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

Advanced PPC #8: Making The Google Display Network Work For You




Joel Bondorowsky: Okay. We are live. I'm so excited about this webinar today. Since I started the series a couple months ago, we covered a lot of different topics, a lot of different aspects of PPC. Finally, we're getting to one that we haven't touched on yet. 

I think it's a bit overdue, which is display advertising on the Google Display Network. The Google Display Network is an incredibly powerful tool. It is one that is extremely complicated. 

I think everyone would agree that it's probably one of the most difficult channels of PPC to run but... once you are able to figure it out, once you crack it, the results can be incredible because no other channel of PPC gives you the reach of display advertising.

I have joining with me today Aaron Levy, Sara Ulrich, and also Michelle Morgan. 

Aaron whenever you're ready, just go ahead and go back to sharing your screen and we can start the presentation. It should be about 20, 30 minutes long. Afterwards, we'll get to our question and answer and panel session.

By the way, so Aaron, you got his name down. He's a director at PPC over at a company called Tinuiti. Aaron oversees an SEM team up and down the East Coast. His role over there is to make sure that every team has everything they need in order to succeed and also manage large-scale trends to future-proof strategies.

He's an internationally-recognized speaker and author of a seven-time member of PPC Heroes Most Influential list. Thank you very much, Aaron. 

We also have Sara joining us, who is a proud University of Florida graduate, a Florida Gator, and has been working in digital ever since her career started. She's still and always been based in SEM but she's built out a niche at Tinuiti with the subject or the area of Google Display, YouTube, and Gmail campaigns, which, by the way, have a lot in common. 

We're here talking about the GDN, but a lot of what you do and a lot of it you need to know in order to advertise on YouTube and also Gmail comes from the Google Display Network.

Last but not least, we have Michelle. She is the director of client services at Clix Marketing. Michelle has honed her PPC skills over the last seven years in both in-house and agency roles, managing accounts in many different industries, verticals, and markets. Her passions are around audience targeting and cross channel dynamics between search and social.

With no more of a delay, we're ready to start. Aaron, I think we're now ready for you to go ahead and take control and share your screen.

How Display Fits into the Marketing Funnel

Aaron Levy: I'm going to talk about how display fits into the proverbial marketing funnel. We're going to talk about why clicks aren't maybe the best way to measure success in display and how you weight views. We're going to talk about some of the old ad units that have been around forever that we're all familiar with. Then, Sara is going to walk us through some of the new ones. 

The nice thing about her role here is that she's gotten to alpha or pre-alpha a lot of the things that we're working on, which is one of the benefits of working with big budgets is you get to be a guinea pig. Sara's going to walk us through some of the things that she's learned toying with things like discovery ads, smart campaigns.

A little bit about Tinuiti. We actually already have a question about someone understanding what Tinuiti is. We used to be Elite SEM. We acquired a handful of agencies. Rather than favoring everything towards Elite and calling ourselves Elite Digital or something like that, we chose to merge continuity, acuity, ingenuity, a bunch of other -uities. 

When you're thinking about display, we'll start with the old marketing funnel, something recognizable that everybody loves. It should be more of a cycle than a funnel but that's besides the point. 


You start with awareness. You're top of funnel. You're prospecting. You'll go down to what we'll call interest, that is people have an idea what they want but aren't quite sure yet, down to consideration, down through conversion. Eventually, you have your nice long-term customer that you've paid plenty of money to acquire.

When you think about awareness, the audience doesn't really know anything. They're just starting their journey. They have no idea. They might have a rough idea of what they want but they don't really know how to find it. These are people who are completely new to your company.

You move into the interest phase. That's a customer who has a little bit of a better idea of what they want, but they have no idea where to get it. They don't have an affinity to a brand. They might be familiar with you, they might not. They're weighing their options out. They've done their research but they're not sure what they're going to pick. 

Then, finally, someone buys something. Then we're very happy and then they are your customer, but the reason why I say, "It's a cycle rather than a funnel," is that, in a perfect world, that customer will keep on going and will become your customer for life and they'll keep on buying and they'll keep on renewing.

The big challenge with display and the reason why this is such a fun thing for me to teach as a career SEMer or to point around to other folks is most of what we do only touches the top and the bottom. 

