English Español Deutsch Français Italiano Português (Brasil) Русский 中文 日本語

Advanced PPC #2: Advanced Analyses to Drive Remarketing Performance




Joel Bondorowsky: Okay. I'm here with Amy, Michelle, and Mark.

What we're talking about today is the advanced analysis that you need in order to drive your remarketing performance. Remarketing as you know, is a very critical part in every paid media strategy. I mean whether it's display remarketing, search marketing or paid social.

From my experience, your marketing could help make or break your campaign after your PPC efforts. It's kind of like a little extra boost that can actually help make the traffic you drive to your site profitable. It can help you pass that threshold to where you can open up your targeting in your campaigns and really scale them. So it's not something to be overlooked at all.

Presenting today we have Amy Bishop. Amy is the owner of Cultivative; it is an agency specialized in paid search, paid social and digital marketing. Amy is a columnist for Search Engine Land.

Also joining us is Michelle Morgan. Michelle is a director of client services at Clix Marketing, which is a PPC agency with expertise in paid search, paid social, display and also remarketing.

And third, we have Mark Gustafson. Mark is the founder of 900 Kings, a digital marketing agency that turns traditional integrated marketing philosophies into multi-channel digital strategies.

Joel Bondorowsky: And Amy, whenever you're ready, you can take over and share your screen and start the presentation.

Amy Bishop: Thanks everybody for joining. Today we're going to talk about analyses to drive remarketing performance. We will cover how to determine which audiences to build based on your data, how to identify which audiences are most worthy of promotion by understanding which have the highest propensity to convert and drive the most value, how to create cross-channel strategies and measure success. And then also, we'll just quickly cover some common pitfalls of remarketing and how to avoid those.

Covering the Buyer’s Journey

So, kicking off, we're going to talk about the general path to purchase. As marketers, we know we have to help prospects follow this path and give them the information that they need as they're following down this journey from awareness of the problem to considering solutions into preference and ultimately purchase, where we hope that they'll become repeat customers for our business.

So, I like to take a customer-centric approach and really think about the information that they need and what they're trying to uncover. So if we think about (people) buying a car. As they're becoming aware, they may just be looking for certain options. We have a vehicle comparison tool, so we can share that with them. And then after they've done that, maybe in the next step, they're still doing some research, so we could share a video with them and continue to help them by providing resources as they move down that journey.



Building Audiences Based on Data

The great thing about this is that we can track each of our resources, so as we're tracking those resources, we can then build audiences off of them as well. So this alone can give a ton of ideas for different audiences that we can build, based upon the resources that we already have available.

In addition to that, we're going to do some analyses to identify other key audiences as well. So if you're doing prospecting, a great place to start is the Google Ads audience and sites report. You can see of the people that have converted, which of those are falling into in-market audiences and affinity audiences. And even if you're not specifically promoting those audiences, these are people that just happen to be converting, but also happen to be part of these audiences. So it gives you a sense of where you may be able to spend more money to drive more value.

You can dig in even deeper in Google Analytics to see more information about these specific audiences, such as the number of transactions that have been driven, the amount of revenue, the average user value. You can really get a sense for which of these might be worth promoting, to try to amplify further.



I also like to dig into demographics. So, instead of just looking at all site demographics, you can dig in and see, you can build a custom segment to say of the people that made a purchase from our website, what do those demographics look like?

And if you wanted to even push that a little bit further, custom reports are a great way to do that. So you can see of the people in this in-market segment, how does that break down between males and females? And if you dig in even deeper, you can see age ranges that performed the best. So if you're starting to build out prospecting campaigns, you can see where the real value lies.

I like to do other Google Analytics custom reports for things like average session duration, page depth, sessions to transaction and session count. Just analyzing things like transactions, revenue, average order value, revenue per user and conversion rates. So you can get a sense for if any of these segments are worth building an audience around, but then in addition to that, if they are, then what is kind of the right benchmark for building that audience?


So if you look here, you can see the revenue per user increases with page depth and there's a certain point where it doesn't really continue to increase exponentially, but looking at between six and seven pages, you see a huge jump from about $9 revenue per user to $32 revenue per user. So that can be an idea for how you might want to structure audiences if you were building an audience around page depth, you may want to look at people that have either a six or seven or more.

