Maia Morgan Wells:
Welcome to another session with ClearPivot founder, Chris Strom. On this episode, we're diving into an important question for anyone interested in understanding the true pathways of their best customers. Join me, Maia Morgan Wells, as I talk with Chris Strom in detail about the differences between web analytics and self-reported attribution. Chris Strom, welcome back to the show.
Chris Strom:
Good to be talking with you again, Maia.
Maia Morgan Wells:
Well, listen, before we dive in, I want to set the context just a little bit because we're talking here about two different ways to track how someone came into your marketing funnel, essentially. We've got web analytics. Probably the most applicable thing that we would pick out of web analytics here would be source or where that person came from, but that only tells a particular part of the story.
From what I've been hearing from you, the way people self-report their pathway might be different in many ways. I guess I'm wondering, can you explain those two types of data for us, set the context a bit? Then also I'd like to know, why are you interested in this question right now? How did this come back up onto your radar in the last couple months?
Chris Strom:
Definitely, yeah. Self-reported attribution is basically just a fancy way of asking people how they heard about you when they come, they fill out your form, or they call in via phone, asking them how you heard about them and reporting their answer. In the past, a lot of companies used to do it.
Then probably with the advent of Google Analytics specifically, when that came out, they introduced the original source tracking of the website visitors. That's things like tracking whether someone came in from organic search, or paid search, or social media, or email. Or if they couldn't tell, they would just say direct traffic. That's the web analytics that started getting reported. Google Analytics started doing that first, probably. We do a lot with HubSpot analytics where it tracks similar sources as well.
For years and years, we actually were telling people, "Web analytics is superior to the self-reported attribution of asking people because, oh, people are never going to remember correctly and they'll probably just make something up, and so you should use the technology of web analytics instead." That was basically what we told people for a long time, and so that's the context. I was getting a little frustrated with the limitations of web analytics themselves too, and having worked in web analytics for quite a long time, I knew a lot of their strengths but also a lot of their weaknesses too.
Over the summer, this year especially, I started seeing other people I know and follow start getting a second interest in going back to the self-reported attribution where you just ask people how they heard about you. They started to share a lot of the insights that they were seeing. They found it was, in a lot of ways, it gave more accurate information, and so that really got my attention. That's what caused me to circle back a few months ago, and I started doing some experiments in adding in self-reported attribution and comparing it against the web analytics.
Maia Morgan Wells:
I'd love to get into what differences you've seen and what kind of details we can learn from that process. I guess first, before we dive into that, you mentioned limitations of web analytics. I know you are very, very well-versed in web analytics. What are some of the limitations that you started to recognize with that type of data?
Chris Strom:
Yeah. There's a couple of big limitations. They can be really great. They can provide a lot of pretty granular data a lot of times, especially when it comes to paid ads, especially Google ads or social media ads, but there's some big weaknesses too. One big weakness is the channel called direct traffic. There's just way too much of it.
Direct traffic is basically where they couldn't detect where the person came from. There was no source data that came through in the visit session, and so the historical answer we would say is, "Oh, well, if someone types your domain directly into their browser and goes directly to your site, that's what the direct traffic is." That certainly would be direct traffic, but technically what direct traffic is, is traffic where the web analytics just couldn't tell where they came from for one reason or another.
Sometimes, it might be that they typed it in. Sometimes, it might be maybe they were searching for you and maybe Google Chrome auto-completed it, but it certainly can't account for all of it. I'm looking at the analytics of one of our clients here. Over the last two years, they had 213,000 website sessions and 30% of those got logged as direct traffic, so that's about-
Maia Morgan Wells:
Wow, that's quite a lot. Yeah, that seems like quite a lot.
Chris Strom:
Yeah. It's like 70,000 sessions reported as direct traffic. There's no way on earth there's 70,000 people typing this company's domain directly into their browser over the last two years. Yeah, direct traffic is basically just it couldn't tell, and it could be for a variety of reasons.
Some specific reasons might be someone had an email that they forwarded to someone else, and then that person came from the email and the tracking links got broken in there. Or maybe it was from a Slack message or some other direct message that they clicked from. Or maybe there was a VPN in the middle, or a browser cookie blocking, or things like that. There's a variety of ways where ... a variety of things where it just breaks the tracking. Pretty consistently across all of the clients we work with, we're seeing 29, 33, 35% of the traffic just being reported as direct traffic across the board. A third of the traffic, pretty consistently, just can't ... There's no identifiable source tracking, so that's one big limitation. It doesn't work about 30% of the time.