The way that we look at display is it is either prospecting or it is remarketing. That is pretty much the only way we look at it. We kind of ignore the whole middle of the funnel. 

We know the customer journey is complicated and it's never just see something and then buy it, but that to a degree is the way that we treat it. Hopefully, we're going to get you out of some of those old habits and explore some different ways to do it, some different levers.

Measuring Display Ad Effectiveness

Talking about things on only a click basis, that is, only evaluating what you can see in Google Analytics. It's tangible, we love it, we're used to it. We love seeing all of our conversions in one place. It's beautiful but just because it doesn't show up in there doesn't mean it didn't have an effect or didn't help. The battle of figuring out how to measure views is really important.


You're statistically more likely to survive a plane crash than you are to ever click on a display ad. You're more likely to get a full house at random if you're playing poker. I think you're more likely to have two different colored eyes than you are to click on a display ad. Long story short, nobody clicks on these things. 

That doesn't mean they don't work. It just means that action has to be recorded in a different way. The other thing that we usually do is that we like conversion; everything must tie to conversion.

Don't get me wrong. We want to make sure that this stuff works and that you get money out of it but you can't look at these things in the same lens. Sometimes you have to evaluate these things in a different lens. 

Not everything is meant to convert immediately or is meant to have an immediate direct response on your bottom line. Sometimes, it just sort of shoves someone along or push them down the funnel a little bit.

The last thing that's a little bit of a challenge, and we're not going to go in-depth into this too much, but those of you that are doing things programmatically or you're going with a third-party buyer or you're using the AdRoll, the Criteo of the world is we put too much trust in them in the same way that we might look at Facebook and Google together. If someone sees an ad on Facebook and clicks the ad on Google both of them are going to say, "We made that sale," and you only sold one item. So, only one of them gets credit.

If someone is searching for something, they want it. Great. You give them an ad. It's the right person. On display, we're guessing. If you're going based on placement alone or you're going based on a person alone, you might reach them at the right time. You might reach them at the wrong time. You might serve them the wrong ad. You might serve them the right ad.

When you start to think about display, especially on the Google Display Network, you can put it into two categories. All of the major targeting formats are either based on people or places. 


You have to use those to serve me a message when it's in the right context in the right place on the web, making sure that it's either a topic that I care about or it's on a website that I visit frequently that you know that I'll be in the right mind-set or you used keywords to build out a theme. You have to make sure that these two mesh so that you get the right message to the right person at the right time.

Remarketing similar audiences. These are the things that us searchy types tend to love the most because it's the most tangible to us, because retargeting ads tend to get conversions.

You can take these things and put them into individual buckets though. So, you think about me. Interests are, again, the things in the middle. My golf, my inability to find my own car, things like that. Those are interests. Then, you can take attributes, that is what I am or you can take actions, which is what I do.

As you're starting to think about this and starting to think about how you can build out your own campaigns, put your audience into these buckets and figure out what mix of them you would want to reach and you'll have a lot more success than going in alone. 

The higher up in the marketing funnel, if you will, the broader the reach is. When you're thinking about demographics, affinity targeting, that sort of stuff is really close to the top of funnel. 

When you get down to remarketing CRM, that's really close to them becoming a customer. It's a little bit of a different way that you can look at how these may fit into the funnel, but don't forget about things like in-market or similar audiences or custom intent. Those are that sort of middle funnel area where people are aware of you but aren't ready to buy yet. 

Fun fact: The in-market audiences were originally built for the automotive industry because that has a very predictable long-term purchase path. Then they have a pretty good idea if someone's about to buy a car or wants to buy a car at some point. 

Moving Away from Clicks

Now that we have a really brief overview, let's talk about tracking and, to a lesser extent, bidding and why you want to get away from being “clicky”. 


Okay. This is a stat from a site called Banner Blindness. Of the people that click on display ads, I'm not sure how many of you are familiar with the Pareto principle but 20% of the effort is responsible for 80% of the results. It's even more dramatic for click distribution. 

Something like 10% of people are responsible for 90% of the clicks. They tend to skew older. Yes, it's usually your parents. Of those people, less than a third of them or more than a third of them only click five ads a month and 50% of people will never click a display ad. 

If you're optimizing to clicks alone, you are essentially optimizing towards 10% of the population, which is not enough people for success in scale.