Analyzing Different Audiences

So, now that we kind of have a sense for some of the audiences that we want to build, we want to understand how many touches does it take, how many pages are those touches occurring across, what content are they consuming, what topics are they interested in and then what channels are doing the best job of really propelling those audiences further?

So the first thing that we'll look at is lag time. What we see here is that 20% of transactions require multiple days. That's for all audiences, but if we dig in deeper by clicking the conversion segments and then looking at the first interaction being Facebook, what we see is that almost 40% of transactions require multiple days. Because we use Facebook for a lot of prospecting in this situation.

So what you'll see if you start to dig into different channels and sources, is that some of those different sources will require more nurturing and of course, that makes sense if you think back to the original customer journey that we talked about.

So we'll also look at path length and what we'll see in this is all users, not specific to Facebook. 45% of transactions require multiple sessions. And so even though those transactions are often happening in one day, there are people that are leaving the site and then coming back, which is a great case for remarketing.

So as we're thinking about path length and the number of days that people leaving and coming back, we want to think about how do we structure our campaigns to make sure that we're engaging with those? And then also we should be benchmarking this to make sure that we are shortening that path and increasing the velocity. So, of course, we want to dig into that with the different sources as well.

The other thing I suggest is identifying high-value content. So this is just looking at the page report within behavior, the behavior report and looking at the page value of each of these different pages, so you can see how should this inform landing page strategy? And it can also be a great source for audiences as well, just to see which pages are most often associated with purchases.

Site search can be great to help with this too because you can see as people are using site search, what content is missing. So clearly there's either a mismatch in our audience or there's content missing on site, that if we did build it out, we might have a better chance of converting more people.

Audiences and Potential Transaction Value

So next we need to think about which of these audiences really have the highest propensity to transact and then how can we zero-in on how to best convert these audiences? So the first thing that I suggest is leveraging the audience report, it's amazing. In order to do this, you actually have to activate it, so to do that you have to ensure that your demographic reporting is enabled. And then you also have to make sure that you've given Analytics access to your audiences. So you do this the same way that when you create audiences, you give access to Google Ads but you also have to check the box for Google Analytics.


You can only enable 20 audiences for the Google Analytics report. It takes 24 hours to populate and the data is also not retroactive.

So once you do that, you have a high level of which audiences performed best and you can see of those, which have driven the most transactions, which have driven the most revenue and what are the highest e-commerce conversion rates. So which of these if you put money behind them, are the most likely to actually generate transactions? It can give you a good idea where you should focus if your budget is limited.

You can also leverage audiences and custom views.  So we can see of the people in this audience, what content are they consuming, which is most likely to lead them to actually make a purchase?

So if you did all of that, coordinate cross-channel campaigns and also report on results, you could go back to this original mapping that we put together. And the great thing about this is that any channel or campaign can be used to drive the traffic through each of these audiences. And a great place to start is again, by using that audience as a custom segment and making sure that you're looking at which of these channels, which of these campaigns are driving the most value for that particular audience.

And then you'll want to think about audience shaping, to make sure that of those campaigns, anytime somebody has taken an action, we're not taking them backward in the funnel. So we want to make sure that we're excluding anybody that's taken lower funnel actions from higher funnel audiences. So in this situation, if we're driving them to the vehicle comparison tool, we want to make sure that if they've already used it, we're not sending them back to that.

And then, in addition, we'll kind of keep moving down the funnel, if they've already used the video, then we want to make sure that we're not then targeting them and trying to send them back to the video or back higher in the funnel. That should also ensure that we're not sending them back up to the top of the funnel either.

I also recommend setting up custom channel groupings to kind of analyze high-level performance of each of these. I also like to track multiple different metrics for each campaign. I like to track the high-level micro conversion goal volume, so those are each of the things that we talked about when we looked at that table.

So things like the videos that people are watching, content that they're downloading, any tools that we have on site, tracking the volume of those conversions and then the contribution to the subsequent stage, which we can do by looking at the audiences that were built off of those micro conversions, using them as a custom view segment over the next campaigns that are meant to convert them to the next stage. And then also looking at multi-channel attribution and lifetime value of the whole ecosystem.

And by looking at all of those different facets, it allows us then to start to triangulate performance.