Another limitation is organic search overreports because it doesn't account for ... Organic search overreports and then things like social media underreport because in web analytics, if someone clicks to your site from a social media profile directly from the profile, it'll show up as social media, but a lot of times what happens is people follow you on social media and they just consume your content on the social media channel for a long time.
Then one day, when they have a need of your service, they think, "Oh, I need this thing. Who does that?" and they remember you from social media and then they just Google search you by name. The web analytics will show that as organic search, but it's actually, it was technically an organic search, but they're just searching for your name because they originally found you or followed you on social media.
Organic search overreports. Social media underreports. Also, word-of-mouth is not detectable at all. Someone's friend tells them about the company, so they Google search the company name. It shows up as organic search, but it's actually word-of-mouth. Or another big one is foot traffic. They live nearby or they were driving by, and they saw signage, or they saw the building location and they said, "Oh, that looks interesting," and then they Google search the name and so it shows up. That's another case where it shows up as organic search, but it was actually from foot traffic.
Then one other example too is email marketing. One of our clients, they recently did a VIP day for their customers where they invited all their customers in for an in person event. They emailed it out to their mailing list. It was a big success. A lot of people invited their friends and they got a lot of registrations. The day they sent out the email, they got 65 signups for it.
Maia Morgan Wells:
Wow.
Chris Strom:
If you look at the web analytics, it shows 445 website sessions from the email, so 445 cases of people clicking the link in the email and coming to the site, but in terms of new contacts, it only showed one new contact from email, and that's because people on the email list are already contacts. Meantime, 62 out of the 65 new people came in the same day from direct traffic and from organic search. It was clearly because of the email because that spike shows up the exact same day the email was sent, and so that pretty much has to be word-of-mouth.
We sent the email to the existing contacts. The existing contacts sign up. Then they tell their friends about it, and then their friends Google search the name or they just go to their website directly and register themselves. The web analytics can't track that dynamic at all.
Maia Morgan Wells:
Mm-hmm. It seems like there's a little bit of educated conjecture going on there just assuming, okay, this makes sense if it happened this way. Does self-reporting help to close that gap a little bit? Where does self-reporting come in and understanding those differences?
Chris Strom:
Yeah, self-reporting helps to close that gap. Say in the case of the VIP day event with the email announcement, one of the existing contacts gets the email. Maybe it's invite a friend and get 10% off, or something like that so they tell three of their friends about it. Then those three friends, maybe they Google search the name and come in, and so the web analytics says they came in from Google search, but if you ask them, they'll say, "Oh, my friend, Patricia, told us about this." Or, "Oh, yeah. I heard about it from another friend," or, "So-and-so forwarded it to me." In that case, by just asking them, you actually get the real source of how they found you.
Maia Morgan Wells:
Mm-hmm. What do you have to have in place in order for people to effectively tell you how they found you? What's the setup process like for that self-reporting?
Chris Strom:
Yeah. There's basically two steps. Maybe you could say three steps. The first step is, first off, just asking them. We do that on the web forms by just using a plain text field where they can type in anything, and we make it a required field. Not for everything, but for sales-qualified leads only, we make it a required field. If someone's just signing up for a newsletter or something, we don't require, "How did you hear about us?"
If they're requesting a consult, or booking a meeting, or something like that, in those cases, we make it a required field. It's required, so they have to fill it out. It's also, we use plain text rather than a dropdown so that we're not biasing the answers in any way. We're not influencing their answer by the options available or just influencing them by having them more likely to just choose whatever the first option is. It's plain text so they can put in whatever they want.
Some of the things people will type in are ... I'm looking at some examples here. They'll type, "A friend, my mom used to work there, I live nearby, so-and-so referred me, someone I met, Facebook, email, friend, Google, online." These are some of the real things that people are filling in. That's the first step is logging their answers just in a free-form field. Then the second step is you do need to then standardize it a little bit. We'll look at their answers that they give, and then we'll make a standardized dropdown list of the most common reasons, and then we'll then look at their free-form answers and then start bucketing them into standardized options.
Some of the standardized options include search engine, word-of-mouth, foot traffic, social media, existing or former customer. Sometimes, print advertising if they're doing that. Those are some of the main ones that we bucket it all down into.
Maia Morgan Wells:
Chris, is that standardization process manual always?