When you think about these ads and you look at the ones that I was served, the three in the middle are all from Google's responsive ad builder. All of them are very click-focused because that is what Google wants to condition us towards because that's how Google works the best. 

I'm not going to say that one way is right or wrong but it's really just an interesting thing to think of where you think of display is usually sold on a CPM basis and we're buying it on a cost-per-click basis and Google is skewing it towards clicks.

Now, I'm not going to go too in-depth into bidding because most people will talk about, "We'll buy it on CPC through the Google Display Network." That's generally fine. 

I did want to touch on Google's pay per conversion. I guess the beta, it's public now, but it's one of those things that sounds so perfect on paper, like, "I only pay Google when things convert." That sounds amazing for me but Google's not going to run that when they lose. 

What we usually find when we're running that, is it actually winds up being less efficient than it would be if we were running smart bidding on its own and optimizing on our own. Google has offered it as a way for people to break in display because they feel a little bit more comfortable, but in terms of scale, it just doesn't work. 

I'm not going to go too in-depth into this but since we're in a mobile world... when you're thinking about how you would weight views, mobile, especially tablets, most of the conversions tend to be view-based because it's really hard to click on links or little banners or things like that. When you think about how to optimize, you'll have to look into views a lot more for mobile.

I'm sure a lot of you are like, "Okay. Well, do I count every view? If nobody touches an ad, can I count it? Do I trust everything that it counts?" The answer is no. When you're thinking about view-throughs, you'll have to do some sort of a weighting element that is there's a percentage of them that you can count versus what would have happened anyway. 

Sara's going to talk through some of the bare-bones basics and some of the things that she's done with some ad units. 

Responsive Display Ads and Other Ad Types

Sara Ulrich: Thanks, Aaron. Looking at the responsive display ads, these resize to fit any format and they claim that they will serve in the format that the user has previously converted on.  Whether it's text, image, and native, they're taking that into consideration, along with where will this ad look best, how does it fit on the page. These responsive ads are pretty powerful in that they can serve a user and meet them halfway with what they're looking to see theoretically, according to the algorithm.

When they first released them, I think everyone was running static ads and really afraid of this kind of dynamic form of serving but I typically advise against static ads now unless you have the best creative team in the industry. Just because these responsive display ads have so much more automation behind them that is proven to work and to increase click-through rates. 

Gmail ads; my favorite. These have evolved a lot over the years. They used to be in a pretty strict template where you could just open to a static image. Now you can kind of put in your image, maybe even a carousel catalog. You can do dynamic. 

There's lots of options for what it looks like when the user opens your email. Now, you can also customize your teaser headlines so when you see it in your inbox, it might say something different than when you open it and it's in full-screen. 

Cool, so Google gallery ads showcasing. They're constantly releasing ways to upload more assets within one ad set. A lot of these new ads, specifically the gallery ads, are allowing up to 10 ad cards where you can have multiple images and multiple pieces of language set in different cards and the user can just scroll through them or swipe through them no matter what the platform is. 

I think this might see an immediate decline in CTR but an increase in conversion rate because you're able to better inform those users and kind of eliminate the wasted spend, if you're looking at things from a click perspective, and if you're looking at it from a view perspective, maybe they're making their decision a little bit more quickly because they're presented with more information.

Discovery ads, so these are still, I would say, pretty new. I don't think it's quite as business as usual to run these amongst your other Google remarketing or Google prospecting campaigns but they are not responsive ads even though they take all the same assets. Basically, if you're running a responsive display ad, RDA, you can take that and repurpose it into a discovery ad from an asset perspective. 

You can serve on YouTube inventory on the homepage within Gmail or then on the Google feed, which is catered to people's searches and interests because it's on the browser on their phone.

It does tend to really not look like an ad. When you're scrolling through your Google feed, if you have five news articles about higher education, and then you see an ad for higher education, it blends in pretty seamlessly.

Avoiding Unwanted Placements on the Google Display Network

Aaron Levy: One of the biggest challenges with display and the Google Display Network is, in theory, anyone can be on there. So that means that there's a lot of good. There's a lot of not so good.


A lot of what we would like to do is get rid of some of this junk before it appears. Now, it obviously depends on how sensitive your brand is, but Google gives you plenty of exclusions of categories to exclude ahead of time.