I also highly recommend putting together custom reports in dashboards by audience. My suggestion here is that I would build out the custom report or the dashboard that you want to see and then use the audience as a custom view segment overtop of it.

So here we're looking at things like transaction and revenue trends, we're also looking at, of this audience, how does it break down into different in-market segments, how is this audience made up of these different in-market segments? And then down here, what's the revenue per user by different sources and mediums, as they're coming back through these other channels as well?


Remarketing Dos and Don’ts

So just really quickly, we're going to go through some remarketing dos and don'ts.

  • Don't target only all visitors, make sure that you're creating multiple lists based upon different engagement metrics and audiences to see what performs best for you.

  • Don't just target the audiences that some blog somewhere claimed were most valuable. Make sure that you're analyzing your performance and ensuring that the audiences that we've suggested perform best for you. And that's the great thing about analytics, there's so much data there.

  • Don't just blanket exclude audiences based upon time on site. A lot of times it makes more sense to use things like scroll data or other indicators.

  • Don't forget to connect the dots between campaigns. Use micro conversions and audiences to ensure that you're not accidentally sending people backward in the funnel.

  • Don't fragment the audiences too far. So if you don't have a ton of traffic, you really have to be specific about which audiences are going to add the most value based on your traffic.

  • Don't treat all audiences the same. Remember that everybody's kind of in different stages within their journey, so building the audiences based upon those steps in the journey can help you to make sure that you're sending them to the right content. It's going to be the most valuable for them and also have the highest likelihood to create a transaction.

Joel Bondorowsky: That was really good.

I'll start with the one question that's someone asked. Does benchmark data that Google provides for Google Ads apply to remarketing internationally?

Amy Bishop: Good question.

Michelle Morgan: They usually will put something that says where the benchmark data has come from. If the benchmark data that is provided is not generated based on data from the country that you were in, then I personally would take it with a grain of salt. Because we do international advertising and every country performs very differently for whatever reason. Not even just remarketing as a whole, but I mean B2B and B2C can perform very differently, across different countries as well.

Joel Bondorowsky: It is an important topic. In fact, I feel like hopping on Google's Help Center and maybe posting that in their forum and see what comes out of it. Because that is interesting.

Mark Gustafson: And I would say if you have a higher up rep, that they might be able to get you country-specific data.

Excluding Audiences From Campaigns

Joel Bondorowsky: Amanda Smith is asking if you should exclude an audience from other campaigns to make sure you don't cannibalize a specific campaign.

Amy Bishop: I highly recommend adding in audience exclusions to make sure that you're being deliberate about which campaigns are targeting which audiences, for sure. I definitely would suggest that.

Michelle Morgan: The only thing I would say is that if you're targeting two audiences, you want to exclude one from the other, but maybe not vice versa because if somebody is in both if you exclude them both ways..you've made a mutually exclusive list and you've missed both

Joel Bondorowsky: Something I've always done and I'm curious to get your take on this, how do you feel like... I always exclude all visitors from my other campaigns, in other words, if they visited my... if I've driven them to my site before and I don't want my other GDN campaigns competing with my remarketing campaigns. So I always exclude all visitors. Is that something that you guys also do? Would you recommend it?

Amy Bishop: I don't necessarily recommend always excluding previous purchasers, but I do treat them differently. So for instance, if you were thinking about somebody that was leaving the site and then they were coming back and they hadn't ever made a purchase, you may want to incentivize that with a coupon. But you don't want to do that for people that have already purchased, because you might start conditioning people that if they leave and don't buy, then you'll get a coupon and you come back. So I treat those people differently.

And then also people that are in our systems, I try to make sure that we're coordinating with email marketing campaigns, because if we can capture that value through an email, that costs no money, then we want to do that first.

CTR as a Metric for Stopping Marketing

Joel Bondorowsky: You ever look at CTR as a metric for when to stop marketing?

Amy Bishop: That's a good thought. I really haven't done that.

Michelle Morgan: I haven't either because usually we're driving toward conversions, that sort of stuff. I think that the click-through rate can be helpful because if you get somebody to click on an ad for remarketing, they've basically just like refreshed themselves in your marketing list, because they came back to your site again.

