YouTube Advertising (How to Do) Part 3 Q&A
- YouTube Ad Campaign Targeting Strategies
- In-stream vs Discovery Ads on YouTube
- Google Display Network vs YouTube
- How to Segment Remarketing Lists for YouTube Advertising
- Biggest Recent YouTube Advertising Changes
- The Importance of Testing Different YouTube Ads
- Structuring a New YouTube Ad Campaign
- Closing Thoughts on YouTube Advertising
Joel: Okay, now I think we're live. We're doing a follow up to our other two really previously successful webinars on YouTube webinars with Cory, Joe and also Bryant. Bryant had to kind of drop out at the last minute, he had something going on. However, I think we're still going to have a pretty good show here. So how are you guys doing, Cory and Joe?
Cory: Doing well, excited to be here as always.
Joe: Yeah, we planned on this you know, so, we're pumped.
Joel: First of all we have Joe here, he is Director of Client Strategy for Clix Marketing. Joe is a regular contributor... and has hosted webinars for SEMrush, Unbounce, Quora and Microsoft Ads. He regularly speaks at conferences such as SMX, HeroConf, ConConf and more. He was named a top 25 influential PPC expert by PPC Hero in 2017, 18 and 19.
Cory's also a brilliant expert at this. He's a founder at Variable Media, primarily focused on paid media across all digital platforms. In addition to that, he speaks internationally on paid media, especially focused on YouTube and Instagram, which is cool.
YouTube Ad Campaign Targeting Strategies
Let me start with some discussion. Well, first, let's start with targeting. What do you prefer as targeting for YouTube ads? Let's say you were to start a new YouTube ad campaign, what is the way you would approach that?
Cory: I usually like to set my campaigns up in a funnel structure. You can either have top and low, or maybe top, mid and low in terms of targeting.
At the top I'm really looking at topic, in-market, keeping things very broad. In that mid area, I'm looking for those users who engage. So anyone who went and watched a video on my YouTube channel, maybe they became a subscriber, like a video, commented, my engagers are really going to sit in my mid-funnel.
In the low, I'll usually have your targeting, anybody that added to cart, anybody that filled out a form, anybody that downloaded a white paper. Anybody that is engaged to the highest level that I can find within my clients different attributes I guess you can say. Different ways to engage at a higher level other than just clicking or watching a video on YouTube.
I try to separate things out because I definitely want to compare the analytics of the CPM and the CPCs to start. What you should find is that in your top and your mid, those CPMs and CPCs should be a lot better than maybe your low. Because in your low you might get charged a higher amount for a smaller audience, but they are most likely to possibly convert for you.
Joel: What about very, very low. Besides your targeting, would you ever target individual placements?
Joe: Yeah, I would. But necessarily the placement, if you're doing just placements only without any additional audience targets on it, the placement alone may not necessarily be bottom of the funnel.
From a funnel structure, which I do the same thing that what Cory does, I'm looking at an audience perspective of setting up the funnel and that's what I.
I would want to capture the bottom first, just to keep those users engaged. Definitely my remarketing audiences, and then even when I'm going up to mid or top of funnel, I am creating audience exclusions, just to make sure I'm keeping my remarketing audiences or my engagement audiences excluded from my pure top of funnel ones.
I want to separate the net new users viewing my video ads compared to people who already know. Because if you're doing an in-stream ad and you're adding your call-to-action extensions, that message I want to put in that call-to-action I want to put for my top to mid to bottom-funnel, it's going to differ for those users.
Some of the good top to middle funnel type audiences I like to do are custom intent audiences, which those are going to be merged with custom affinity very, very, very soon. They announced it almost a year ago, but it's actually coming into fruition now. Taking, keyword lists, at least 40 or 50 and then making custom intent audiences, which are pretty much your version of in-market audiences if you don't like the default ones on there.
But with YouTube, it's based off of search queries on Google.com. Broad match related, so it's not exact like you think, but again this is YouTube, so we're okay with a wider reach.
Look at your converting search queries, your converting keywords. Build a list of competitor keywords, and start creating your different custom intent audiences, and test those out. Because we've actually seen sometimes those convert at a higher level than some of our lower ones.
Joel: I have some questions related to this actually, this just opened some stuff up. One guy, this one person here, he's in eCommerce, he's selling a particular product. He's asking, would it be a good idea to target individual placement if it was, let's say a video showing reviews for that type of product?
Cory: Absolutely. I think that getting down to the placement level is very important. If...you do your research and you go through everything and you find those specific placements, I think that's one of the best ways to target. But, if you don't have that yet, and you're kind of unsure, a great way to do it is to go custom intent, like Joe mentioned.