Depending on your brand, it may be better just to exclude any and all political stuff. Yes, you may be constricting some of your volume but it can save you a lot of heartache if you're put next to a story that you don't want to be a part of. 

The other things that I want to look for when I'm doing a placement report is, (this seems a little counterintuitive), but I'm always skeptical of anything with an abnormally high click-through rate. If you see something that has a 10, 20, 30% click-through rate, that's not normal. That's probably spam, so you might want to get your ad out of there, save a little bit of money.

Then, obviously, look for high CPC or foreign language sites that don't have anything to do with your brand.

Last but not least is apps. In my opinion and I think most of us here at Tinuiti will agree is that you should treat apps the same way that you would treat a placement. Don't look at them as bad as a category because that's where the vast majority of people's time is spent on phones. 

Look at them the same way that you would look at your placements. There are categories of apps that you don't want to be on. There are apps that you do want to be on. I think the stat is something like 95% of American's time on their phones is on an app. 

If you get rid of them all, again you're getting rid of most of the market, but if you find a way to optimize them, you'll be in a place where in you're really good position to win.


Responsive Ads with Video: Overview and Best Practices

Joel Bondorowsky: I think I will start with this one question I have regarding responsive ads. "Recently, I started noticing this new feature that Google has. They're now offering video inside of responsive ads. I mean, have you tried that? Also, where does the video show?"

Sara Ulrich: Responsive display ads do allow video now. You just link directly from your YouTube channel, so it's the same way that you would run a YouTube ad. You can just put the asset inside the RDA template. They show exactly where the image would show, so wherever that landscape chunk is of your ad where your static image would be, the video does play right there. 

From what I've run with these, performance was not necessarily better, wasn't really necessarily worse, either. I think it can be more informative for someone who is taking views into consideration and hoping to move this user down the funnel. 

I think if you are looking for last click delivery on this ad, maybe people don't have quite the patience that they would for static where they see it as a little bit spammy if they're scrolling through a browser and something starts automatically playing to them.

Joel Bondorowsky: Yeah. Okay. Any best practices for these video ads as far as length?

Sara Ulrich: Keep them pretty short. You don't want like a 30-second because you won't see the whole thing, obviously. Ideally, they would look more like a bumper ad between 6 and 15 seconds. I would run them alongside static just because that's kind of tried-and-true, in case your users don't respond well to live stream on their browser. I would just make sure that you have some static images in there as well.

Smart Campaigns Overview

Joel Bondorowsky: Okay. Great, great. Now, here's another very good question regarding smart campaigns. "I mean smart campaigns came out recently. From what I've seen, by the way, is they just target a lot of stuff really quickly. I mean are they really, really that smart or is it still good to handle things manually? What are your thoughts?"

Michelle Morgan: We did run some campaigns with smart display campaigns when they first came out. It was one of those things where we went into it knowing that it was going to be very much a learning process and that we were going to have to give it some time to figure out what we wanted to do and where it needed to serve things to actually get the right results that we wanted. 

Honestly, we were in it for, I want to say maybe three months or something. By the end of it, we were hitting pretty much exactly at the target CPA that we wanted. It was a lead gen campaign that we had. The quality of leads that were coming from it were not really any different than the leads that we were getting other places.

I think the biggest thing around smart campaigns is that I would say it's the same thing that Aaron was saying about one of the other campaign types is that it can be part of a strategy, but I certainly wouldn't make it the entirety of it. Use it. Let Google learn a bit, but make sure that you set it up for success. Our target CPA was a hundred bucks. 

Aaron Levy: First of all, I agree with everything that you said. Budget size aside, the best practices are the same. The only other thing that I'll add is a lot of folks were and a lot of our clients are really excited where it's like, "Oh, Google will do the bidding for us. Let's just set the CPA really low and see what it'll do." I'll tell you what it'll do. It will turn itself off.

Optimizing Google Ad Campaigns

Joel Bondorowsky: So, let's talk about “dumb” (regular) campaigns for a second. Let's see. Once you've launched a “dumb” campaign, you're using combinations of audiences, topics, all the different features they give you. How often do you touch it once you launch it and what sort of maintenance do you do? What sort of things do you adjust?

