Unknown Speaker 0:10
So thanks, Fred. And I’m gonna flip the stage now and take over a little bit. So everyone, yes, for those of you who’ve never heard of Brad, this is Brad, he goes way, way back in PPC. Usually it can’t shake, shake him off my tail when we go to conferences, right, but he’s gonna be amazing. So stick around. Yeah, it’ll
Unknown Speaker 0:27
be fun heavy fun.
Unknown Speaker 0:28
Thank you, Fred. All right, well, we’ll chat in 25 minutes. All right, so I’m going to finish off. So I’m Brad Geddes. If you don’t know me, I’ve been doing paid search for way too many years these days. And I’m going to get into ad testing as soon as zoom decides which screen I should be showing. Alright, and so with that, let me just readjust. Oh, zoom, you’re a funny creature.
Unknown Speaker 0:57
There we go. Alright.
Unknown Speaker 0:59
Um, so when we look at why people test and what they’re doing, right, mostly, we see people test because they’re supposed to, or because our agency says they test monthly, whatnot. And it’s not always testing to stay ahead of competition, or to really find interesting insights. So this was an advertiser last year, right? When, when when COVID first came down. And while this was, you know, last year’s data, it’s still very applicable 10 years ago, or next year, when there are external events, which significantly change what you need to message. And it doesn’t matter if this is the Superbowl spring break, travel seasons back on whatever it is, write events happen, that change how people either view your ads or view the offers, suddenly, your offer doesn’t look as good next to a competitor’s offer. Right? And so it’s not just about, hey, we need to test to see, does this message work better than this one? is also this understanding? What’s the universe that’s happening around us? And how does our offer fit into that universe? So we think about, you know why most people test its new products or launched those changes to the business, because we’re supposed to do it. And some will do it, because reaction to data trends, assuming you track your data trends, right, others, if you’re tracking competitors, that will do it in change your competitors ads. And finally, you have those who say, you know what, we watch these things, and we react to them, we also just test to stay ahead of everybody else and be the market leader. So when we see this, right, we see, okay, we have a big client, and they haven’t changed their ads in a while, and they just want something new, or a competitor made a big change, right? Step one, is often to determine the actual competition, right? I’m sure you’ve all had clients where they say, yeah, yeah, we want this goal, but you must be Y, whoever y happens to be your x. And you you often go to auction insights and say, well, they’re showing up like 3% of the time. So you can use you know, Google, this is a one of my favorite Google Data Studio templates, where you can see every campaign if you’re broken up by geo, or time of day, you can even see who are your competitors for every campaign? Or we asked if it’s a larger company, give us like an ad feed or something to really dig into first, who should you pay attention to? Because if you just go to the SERPs and look at it, a lot of times you’re not seeing who is the actual company you need to pay attention to. Right, so that’s one of those first steps. Second, then, is really going through in examining those ads. If you’re using a third party system, you might already have a list of all of them in there, and then start taking notes, right of what’s unique about certain ads, what really stands out about the ads? Is someone being really creative in their ad extensions. Right? How many people are using headline threes? I’m still amazed at how often we see some pretty expensive industries. And no one has a headline three, create a jet. And it’s not that hard to create them. And they’re very useful. They they get difference in your ads. But it’s very common to not see them still. Right. So it’s okay, here’s our notes and our checklist of things you want to think about. Right? Then always and I think Brooke and Fred, everyone sort of mentioned the funnel stage a few times, right? But then we go back to the funnel stage and we say where are they in this funnel stage? Now once we have that part done, right now we want to say Alright, here’s where we’re trying to advertise. And the reason we want to go back to the funnel stage is we want to say, what are our major considerations? Are we thinking about benefits and features? Should we be thinking more? Alright, this is awareness, they don’t know us yet sweet and want to talk about the benefit of our product, or this is purely on Amazon. This is other a bottom of the funnel, therefore, we really want to focus in on the retention standpoint, or do we need to get purely into features?