Joe: You have to be careful with placement targeting too. Look within your campaign settings because by default for in-stream ads, Google's going to check the box of "also show my video ad on related placements on the display network".
So if you're picking a specific YouTube channel and that YouTube channel has a corresponding app or corresponding website, Google has the right to show your video ad on the display network too, even if you've only selected video placements on YouTube or YouTube Video. You really need to monitor placement targeting the minute you launch an ad group.
If you don't see specific placements that you want, run your other specific like custom intent or custom affinity audiences for now, look at your placement reports and if you see ones that are good engagement or just flat out conversions, then you know those specific placements have video ad space. Then export those out, and create a different managed placement ad group or campaign for those, and then you know exactly where your ads are going to appear.
Joel: A question related to custom intent keywords. How do you create them? Do you go after competitor domain names? Or you go after just things people search for? And then how do you find those keywords? Do you use the keyword planner, maybe?
Cory: My custom intent audiences are really a mix of everything. I kind of throw things at the wall when it comes to custom intent. There's keywords, there's websites. I do break out a competitor-specific list, and it's got competitor keywords, competitor websites, et cetera. And then I have another list that is more focused on just a category list.
But I'm very diligent in checking that placement report. Because when you set such broad parameters from a competitor and from a category standpoint, you usually get some placements that you're kind of like, “I don't know if I want to run in front of this video”. Or potentially just running in front of music where they could be just listening to my ad and not seeing it. I treat custom intent with a very short leash, in terms of being able to look at those placements and understand where my videos are running.
Joel: One thing that I've always wondered about is how do you know if they're overlapping, or maybe competing with each other. What is the way that you kind of test these lists and make sure that they're not too similar, or what's your strategy there?
Cory: It's all about the placement report for me. The question that you're asking is really me setting a broad net with that custom intent and then checking the performance and seeing how it aligns to somebody who engaged, somebody who was targeted. It's a lot of comparison.
Joe: And that's one thing I typically don't, from a custom intent audience standpoint, I don't add those exclusions. I don't exclude a custom intent audience from another one. Because they could be part of that audience that's looking for my competitor, and also interested in whatever service I could be promoting.
I want to look at the engaged user. And then even from the funnel strategy, once they are engaged, those earned actions that we've mentioned in previous webinars here, people who liked videos, subscribed to your videos, added you to a playlist, whatever. We can create audiences off of those actions, so once they do, there we go, I got them in another list, they're excluded from my top of funnel campaigns and now they're seeing something different.
Competitors; one thing, because when you go and you create these custom intent audiences, you're doing it in the audience manager within Google ads. If it's display and a YouTube custom intent; they're two different ones and they have to be two separate audiences.
When you're creating it from a display one, they actually give you keyword suggestions on the side, so I actually like to go and tinker with it on the display one first, and then once you would go back and create it from search queries from Google.com, which is the YouTube one, they don't give you those recommendations.
You can have different ideas when you look at it from a display ad standpoint first, export that one out, and then copy and paste those keywords in the Google.com one, and then use that one for YouTube.
Joel: Okay, that's very insightful. But actually you brought something up, and I think that was a good point. Excluding audiences is something you have to be very careful with. Because one person could be many, many, many, many things. For example, almost anybody could be a music lover, right? So if you exclude an audience, you could actually be excluding a lot of very qualified people.
It's not like excluding a placement, because a placement is a placement. Placements don't overlap, right?
In-stream vs Discovery Ads on YouTube
Now, what are the pros and cons as you see it in using in-stream or discovery ads?
Cory: Discovery ads are fairly new. But when you set up a discovery campaign, you're kind of limited in terms of where that discovery campaign is going to run. It's either going to run within the news area of Google, Gmail, or potentially YouTube. You can't necessarily silo out those YouTube impressions.
I think what I like about discovery is the fact that you're in a lean-forward scrolling environment. That usually indicates that somebody's willing to engage and probably watch something for longer. I don't think you get that with in-stream where you're running something before somebody's content, or in between somebody's content. Nobody wants to be disturbed, and so that impression is just different.
I think the ability you do have with in-stream is to be able to run long-form content and a variety of different units, whether it's a bumper, whether it's a non-skippable, skippable for whatever length. I feel like you have more targeting abilities, you have the ability to learn more about your in-stream campaigns, and you have video, which gives you a plethora of consumer behavior information.