Sara Ulrich: “Dumb” campaigns you touch a little more often. I would say if they are on automated bidding, like a TCPA (Target CPA), you are still waiting at least a week before you're bringing your TCPA go up or down based on volume, bring it up if you want more, down if you want less but ultimately trying to hit that goal that you've set for yourself for that campaign. 

Aaron touched on the manual placement exclusions. I think that the system is only so smart in excluding entire categories of content. You do have to comb through those and make sure that you're not serving on sweepstakes pages or anything that looks spammy and maybe has a few too many leads under its belt. 

Joel Bondorowsky: Right, because, I mean, let's not forget, the one thing that they can never automate are the ads and the creative message is only going to come from a creative person.

Sara Ulrich: Right. Totally.

Michelle Morgan: Yeah. In terms of the frequency and kind of changing things, I would absolutely agree that any of the algorithmic learning, smart bidding, that kind of stuff; give it time. If you start moving that around real quick, it's never going to learn anything. You're not teaching it what to do. You're changing everything and it's constantly trying to adjust.

The only other thing I would say is I usually treat it like basically just like a brand new launch anywhere else. The first few days, first couple or three days, I'll log in and just make sure that everything looks normal and okay because sometimes you can launch something in one of those placements that's just total garbage, will take up all of your budget. You got to go in and just make sure you're culling those out relatively quickly. 

Once you've got the kind of baseline things out of the way, if you weren't able to proactively add in negative placements, that kind of stuff from other campaigns, I think the first couple days are just making sure that everything looks smooth and nothing looks ridiculous. Then, after that, giving it time to learn and gathering more data to make decisions on.

Running Gmail Sponsored Promotions (GSP) Campaigns 

Joel Bondorowsky: Gillian Murphy was asking regarding GSP campaigns. She's asking about how Google manages Gmail ads with Gmail versus as in the GDN Network.

Sometimes, what I think a lot of people see is that they launch their campaigns and it ends up getting dominated by really, really, really, really clickbaity placements and maybe the ads are appearing on Gmail a lot. 

It's just kind of dominating everything. You're not getting any of the really, really good placements. So, what are you doing? Do you like segment them out? Do you have like one campaign or, sorry, one group maybe or one campaign is targeting mail.google.com as a placement while another one is targeting the apps in order to, let's say, get the good stuff running?

Personally, responsive ads, they're great, I'll tell you what, and you're very correct about how they work really, really well but sometimes there's certain banner sizes that only work best if you have a very creative person design the ad. Let's say that 300 by 600 size. It never looks good in responsive.

So, what I'm asking is about what sort of things do you segment when you watch?

Sara Ulrich: Yeah. I think if you want to run within Gmail, the mail google.com placement, it's best to negate that and your Google Display campaigns and then run a GSP campaign because that way you have more control over your Gmail ad. You're curating it to be a Gmail ad. 

You know what assets are in which spot, whereas if you are targeting that placement in a regular display campaign, it's going to take all those assets and serve it. It might not be exactly how you wanted it to look or how it could look in its best format.

I would say if you have the budget to do both, do both and definitely separate them to where mail.google.com is exclusively a GSP campaign.

Aaron Levy: If there's a place that dynamic retargeting makes no sense, it's in Gmail because the subject line is okay, it's Nike shoes size 9W black whatever is the subject line. I'm like, "That doesn't help." The ad unit itself looks pretty good but considering its feed-based and it's pulling from a product feed relative to other email subject lines, it feels a little weird.

Joel Bondorowsky: Yes. Okay. We're running a little over. We can continue this on Twitter, by the way. #SEMrushLive is the hashtag. I'm very happy that everyone joined. Thank you, Aaron, Michele, Sara. It's great having you. 

Any one last trick or tidbit you want to give anybody before we sign off?

Aaron Levy: I will give one answer to a question that I saw but we didn't get a chance to because it's hyper-niche. Someone asked about CBD or hemp products. There are networks for that. Google is not one of them. 

Generally speaking, anything controversial Google just says, "Nope. Not going to mess with it," but there's a couple networks out there, so if you just do a little bit of light Googling, they're pretty lightly used. They're worth it.

I'd say my one tip in spite of what we just said about Mr. Hydrochemical, is when you're doing display, think about people. Don't think about sites, don't think about websites, think about people and what you're trying to tell them, what you're trying to get them to do. That will make things make a lot more sense.


Check out other webinars from this series