Unknown Speaker 5:42
Now beyond that, we want to say, Okay, how strong you want to be with our calls to action. So if we’re at the very top of the funnel, we can’t be like, buy this now. If it’s at the bottom of the funnel, right, now, we can be a lot more straightforward, like, call us schedule appointment. So you want to make sure you’re looking at that funnel and matching your CTA matching your your benefits features with the funnel. So we say what’s a common ad layer? Right? It’s often related to ad group, headline to benefit. CTA headline three supports headline two, etc. Right, pretty common starting place. And the reason this is a good starting place may not be your ending place a good starting place, is we’ve done this study four different times now, where we go through and we have people figure out, what’s the benefit? What’s the feature? What’s the mixture, that makes a huge difference in how the ads perform. And we very commonly see that mixing calls to action and benefits lead the pack as far as what’s the best conversion per impression, which is why we often say this is how we want to, you know, start with the ads. Now, this gets a Sunday. Alright, so we want this kind of a CTA and this kind of a benefit at this stage of the funnel to figure out what our message should be because of X Factor change. Right? So then we say, what do we test? Right? What are we doing? And this is when we often ask those big questions. Should be include price, what CTA should be used, or you’ve tried three benefits, what if you say in authority statements 100,000 served? Right? What happens if we want service versus price? Right? If for Gucci, we want to talk about luxury and exclusivity. If we’re selling something that’s sold everywhere play want to focus on a really low price. Like if we’re in b2b, it’s how do we talk to the business people while telling the consumers don’t click on our ad go away? Right? So that’s that big step? And that’s a sort of a bigger picture. Then, where do we want to test? Right? What source segments? And how are we going to do this? Do we need to segment by device? Sometimes not? Sometimes it’s a definite must. We never want to combine ad types, of course. But then, if we’re going to test this across our brand, and third party brands and info, words and locations, right, we want to make sure, even though we may set it up that way, we’re reading that data differently. So now we have an idea of what to test where to test how to test it. Then the last question is, are we doing single versus multi agric? testing? Let’s so single Europe testing says we’re putting two or more ads in an ad group. And we’re gonna find the best message for the targeting method of the ad group, be it a keyword an audience, whatnot. Now, if it’s a big brand term, that’s the way to do it, because you don’t want to mess with your brain with you don’t want to co mingle with other data. If we’re saying, this is for 5000, longtail ad groups or 3000 different product terms, we might do a call multigrip testing, which is we’re testing price versus not
Unknown Speaker 9:15
price. What if we have location versus not location, right? We’re testing that hypothesis across a large selection of ads. In this particular example, it was a company who had a crazy amount of keyword insertion, and they’re like, Google, our reps always tell us use keyword insertion, the keywords are always there. So just wanted to say well, and doublex by SharePoint means does not include. One is say, Well, what happens to your data if you have keyword insertion in the ad, versus you do not have keyword insertion yet? Right, that it’s a pretty simple question. And across some and this is, you know, over 4000 ad groups, they have significantly better data without keyword assertion. not always true, right? Sometimes You find it better. But that’s that bigger level question that’s really useful to test. And so I love multigrip testing, because it tells you, this is not the best ad. So for an ad group, it’s the best for this entire section of ad groups. Now, this is not the way to test for your most important aggregates, you have an ad group your boss watches every day, right? You know, it’s one of those are you don’t run this in that right? This is where you don’t want that bigger insight. And the reason we like this is you say, where do we use this? Well, of course, you know, in your paid search, ads, email, subject lines, email buttons, title tags, we have customers, we’re a testing platform and recommendation engine, we have customers who are only SEO companies who just plug in the Google Ads data sometimes, you know, run by different agency, just to test for title tags, right landing page testing, Image button, social CTAs. If you’re doing TV buys, how does that TV ad fit in with your search ads? Right? So there’s a lot of good ways to use that data. When we say, right, what are you trying to do, and that’ll affect the metric you want to test by. So in many people test by CTR is the most common test that exists. It is a horrible, horrible test. I mean, Fred mentioned earlier, use proxies for data, even if you don’t have actual revenue, use proxies. Right? One of those that that will often use is just a two page visit, or a visit over 60 seconds. And that can be an event, which means it can be a conversion, which means you can now do conversion per impression testing or conversion rate testing, not on everybody which Ctr. If they 100% people click and harbor some people bounce, ah, your winner, right horrible winner. If you say now, this is an engaged user, we’re going to test by right just by doing some sort of on site engagement process as a conversion. Now you can actually test by real data again and use it. So there is no best testing metric, right? It’s more What are you trying to accomplish? So and the reason why I feel these steps is, here’s some tests we’ve seen recently do well, one, a large brain, I’d say a large brand, usually over 5 million in AD spending a month, they do this test, every six months, they have a very large branded departments. And the brand apartment always pushes us our brand and every one of your ads, we don’t care what the keyword is. And so but every six months they test and they almost always find not putting our brand in non branded keywords is better for our conversion rates better for our profit. Yet, we have a different company, another large one, they do this also same test, but every six months. And they find that putting their brand in non branded ads increases their conversions and revenue. Now both of them are big TV spenders, both of them spend a lot on on on search. They have very different industries, right? And so you should never look and say, Hey, this person says always use branded non brand or never use branded non brand. You should do it yourself. Right? These are suggestions of what to test. Because we will commonly find people who either always use or never use brand and ads. And instead of saying this is how we do it, we should be saying what is our data when we do it each direction right each way, which is a good way of doing some good multigrip test.