So if you're finding that you're running into campaigns that don't get a lot of conversions, but you still need people to keep coming back to the site or even something where you have really long buying cycles, that could be useful or really big purchases that people don't make on the site. Like people aren't going to buy enterprise-level software on a website. They're not going to go and just click purchase.

So it could be something that that could be a good indicator for those much larger purchases or just longer sales cycles in general.

Mark Gustafson: Just like Michelle was saying, depending on how long that buying cycle is, I'll have like probably three or four kind of unique selling propositions that I can pull out of an offer and have specific creatives for those. And if it's a longer process, I might say, okay well, if they haven't taken this action in the last 30 days, let's have them the next 30 days be a very specific campaign. But if it's like shorter buying cycles, I'm just throwing all four at the same time.

Micro Conversions and Actions

Joel Bondorowsky: When you're doing conversion optimization and when you're trying to understand the value of users along your sales funnel, you might also look at micro conversions. Do you find that micro conversions and high-value actions correlate? How different are they?

Amy Bishop: So I think that micro conversions are a good indicator of where somebody is in the funnel, which will then allow you to provide the right content to help get them to the transaction.

Michelle Morgan: If someone actually uses a free calculator tool on your site to see how much time they can save, that's somebody who's actually pretty far in the cycle, probably, trying to figure things out and shopping between solutions.

But somebody who just comes to your website and maybe looks at a solutions page and then leaves, that's still further down the funnel than somebody who just views the homepage, but it's not nearly as far as somebody who uses that tool.

Joel Bondorowsky: Did you guys ever use like Google event tags to create actions? What tags did you create?

Amy Bishop: So things like downloads and I like to be specific about when you're tagging those things, not just tagging download, but specifically what the content piece is. Because ideally you have multiple different content pieces that are throughout the funnel.

Video views, if they've used things like live chat, even pulling in things like the fact that they've called, pulling those things in as an event can be helpful too. Any sort of action resource or tool that you have on your site, that you can track is worth tracking it as an event.

Michelle Morgan: One of the ones that I like, just the way that they use your site, like if they do site search, you can tag them for that. Using Google Tag Manager you can tag people when they fill out, start to fill out part of your form and then don't fill the rest of it out. So you can see not only if that person was interested enough to start filling out the form, but you could probably also then start to see like, okay, first name, last name they were cool with, but then whenever they put in phone number, that's when everybody backs out. You can infer things about your audience from where they abandon the form as well.

Remarketing Tools and Platforms

Joel Bondorowsky: I have a question again from Serkan, how different do you guys think Criteo is from Google Display remarketing?

Michelle Morgan: Personally I haven't really run Criteo in a while and I feel like that mostly is because I have found all the advertising solutions or all the remarketing solutions I need through Google ads. And then other platforms as well. I think somebody asked about other platforms, it's like Facebook, utilize Bing ads, can utilize LinkedIn, all of these different platforms, that sort of thing. They each have their different strengths.

I personally prefer doing everything through Google Display or any of those other platforms, like I said, just because it has all the levers that I want to pull and I get to see all of my data broken down however I want. It lets me play in my sandbox and still target everybody that I need to.

Joel Bondorowsky: So Climb Marketing is asking what experiences do you have with other marketing platforms, pros and cons and what are your favorites?

Mark Gustafson: I've used Criteo earlier in the year and I haven't (used it) recently. And I think it does an okay job and it just depends on how much time you have. If you are dedicated to one account full-time, like for sure and you have the time, pull all the levers.

Amy Bishop: My response is a combination of the two of them. I definitely think it's worthwhile to test different platforms and I've tried a few different platforms for remarketing and trying to dig into additional inventory, things like The Trade Desk and Ad Role.

But to Michelle's point, I typically have seen the best returns through the Google Display Network, as well as some of the others, like Facebook does really well, LinkedIn can be good for B2B as well. So I have generally not been a big proponent of expanding into those platforms, but it's definitely worth the test, especially if you feel like you're maxed out on another platform.

Mark Gustafson: And just consider Criteo's business model, like there's a middleman you're paying, so naturally it's going to be more expensive than if you can do it to the same capability as them.

Joel Bondorowsky: And that's a something about this profession that'll always be there, the actual ad copy and the creative aspect, no matter how good the platforms get with targeting and with the features that they give you, we're always going to have to come up with that and it always makes a difference.