Joe: With TrueView, the skippable ones, there's a good chance that you might not pay for that view, so that's free impressions there. If someone's on a discovery ad, the most common way is the YouTube search results or recommended videos, that's the most common way you're going to see the discovery one, a user has to click on it. So they're pretty much guaranteed that you will be charged for that view every time, but the intent is typically higher.
If your targeting is great, and you have a how-to video, and you have that video as an answer to someone's question of, "how do I fix this?" And you have the video for it, if you're the top one and they click on it, that user is a very high intent user, and that's where the benefit could happen.
Cory mentioned the experience is totally different, and that's true, where in-stream is going to be more disruptive. They wanted to watch their music video, their cat video and you're getting in front of them with an ad.
With an in-stream one, it could be a call to action extension, your shopping cart, you can send people to your website. With the discovery one, they're on the video watch page. So then you have to look at updating your video with either other recommended videos from your channel or end screens. It's more...to keep them engaged on YouTube, so that intent is different.
Google Display Network vs YouTube
Joel: Okay, that's very, very important to understand. Now, question, actually we're talking about GDN and YouTube there that how the targeting is kind of similar, what would you say the differences are between YouTube and GDN?
Cory: I feel like you have a lot more that you can use in YouTube, just from a video standpoint. Like the engagement metrics there I think are great. You still have ad sequencing which is very unique in terms of, "Hey, they watched this video and this video, wow, that's a high intent user, I want to show them this video because that's my closing video." I think that just because display came first before YouTube it's got a lot more targeting capabilities that YouTube probably doesn't have.
I think Joe brought up a great point earlier, where it's just like, use these tools in Display, export them, and then pump them into YouTube. If you've been with Google for over a few years, you've seen how things get rolled out to other platforms later, but some of those cool tactics usually stay in some of those legacy placements and places.
Joe: Google makes it look like you get all the exact same targeting options for Display and YouTube, and you don't. It goes back to that Display Planner thing. Even when Display Planner was visible, you looked at all the websites, whatever type of topic you were searching under, all of them had display ads because they're part of the Display network, but then only a random few of those actually had video ad space. It's much more specific from the YouTube perspective.
Cory: One more thing when it comes to that YouTube/Display comparison. When you run Display, and this again, doesn't have to do with targeting, you only have the option to click. If you run a YouTube campaign, there's so many other options.
You have the call to action overlay, you have a card where you can include multiple different click-throughs, you have in-screens, you have the ability to re-target somebody who goes to your YouTube channel, they might not even keep engaged on that ad, but if they go take another action somewhere else, you can re-target them. Giving somebody more options to engage and being able to re-target, that's what I think is really powerful about YouTube.
How to Segment Remarketing Lists for YouTube Advertising
Joel: On the topic of remarketing. How do you segment your remarketing lists after your YouTube videos? Obviously, with only 100 visits you can't really segment very much. You just have to market everybody together. But let's say you have hundreds of thousands. What would you do?
Cory: I think the first thing it comes down to, especially if you're running on YouTube is your video content. If you've got a plethora of video, and you've got all different types of video, videos for this product, that product, a founder video, a review video, all these different things, you can really get very siloed and very segmented in terms of your retargeting, specifically through your video.
When it comes to my website, I think Google has done a great job with Google Analytics to bring in very different type of goals. There's a lot of different things that you can create in Analytics that you can bring into YouTube like, I want to target someone that's been on my site for at least four minutes. I want to target only smart goal users. I want to build a lookalike that have been to seven pages and stayed for eight minutes.
Those are the type of segments that I build because I want to compare those unique segments against my typical, add to cart, product view, website visitor. And then once I have those, I want to go ahead and compare that to my video retargeting. People who became an earned view, subscribers, people who like, people who comment. And then people who completed 100% of my video is also very unique.
I take all those different types of retargeting audiences, and there could be people that flow across both of them, and then I start to compare the results. How are the results, how are they clicking through, what are the CPCs, what am I being charged to reach these users? And then finding out what's most valuable to execute after that.
Joel: So basically, you start with a lot of lists and then you kind of narrow it down.
Cory: Yeah. And that's all dependent on how big, right? If a client's getting half a million users to his website all day, it's going to be a lot of analysis.
If we're only getting a thousand users to our website, we might just be using with website visitors, people who stayed longer than three minutes, and people who became an earned view. But it really depends on the content you have, as well as the type of volume that you're driving.
Joe: Yeah, and from a retargeting standpoint, I always like to look at what content users have. I have a lot of clients who they have video content, they're not getting a new one for a while, they're kind of stuck a little bit. So okay, what is the message in this video and who would be the right targeting audience for it?