Chris Strom:
We started out doing it manually, but we've also been setting up some workflows in HubSpot to help us automate it. We've set up a couple of keyword-based workflows on contact properties, so if the, "How did you hear about us?" free-form field contains terms like Google search or research, tag the dropdown field as search engine. If the free-form field has things like friend, sister, mom, family, then tag them as word-of-mouth. If the free-form field is Facebook, Instagram, IG, FB, LinkedIn maybe, then tag the dropdown field as social media.
Maia Morgan Wells:
Okay.
Chris Strom:
We can automate it a fair amount. Just setting up a couple of keyword-based workflows like that will automate about probably 85% of it. Then we can pretty easily go in and just manually fill in the rest of them.
Maia Morgan Wells:
If somebody wrote in some kind of crazy answer or abbreviation, you just have to do those manually?
Chris Strom:
Yeah. Yeah, sometimes they get confused and they put what they're looking for in the field or something like that. Then you must mark it as an unknown, or maybe if they fill out the field wrong, you can then look at the web analytics as a fallback. Some of these, I'm looking at some of them here ... Sometimes, if we don't understand their answer that they filled out, then we just look at the original source. If they put in something weird in the, "How did you hear about us?" field, but the web analytics say, "Organic search," then we'll probably just mark it as search engine in that case.
Maia Morgan Wells:
You're getting a little bit into how to compare web analytics and self-reported attribution. Do you do much comparison after the fact? Once you've set this up, now you're getting self-reports. You're able to compare that against web analytics in certain ways, which is what you're just telling me. Can you expand on that just a little bit more in terms of what you've discovered in using both of these types of data together, and how that's propelling you forward?
Chris Strom:
Yeah. Yeah, so I was doing some research and analysis on this. I looked at three different clients of ours and compared the results of the web analytics to the results of the self-reported attribution, and I definitely saw some ... I learned new things on both of them, on all three of them. Some of them we've been working with for a long time, but even after all that time, I discovered new things from this that I hadn't seen from many months of just looking at the web analytics.
For the first example I looked at, this client is a plastic surgery and med spa clinic. They were one of the first ones we set this up on. A big learning that I found from there was that social media, for them, was actually twice as effective as we had thought it was.
Maia Morgan Wells:
Wow.
Chris Strom:
Literally, twice as effective. It was the organic social and paid social together in the web analytics showed up as a very low percentage of their total sales-qualified leads they were getting, but when we looked at what people were actually saying, it was more than twice as many. That was a big shocker because they and us had actually concluded that maybe social media actually wasn't really worth the time, and they actually cut their budget on their social media marketing-
Maia Morgan Wells:
Oh, wow.
Chris Strom:
... pretty substantially as a result. Then as a result of this, I found, oh, wow, maybe we shouldn't have done that. I guess we should circle back and invest more in social media, go back to what we had been investing.
Maia Morgan Wells:
Did they do that? What's been the result of that conversation shared with that client?
Chris Strom:
Yeah, we've gotten more serious about social media again and we're putting more effort, and energy, and focus into it. We're running more ads on our end. They hired an internal person in their office to shoot footage and produce content in-house as well because they're in another location from us, so we can do that part. Yeah, they have someone in-house making a lot of organic content now, and then we're spending a lot more time on ad campaigns on our end.
Maia Morgan Wells:
Wow. Sounds like a huge discovery. I'm going to have to check back in with you on that and see if it helped the results even more to reinvest in social. What else did you discover? You said you had three examples that you looked at. Was there something as juicy for the next one? That seems like a really great discovery on the social media thing. What else did you find?
Chris Strom:
Yeah. Yeah, two other examples I have here are from two different senior living communities we work with. The first senior living community, the web analytics were reporting most of their contacts from search engines, organic or paid. Then, a fair amount from direct traffic. Then, some from offline sources, which is where we just manually add them in. That's not super helpful. It tells us that the search engine marketing is going well, but it doesn't tell us much beyond that.
When we compared that against what they were telling us, we found that word-of-mouth was actually a pretty significant driver of their sales-qualified leads. Eyeballing it here, about 20% of their sales-qualified leads were coming from word-of-mouth, friends, or family members, or people like that. Then also, a significant amount maybe combined between existing or former customers coming back, as well as foot traffic or people living nearby. Those two combined probably were about another maybe like 12%. No, probably higher than that. More like 20% of their sales-qualified leads coming in. A lot of word-of-mouth, foot traffic, and then existing customers returning. All three of those categories were invisible to us from the web analytics.