Mark Gustafson: The beauty of this segmentation is crafting specific messaging to these actions, like Amy had that list of all the common pitfalls. I think a big common pitfall is using the same creative and just segmenting everything.

If you have the same creative, sure, you have better insight into the data and you can target better, but I mean you're missing out a ton by not customizing that messaging.

Joel Bondorowsky: How do you decide whether to show different messaging based on time periods? At 30 days, 60 days, 90 days, versus showing multiple messages at the same time. It's a very good question except I would make those time periods much shorter, to single digit instead of 30, 60, 90.

Michelle Morgan: I don't know if I would necessarily agree with you on that because it depends on how big your audiences can be and I think that's the answer to the question anyway, at least from my perspective is that it depends on how big your audiences are. If you can create audiences that are big enough to target multiple of them with single date ranges, then fantastic.

But I have some pretty big sites that have lots of traffic. So the 30, 60, 90 works pretty well.

Searching Your Marketing

Joel Bondorowsky: How do you search your marketing?

Amy Bishop: It depends on the client and how much data that they have, so if they have enough traffic, it can make sense to separate out RLSA campaigns. If they don't have a ton of data, then I just layer on the audiences' observation only, but in an ideal world, serving up separate creative to those people is kind of best-case scenario. But a lot of people don't have enough data to be able to do that.

Michelle Morgan: If you layer in your RLSA audiences into an existing search campaign, you basically just have them in there as an observation layer. You can use what Google Ads calls if statements, that says if somebody is in this audience, show them this message.

You've probably seen those before, just the idea of saying if somebody has seen your video, you can basically reflect the information to them in that copy. It all exists within one ad unit, so it kind of dynamically changes based on the audience.

Mark Gustafson: And I would say that I think search is the only exception into where I would do that. Where you would keep people who've been to your site or taken certain actions in the same campaigns, but honestly, there's a lot of volume issues there. And like it just depends like how much work you're going to put in.

And one thing to keep in mind, like not even in search, but just across the board is like how much are these people sucking of your budget? So that's the real reason why I think from like a social standpoint or a display standpoint, why you need to exclude, is because you don't want these people in like cold audiences, in your cold targeting campaigns and having warm audiences suck it all. Because if you have any kind of automation going, they're going to find those people and it's just going to suck your campaigns dry.

Joel Bondorowsky: You said something earlier about observation-only, if the stats are low, you could watch it, you can make minor adjustments to it. I'm kind of thinking that's a good way to test remarketing lists. In other words, you can create so many remarketing lists, have them running in Adwords, have these observations only to try to figure out how they do before you actually target them on their own.

Remarketing Takeaways for 2019

Joel Bondorowsky: We have a couple of minutes left. So before we close, Amy, Michelle, Mark, I'm very glad to have you, this is like a very good conversation we've been having.

Before we leave, like I'll give you each an opportunity, if you could just give one takeaway for 2019, whether it be about marketing or not, just something people can take with them as we close for the evening or morning for other people.

Amy Bishop: Dig through your data and make sure that you're analyzing to identify the best audiences based upon the data that you have. And then make sure to set up those audience reports, so that you can monitor those.

Michelle Morgan: I'll say on top of that, I think one of the things the idea of understanding what all of your micro-conversions are, what are the different indicators people take on your site, what are the different things they can do? And then where do those actually fit in the funnel?

So take some time and do homework on your own business. What are all the conversion actions you need to track? What are all the ones that you need to make sure that people are engaging with? Which ones are people engaging with? And kind of map those into a funnel and find out if you have any holes that you need to fill.

Mark Gustafson: Every single auction that we can buy ads in is getting more competitive right now. Go back, seriously and check out Amy's deck where she has funnels built out. And the purpose of having them broken down like that is especially if you think of it through a social lens, like you can have very cheap top funnel campaigns and then qualify them to spend like higher CPC or CPMs further down your funnel. And so it's just really helping you be as effective as possible with your budget.

Joel Bondorowsky: Again, I thank every one of you for participating and this was like really a terrific webinar. It felt like we had good chemistry and I hope to be speaking with you all again sometime soon.

Michelle Morgan: Sounds good, thanks everybody.

Amy Bishop: Thank you all.

Mark Gustafson: See you.


Check out other webinars from this series