Doesn't mean you need to blast it out to everybody. If you do have a generic branding video that can apply to any service or any product that you have, then look at one of your retargeting or remarketing users. How can you change the ad? Can you switch up the message in the call to action extension to speak differently to each of the audiences that you're retargeting, what's important to them?
If you're eCommerce and you're doing true view for shopping, you're showing your products on those videos when you're remarketing, it's a generic brand video? Cool. It's the same sale, no matter what you're buying? Cool. But then what products are you remarketing to those users who visited your website?
You can still switch up the video ad, even if you're retargeting, or your main bulk audience and your creative is the same. It's just the message that you put in front of the users when you are retargeting them can still be different, and you can still ad test that out.
Biggest Recent YouTube Advertising Changes
Joel: Okay, that's very insightful, to be honest, and I think it's good for everyone else out there, as well. What are the biggest changes that you've felt with YouTube in 2019? And I remember I think the first time we talked was shortly after YouTube Premium came out, and I was thinking, Okay, maybe this is going to have some sort of influence, because people don't see ads, right?
Cory: I think the biggest change for me on the YouTube platform I've seen in 2019 is non-skippable ads. Non-skippable ads used to be something that we had to reach out to a rep, and buy, and a couple years ago, they were at $40 CPMs only running in front of premium content, and you weren't getting the necessary reporting back on it, right?
Now, being able to buy in-stream in the platform, and get a fair comparison against this 15 non-skippable versus skippable ads, this is where we've seen, oh, this is very interesting, now the CPMs are closer to $10.
The CPCs are still outrageous, but we've got to think about the user experience. They're dealing with this 15 second in between content, we can't expect them to click out, unless we've really done our job within 15 seconds.
This has been the biggest change for us, and we mostly use this 15 second as a frequency tool. It's a great frequency tool to add on to what we're doing from a non-skippable standpoint.
Because in my opinion, the skippable, I'm sorry, the skippable, the true view, skippable, you skip an ad, that is just for longer form content for us. Really engaging somebody, giving somebody the opportunity to skip and say, no, I'll stay with this engagement.
And then adding on top of that, from a retargeting perspective, the non-skippable frequency 15 seconds in between videos to get that reminder in. "Oh, I did see that. Oh, I do need that. Oh, okay."
That to me has been the biggest change along with the implementation of discovery ads, and ad sequencing and a lot of different placements of ad units they've been adding to YouTube.
And one more, I think, big advancement this year that I am so thankful for that I think is going to be so impactful for those in creative specifically, is that if you go into the video area inside of Google Ads, and you click on a video, the analytics has improved, I think, by 1000%.
Before you could only see how long somebody was watching my video. Now you're able to segment how long somebody stayed through the video by platform, so you can see how long they stayed on TV, how long they stayed on their mobile device, how long they stayed on desktop. But also by age, as well as by gender. So I know that, oh, my older audience stays all the way through this video, but my younger audience can't get through it, what's the problem, right?
It develops this conversation and consumer research that ultimately, I think advances your brand that you don't see in the cost that you make in the impressions and the clicks. It's consumer research that I think that drives the better performance in the following weeks and the following months and that's a huge change that I don't think any other platform has.
Joe: Yeah, with the YouTube analytics, too, you get really good device data, as well. If you're doing a ton of content marketing and you're pushing out a lot of videos, you get traffic source data. So you can see if your YouTube videos are embedded on other websites or anything, you get that data of how users are engaging in it, and where your views and stuff are coming from.
And that could be another option for you to realize, oh man, this specific website is running my video, it's getting great engagement, ton of views and everything, maybe I should use that as a managed placement, too. Because that wasn't showing up in my typical Google stats. So that YouTube Studio stuff, yeah, it's been fantastic.
The Importance of Testing Different YouTube Ads
Joel: What about testing ads? What is your strategy behind, let's say, if you're testing something?
Cory: The number one rule that I will tell absolutely anybody: variation. Make sure that you don't just run one video. Please run multiple videos. Even if they're the same length, make them different.
I encourage them to make the beginning different. If you make the end different, not as many people make it to the end of your video than start your video. So it's harder to learn about variation at the end of your video, and much easier to learn about variation at the beginning of your video.
But if you can create multiple links, potentially have multiple actors. Potentially create two different types of video, maybe a funny video and a normal video, this is where you learn about your audience and what potential video you want to create in the future. Now, if the budget's not there, open up your camera, do a face-to-camera, just talk to people. That can always work, we've seen it happen.
Joe: From a discovery standpoint, you have the option, the main part that the user sees is the headline, the description if it's desktop only, and then the custom thumbnail. So those are the elements that you pretty much can only test out.
