Optimization and Refinement
in this chapter, we'll talk about optimization and refinement. There are a number of things that you can test when it comes to google ads. You can test things like personas or audiences. So as you are building out your target personas and target audiences. That's one aspect that you can take for testing. Other things that you can test our copy ad copy itself. If you're running display ads, maybe images. You can also test the different campaign types if you were to go beyond just the search kind of campaign. But perhaps marrying shopping ads versus search ads that drive people directly to product pages, you might want to test scheduling. So maybe it's time of day, maybe it's a specific date, maybe it's a specific seasonality. You can also test things on the home page or landing page as well. So some user experience testing as they arrive into the website from various ads in various campaigns and then finally you can test behavior. So the beautiful thing about digital marketing overall, ...
but specifically when it comes to google ads is you can apply all of these variables into an optimization strategy. If you're looking at the ad in front of you, what are some things, you can test, there's a whole list below that and essentially you can test every single part of this one. Search ad which is really just text. Right? You can test phrasing and language. Maybe you want to test the call to action. Perhaps you test the keywords that people might be using to trigger an ad like this. You can test for various demographics of the audience that might be interacting with the ad. You might test for different devices perhaps looking at how people interact with this on mobile versus desktop. You might also want to test destinations. So you can run the same ad but send them to two different landing pages to different U. R. L. Destinations and see how that performs. You might want to test bidding and budget. So bidding for a specific keyword or phrase at a price point and then maybe seeing if doubling down on that, investing more money into that would generate better results for the business. You can also test geographic locations. So maybe you only want to run this ad in a specific region and see how that performs versus the rest of the world. So a number of things that you can test with. Just a simple search ad. Other things you can test are the destination, your URL or the landing page of where you're sending people. So a number of things that you can test on landing pages would relate back to the relevancy of your ad as well as the relevancy of the keywords. Something's the tests on the landing page could be the call to action. Might be the actual placement of the buttons that you have there for the call to action. You might also want to test language and how that mirrors the ad or maybe describes the product or the service in a way that's related back to the ad or perhaps something completely different. You might also want to test images that they might see when they arrive on the ad. But a number of things and I'll give you some examples of some landing page A. B testing that has been done previously. What you're seeing in front of you is both a control version of a landing page and then a variation of the landing page with a slightly different image. And this was a landing page that was developed during the Obama and biden campaign a few years back the presidential campaign where people who submitted their email address would be entered into a sweepstakes to possibly win the chance to have dinner with President Obama, he wasn't the president then, but the presidential candidate at the time. And so on the left, there's a picture of President Obama on his own with some people around him but you can't see their faces and on the right, you see a group of people engaging with each other with the president um sitting there where you could see people's reactions and what they found was that the variation showing people's responses and reactions performed better. But they also found where that people could imagine themselves, they're sitting there and having dinner with him which made this variation of the landing page more compelling and there it allowed people to want to give over their email address and sign up to be a part of this experience. Another typical A. B testing strategy which happens quite a bit on landing pages is simply changing the color of the call to action button. So what we see here is just the same exact landing page. Everything is the same. The call to action is the same. The placement of all the elements of the, of the landing page are the same. The only key differences. The button color one is green and one is red and based on that it's a viable test to see which color might inspire. A greater engagement rate or higher click through rate once people land on that page in the cases where you might have more than one thing that you want to test, it's called multi variant testing and this is where you would have multiple versions of something that you might be interested in testing. So what you're seeing on this screen now is a variety of different buttons that were used during the presidential campaign. From donate. Now to variations of that same sentiment, please donate, why donate donate and get a gift or contribute. So different calls to action. Different languages used for the same type of call to action with everything else saying the same, the button size, the color and even the placement of the button. And what they found was that depending on how engaged a potential donor was at that moment, people responded in different ways to the button. So if someone had not signed up before donating and getting a gift worked best for them because not only were they asked to donate, but they would get something in response to. These were people that had not get engaged with President Obama or his campaign for people who had signed up but never donated ever before. So they were already on his list, may be interested in supporting him but hadn't actually converted yet. The button that was labeled please donate did the best and it is significantly better than other buttons that were also shown to them and then for the group that had previously donated. So these were existing supporters of his campaign. The button that did the best for this cohort was contribute. And so it's an interesting test to be running, especially if you two have various segments of your business or various ways that folks have engaged with you in the past. It's a good practice to be implementing multi variant testing for landing pages and buttons. But it's also a strategy that you could take when it comes to keywords using the same kind of sentiment. Right? So if the call to action button, please donate worked best for people that had signed up but not actually converted running ads to that same cohort that would use the phrase please donate could also work very well to get them into the site to take that action. So just being able to apply these insights across different stages of the journey as well as different stages of the funnel are helpful. Some other variations of a B testing are to actually switch up the layout of a landing page. What we're looking at now is a control versus variation and a version versus the b version of simcity the game. And so what they found was that when the two options were next to each other without a banner on the top of the landing page, there was a 43% increase in checkouts. So just getting to be able to check out faster, getting to the product faster without actually having to see any other kind of marketing messages after people have arrived was a much more efficient and effective way to get people to purchase the game. Some tools that you can use for optimization are Google optimize, which automatically integrates into google analytics and so with google optimized, you can set your experiments which are essentially just the tests that you be running. You would indicate what you're testing, whether it's an a. B test or a multi variant test, you would indicate where these you RL's lived and exactly what you would be tracking for these tests and google optimized would track that um as well as recommend to you what's performing and how you can improve home pages or landing pages or parts of your website to be able to generate the best results. Another very common tool that's used for A B testing is optimized lea especially as it relates to web sites or even landing page is very similar in the same way where you would use optimized to create a, an experiment or a variation on what exists optimized, we would then track all of that information and report back to you the clear winner. I've actually used optimized lee to test entire websites which is kind of complicated but literally being able to set up an a version of a website completely in a B version of a website and then using a third party platform like optimized lee to be able to look at traffic patterns, behaviors and conversions for an entire micro site or entire website. So these tools are scalable and encouraged for you to use, especially when it comes to figuring out what's going to work best for your business. So a good exercise would be for you to pick something to test that will be in your quiz. I'd love for you to take a look at this ad and figure out what you'd like to test what you'd like to change and maybe a hypothesis of what you think might work better as you are thinking through how to optimize and improve the ad in front of you