Unknown Speaker 13:54
We are still seeing even today that uplifting and positive messages are just beating apart the we’re here for you what unprecedented and things we’ve heard way too much of the past year that anything a positive people are looking for something positive right now. And adding positive does incredibly well these days. Expect next year. That won’t be true as true. Right? safety is still a huge concern for everyone. It may not be your leading headline anywhere anymore. But what did you do to keep employees safe? What do you do to keep visitors to your business safe assuming you have visitors ecommerce course quite different, right but safety is still doing really well. And humorous and colorful message right now are really doing well. Again, something that stands out a little bit different. Most ads are headline one the keyword headline To best prices best, whatever, headline three, call us by now shaking up but formula is really used for standing out. And at a time where you know we’re talking was earlier today, we see the same walls every day we see the same stuff outside every day we see the same people every day. There’s very little variety in a lot of people’s lives right now. So searching and seeing something that stands out that’s really different is causing really good interactions right now. So things to think about with your with your testing. Now. If we’re doing responsive search ads, right, one of the big things to look at is, are they winning, and they will win? Sometimes they will be your best at times. But to if they’re not your best ad, what are they doing on impression basis? Right. And we find that if an RSA, even if it’s been in the Eiger for six months, and Google should have learned this by now, right? If an RSA is one of your worst performing ads, it will often still have by far the most impressions. And so with our essays, responsive search ads, they can be useful. I don’t want to give the impression they’re not. But watch them, they still need headlines related to the agroup. And I’ll get into that in a second. Right? Because then this account and in many accounts, right? pausing there, RSA led to more than 1000 new conversions in a month, right, just by getting rid of the RSA. No, again, it wasn’t doing well, that’s not always true. So often, we want this sort of big snapshot, or we say, Alright, here are ad groups that have an RSA in them. Which one is RSA winning and impressions and conversions? Or is our ETS winning, and getting that bigger level picture of how they’re performing across the account? Because sometimes we’ll see an accountant are saying you’re doing quite well. When they are doing quite well, it is they have one really common theme to them. And that is generally the creation. So we did as we’ve done a couple studies looking at accounts with responsive search ads. And in many cases, it is the exact same responsive search ad in every ad group in a campaign. Well, search again, is about saying, Hey, we know what you want, right? Like you searched weaken. We want this come here, we know what you want. So when you have the same RSA in every ad group, you lose the relevancy of that headline one, which is often so unique to the ad group, saying we understand what you want Come over here. Right. So when you create responsive search ads, you might use the same call to action across all the ads in that group. You might even use the same benefit statements, authority messages, whatnot, thinking about how do we make the responsive search ad relevant to the keywords in the ad group, it’s so important to just make sure you add a couple lines that are say, that are related to the keywords why someone did the search in the first place. And it’s it sounds so obvious when you talk about it. But Google’s made this thing our sales are so easy, just make them and walk away your don’t do anything ads anymore. That it it often loses what makes search Amazing, right, which is I looked for this, I saw so and so has this great there’s a match, there’s a conversion. And so just be careful about you reusing the RSA across a lot of disparate ad groups, and losing the relevancy of the user. In addition, RSA is can have 1000s upon 1000s of combinations. When our essays generally are doing really poor in an ad group. The second most common reason is the ad group has two or three RSS in it. So let’s say you get 10,000 clicks a month in an aggregate that’s a that’s a pretty well traffic died group right. And one RSA with 5000 combinations, it could get two impressions for each combination. Now you add a second RSA with eight headlines, and suddenly each could get one impression or combination, you add a third one, and now you’re talking there’s over 100,000 combinations of what Google could show for an individual impression. It messes up their machine learning rate. So if you’re going to use our SES, only one per ad group, I can’t say Started off every time we see two or three, something just goes wonky with the ad serving over time. So just you know, some things to think about. And then we’ll get to q&a with everyone write test, not just because you’re supposed to, or because the agency doesn’t monthly write, test to stay ahead of competition to know your data, when you see data change, and it’s generally a negative, right? Look to the ads and what the competitors have done. External factors affect ads so much, right? Because I especially now, right, we’re, everyone’s sort of expects everything to be up to date, real time. If your ads are from a year ago, and there are many ads running today that are from two or three years ago, right? They’re probably not relevant, they’re relevant, but they’re not as good as they could be from aesthetic standpoint. Right? So look at the competition, look at the ads, make your notes. Always consider the funnel, right? If someone does not know your brand, saying, buy this now, it’s often a mismatch, right? isn’t gonna do as well. So think about you’re here in the funnel, let’s meet you here. Let’s give you a path to next step. Right? Then it’s alright, are we doing single leg testing or multigrip? testing? Do we want to learn it for a single agroup, which is often important, or across a large swath? And then learn from your data? Right? We see very successful marketing departments. They share information, we always talk about silos, right? And breaking down silos. There’s so much information in your ads you can share with other groups, right? What is making a big difference in your metric and help your other team write title tags, or headlines, your content teams are so much data in them that other people can learn from? Alright. And then responsive search ads. We have seen a lot of big brands pull back from them significantly. But they’re they’re probably gonna get better over time. And Google’s machine learning gets better every month. Right? So it can be really good for CTR, they’re generally lose about conversion rate, but hopefully someday, that’ll change. Let them run long enough to judge their success. You can’t run them for a week. There’s not enough data there. Right? But if they’re not performing, don’t be afraid to remove them and try again later. Right. We always pare back what doesn’t work. And then we start we need to try this again and see what happens. If you’re looking for software like Fred, we’re a software company that helps a lot of these things. And with that, I will say thank you. I am going to stop sharing my screen.
Transcribed by https://otter.ai