If you want to do discovery, and you want the higher intent, you want to make sure people are definitely seeing it, test out a variety of different thumbnails and different headlines. What's going to capture that user’s attention that gets higher click through rates on that discovery ad.
Structuring a New YouTube Ad Campaign
Joel: Now they also have another question. How would you suggest to structure a new video campaign from start and through the campaign? Do you set a wide net of topics, et cetera and then optimize for the best performance targeting?
Joe: That might go back to the funnel structure that we already talked about.
Cory: I'd probably mention the whole budget situation, right? If you have a smaller budget, be very meticulous with your choices on audiences. Don't just open up a large custom intent and spend everything in one day.
Be super meticulous and strategize if you're going to create a YouTube campaign. Don't just throw something up there because it's not going to be good, you're not going to be able to analyze it the right way.
Take your time if you have a small budget, and really create a placement list, target the right keywords, and learn from more of a mid to low funnel perspective. But I think if you have the big budgets, one recommendation that I would say is to stick to the funnel strategy as we talked about, but also look at those first two weeks as a great learning time.
You don't need to spend all that budget in two weeks. Spend the first two weeks just analyzing the data. Finding out what audiences work the best, then spend the bulk of that budget against what you learned, maybe with 10% of it, 15% of it, that would probably be a really good strategy for somebody on the higher end, and then somebody else on the lower end.
Joe: Go to the Google, or the YouTube whatever support pages on the targeting on these things and really read everything in the fine print. Because a lot of times Google's not lying to you. Just regular affinity audiences, Google flat out says they are very broad, and they are TV-like demographic audiences.
If you don't have what I always say, a Coca-Cola budget, and you don't want to spread it out to a wide thing? Then don't use affinity audiences as well. They're not trying to pull one over on you. It flat out says it on the description on the Google support pages. So, read how these targeting options can actually work.
If you have a small budget, I'd stick with remarketing and your Custom Intent. Anything with basic topics, basic keywords and basic affinity audiences, they are always way too broad, and you need way more time to let those run and test so you can start excluding out those options before we actually see some of those broader ones work.
Closing Thoughts on YouTube Advertising
Joel: I'm going to ask one last question and then I was thinking for the last few minutes. Any best practices you can think of, let's just blurt them out because I have some ideas.
Now, best practices, I'm thinking about something like, I did this, I think you guys probably do too, but I want to share it and I think it's good for the audience. Exclude your visitors or viewers from your campaigns and just manage them off of remarketing.
Cory: Yes, we do this with our top and our mid-funnel. We definitely want to exclude any of our website visitors regardless of what they do. So our website visitors, our low funnel is removed from our mid funnel and our top funnel all the time on everything that we do. Because we do want to silo that low funnel, and give them the unique experience they deserve because they engaged at a high level.
Joel: What if somebody wants to get thinking about things they should do or shouldn't do?
Cory: I think the biggest best practice I can give is just variation; test your video. The big ability that is not seen on YouTube is the consumer research you get back by seeing how long somebody stays with your video, when they drop off, and what version they potentially like. This version versus that version. I think version testing is my biggest, best practice that I can give.
Please do not build one video and then run it and say YouTube doesn't work. At least run two videos and say, oh wow, this video did better than that one, here's the reasons why.
Joe: Look at yourself. Do you hate seeing the same TV commercial three times in a day? So why the hell would you run the same YouTube ad for six months straight? You've seen it on YouTube when you're on there, the same ad over and over again. Don't do that, if you hate it when it happens to you, why are you doing it to your customers? There is frequency capping so you can cut it off after a certain amount of impressions.
One of the things I would do is you can create audiences and layer them onto your search campaigns. Look at your trends, besides the initial video view, and we already talked about next marketing campaigns, do the research and actually see how YouTube impacts your other channels.
If you're creating audiences that were originally driven from those YouTube campaigns, use UTM parameters, then you can potentially use, if I drove X amount of traffic from my YouTube channel into my call to action extensions, that was tagged with a specific parameter, then I can use that for remarketing on Facebook or anything else, as well.
There is an option to use YouTube in deeper levels in the funnels and look at the value that it actually has beyond just the initial video view. That's my recommendation.
Joel: This has been a lot of fun, and I would like to do it again, we can go on and on and on. And again, every time it keeps getting more interesting.
Joe: We'll do part four.
Joel: Let's do part four, yes. So, it's really been a pleasure. Stay tuned, we're going to do this again. And again, thank you for joining and let's keep the conversation going. It's been a pleasure for all of us.
Joe: Thanks guys.