Maia Morgan Wells:
Well, and so does that relate to what you were saying earlier about maybe overreporting on direct traffic?
Chris Strom:
Yeah, yeah, if it's saying ... Since direct traffic is basically people where the web analytics couldn't detect anything, then that could be one of the sources.
Maia Morgan Wells:
Like maybe somebody, having heard about it from their sister, and so they type in the name of the community and it comes up as a search, but really, it was word-of-mouth. Is that what you were saying earlier?
Chris Strom:
Yeah, it might show up as organic search in that case.
Maia Morgan Wells:
Okay.
Chris Strom:
Even organic search shows up as direct traffic sometimes, but yeah, in either case, whether it shows up as organic search or direct traffic, that shouldn't be how it's attributed.
Maia Morgan Wells:
Right, yeah, because the source is the sister.
Chris Strom:
The attribution of the source was the sister, yeah. It was not actually viewable.
Maia Morgan Wells:
Wow. Yeah, this is really interesting. What did you discover about that third example? Now I'm hooked. I want to know what else you found out.
Chris Strom:
Oh, yeah. Yeah, so the third example was a different senior living community. This one is even more dramatic. In their web analytics, between organic search and direct traffic, it reported all of their leads as coming from one of those two sources. 100% from those two. Nothing else.
Maia Morgan Wells:
Wow.
Chris Strom:
Okay, so that's not super helpful, especially because direct traffic was about 35% of that, and that's just where you can't see anything. When we started doing the self-reporting, we found that, in fact, organic or paid search combined was less than half of the actual attribution, so maybe about 40% of their total sales-qualified leads were coming from word-of-mouth specifically.
There were almost as many people saying they heard about them through word-of-mouth as there were people saying they found them through search engines. Then in addition to that, so that was like probably 40%. Then another 20% reported coming from foot traffic or just seeing their sign as they drove by.
Maia Morgan Wells:
Wow.
Chris Strom:
In that case, word-of-mouth and foot traffic combined were actually bringing in more people than all search engine traffic all together.
Maia Morgan Wells:
Wow. That's so interesting. Just even taking that one example, this last one, what kind of adjustments would you make as a demand gen agency or as any type of marketing staff? What kind of adjustments can we make as marketers when we find out this type of information?
Chris Strom:
For the plastic surgery and med spa practice we talked about, they and us have now redoubled and gone back to more of a focus on social media. Then for the senior living communities here, seeing how word-of-mouth is such a significant driver of their business, to the point that it's almost half of the total business, now we're thinking a lot more about customer referral campaigns and customer advocacy campaigns. Still focusing on attracting net new people through SEO and ad campaigns, but also thinking we probably need to put a significant focus on connecting with the current residents and engaging them for advocacy more, and referrals more, and talking through with the client what we can do there as far as engaging and enabling the biggest promoters from their existing residents.
Maia Morgan Wells:
Right. Chris, it seems like self-reporting has been a great source of data. Do you feel like web analytics still has a place in what we're doing as marketers? If so, what is that place and how could we move forward using either both sources, or should we be focusing in on self-reporting only? What would be your opinion on that moving forward?
Chris Strom:
Yeah. I think web analytics themselves are still very useful as well. What I've been experimenting with now is doing more of a blended approach almost where we ... Sometimes, the web analytics don't turn up much, but the self-reporting does. Sometimes, the self-reporting doesn't show that much, but the web analytics do. Sometimes, when we ask people how they found them and they just put website, or they put online, or something like that, in that case, the self-reporting, that doesn't really tell us much, so then we look at the web analytics and we say, "Oh, okay. Paid search," so we'll mark them as search engine. Or, "Okay. They just said online or something, but the web analytics says organic search, or Facebook, or something," and we mark it as search engine.
Then also, for things like search engine ads and Facebook ads specifically where you're running structured campaigns with UTM tags set up, with your campaign labels and things like that, web analytics oftentimes works really well for tracking the details of those paid campaigns. There is value in both, but entirely relying 100% on just one, I found, can lead to blind spots.
Maia Morgan Wells:
Well, Chris, thank you for this really informative discussion about web analytics versus self-reported attribution. I think there's a lot of interesting nuggets we could take away from this conversation. We appreciate you coming back onto the Market Hero podcast. Thanks a lot.
Chris Strom:
Yeah, thanks for doing this. This was fun, Maia.
