? Like it definitely took us off guard, I think in some cases, but also gave us a plethora of data to really use in our planning purposes. Because we now know, based on that velocity, we know how much time it takes from when we launch a campaign to when we should expect to see sales from that campaign and we get out of this kind of what I would call a balance sheet approach to marketing.
Chris Thornton (20:53):
Well I spent this much this month and I got this much sales. That's not how it works. Right. And so I think attribution, the measuring of attribution that we've been able to do now, we're able to see that yeah, the money I spend today may not really come to fruition for 90 days. So if I am doing great two months ago and I caught my media spend, it's really not affecting what's happening now. It's going to affect what happens to me in 90 days from now. And I think that was a huge eyeopener for us.
Lisa Altshul (21:27):
Yeah. That's a really good point. And for us making sure that, yeah, we might not be as hyper-fixated on like, what was the cost per lead for this channel because we do know that that later cost per sale might be better, but what are those things rolling up to it though? You know, making sure that we are still capturing those metrics because it does take longer, and kind of being aware of that full process. When we do know that something we do today might not really see a lot of the changes until 90 days from now. I think just having that knowledge and that baseline understanding of all of that data helps us be smarter marketers. And for surprising me, I mean, Molly kind of echoing what you said was just how big that difference was between the channels that we were seeing, some of the really low cost per leads and what that actually meant was pretty great because we were able to make some really big changes and not just kind of overall big picture, but for the brand, at the different geos, we were even seeing different performance.
Lisa Altshul (22:30):
So when we'd had an idea of kind of like here's our media mix for the brand, that's actually not always going to be the right thing. Like some geos might have a completely different media mix necessary to get us to our goals for that area just based on the audiences there. And even when you break that at a more granular level, when we start to do audience testing, what works in certain geos might not work in other geos, and being able to track that whole process and actually really come to like a data driven decision about what we're going to do in different places has been really eye opening.
Molly Pilgrim (23:03):
Yeah. And just to tack onto that is, where the tactics fit in the marketing funnel too, is making sure we know the impact of each tactic and the purpose behind it, because we could see a really low cost specifically for our upper funnel tactics that aren't going to convert for, to Chris's point, 90-220 days. So each tactic definitely has a strong purpose.
Lisa Altshul (23:33):
And I think we've hit on every bit of velocity and every bit of the duration of the lead. But I think to me that was, I just want to drive that home. That was the most surprising thing to me as well, not only like how long it takes from the very first interaction with it for the user before they turn into a lead or a sale, but also just the variance by tactic.
Alex Lowe O'Connor (23:46):
And I mean, we can change up our messaging based on that. We can hit people at different parts of the funnel based on the fact that we have all of this data to support it. And I just, I think that that's really, really critical and the fact that we all have resorted to that as our most surprising thing, I think it's honestly really, really telling. I love it. All right. Next up is really just focusing on what is next in terms of attribution and data-driven actions for the team. Like, what are we thinking about next? What are we going to focus on next?
Molly Pilgrim (24:27):
Yeah. I'll touch on this one very quickly. I think it's testing. Things are changing every single day, from as little as whatever it's CTA is to as big as a specific tactic. I think we constantly have to be testing and that's going to help us evolve. It's hard for me to predict, Chris will probably have some more thoughts to this, but you know, given the landscape we're in and we're making changes so frequently, I think we've got to test and make sure that we're staying up with how each platform's performing.
Chris Thornton (25:04):
Yeah. I'll give the short version, but yeah, it's definitely a kind of constant state of evaluation. And again, I always think that it's important to remember, you have to be able to trust the data and kind of act on the data. And a lot of times what I find is a challenge in starting something like this, you know, you always hear the Wanamaker quote, right? It's, you know, 50% of my advertising is working. I just don't know what 50%, well, now it's more, 50% of my advertising is working, how do I convince everybody to stop spending money on the 50% that doesn't work? And I think there's a lot of that. That is a reality where you get kind of in that typical cadence and things that you sort of lean into that you have to kind of trust the data and go, well, you know what, that is changing. That's something we need to modify and take a different approach.
Alex Lowe O'Connor (26:00):
Yeah, absolutely. I love it. Chris, if I had a nickel for every time I've heard you say 50% of my advertising is working. All right. Then we just have a couple moments left, but I wanted to wrap it up with one last question really around just giving a company advice. If they haven't made any strides towards attribution and getting all of this set up, what do you think their first step should be? What would you recommend?
Molly Pilgrim (26:29):
So if I were talking about from the very, very, very beginning, make sure you understand where your traffic's coming from. I mean, I think that's as basic as it gets, you have to know where your traffic's coming from and then ultimately you're able to make decisions based on that. But I think if you can get as far as tying a specific person to a specific source and the other touchpoints before, it's incredibly impactful.
Chris Thornton (26:55):
Yeah. I think it's really, you know, again, along those same lines, it's really defining the categories, right? What are the touchpoints that aren't going to be important to me? And then I want to measure, how am I going to categorize those so that it's not just an indefinite list of things. You want to kind of try to organize it in a way that will hopefully make sense longterm and is scalable. And then the last thing I would say is really, again, just, define that model, right? Know which touchpoints are more important to you going into it so that you can effectively evaluate, you know, and all touchpoints are not created equal. And so how do you want to weight the different parts of the journey that your customers are going to take on the process to make the purchase?
Alex Lowe O'Connor (27:43):
And I think my big thing, and this is something I definitely struggled with at the beginning is like keeping an open mind, like the data could tell a totally different story than what you're expecting to see, and just being willing to listen to it and to dive in and to know that at the end of the day, like your data is not going to lie and we could have totally wrong assumptions of our audience, of our user experience, of all the things. And if we've got all this investment going into our data, we should listen to what it says. And I think that's probably the biggest thing for me. Perfect. Thank you guys. Thank you guys so much for doing this. I think we've outlined some really good foundational next steps for setting up attribution, for understanding it, and for putting it into action. But at the end of the day everybody's data's going to tell a different story and everybody's journey through attribution is going to be a little bit different. So I hope that this is a really good starting place for a lot of people. Thank you.