What does it take to write effective ads? What are methods to measure performance in e-commerce or B2B? What do we look for when hiring a PPC agency? Andrew Goodman, author, speaker, and founder of Page Zero Media shares best practices with OMCP:
What are some best practices for writing effective digital ads?
Well the place to start is to understand that this is a test, in a testing environment. You’re rotating the ads, and so it’s not about just the ideas that go into the ads, its about how to structure the tests so that you actually understand which one won. So from there, we’re going to have to pick which key performance indicator we’re testing to.
I recommend that if you got revenue data, you be a little bit worried about that and don’t necessarily test to return on ad spend. Because order size in let’s say, e-commerce, can be spiky. So spikiness can ruin your test in the sense that it could be a random outcome. So you know, something like conversion rate, the highest converting ad, is one of your key performance indicators. At the same time, quality score gives you better results, meaning a huge component of quality score is click-through rate, ad relevance. So these components mean that we kind of have to look at click-through rate as well. One way to go about baking this all together is to look at revenue per impression as a metric, or just to look a lot at conversion rate and then try to get the highest converting ad that also has a good click-through rate, if that makes any sense.
Is there something in between click-rate and e-commerce conversion that might be an indicator of a performing ad?
I think we’re always going to go with a purchase to be honest, when it comes to e-commerce. And that gets a little bit to how we structure things. Granularity I think is seen as this holy grail of specificity. But it cuts down on the data that we have to work with in any given situation.
So reasonable aggregation might trump really hypergranular things that you can look at and you never have data to conclude a test from. So I would say that yes, outside of e-commerce if we have long sales cycles, we have problems with understanding how often something meaningful happens for a large B2B e-commerce, sorry enterprise software, HR consulting, something big, nuclear reactors what not. Then we need to go by those proxy metrics until we do land that big fish. So that might be the white paper download, or the engagement. So we might need to import from Google Analytics some engagement metrics, which pages are viewed, or number of pages viewed, or something like that.
What are some of tips and best practices about writing effective ads?
Well I think one of the big myths is that we have a lot of white space. Well, no one thinks we have a lot of white space. But sometimes we need to remind ourselves that this isn’t a high-flown narrative.
The ad industry as a whole will tell us all about the 1984 ad by Apple, or go back in time as Adam Parrott gave a great talk on the tried and true positioning of traditional copywriters. But you got a full page Time magazine ad, we don’t have a full page for that. So the Volkswagen Beetle ad might look very different in this tiny space.
So instead of persuading we are really filtering to some degree, so the kind of verbiage that actually might weed out inappropriate buyers. And then I think it’s really about some of the important details, not minor details, ’cause that just gives us minor results and random outcomes. But something like the nature of the call-to-action, important details that resonate with that particular consumer. And it is again though, just about testing. So to bring some principles in your toolbox as testing principles. Not to say that, oh the ads that show that we have a wide selection are always better than the ads that just talk about our key selling propositions. It’s like, try the different styles of ads, use the the ad that in this category, discusses four brands that you offer. And the worst that can happen is that ad loses. So it’s more about deciding on several triggers that might be significant triggers, and just running those tests.
When bringing in candidates against an incumbent ad, are there any practices for protecting my incumbent if it’s performing?
So yes and no. There’s some technical things that we can do to start, all the way to the end of getting around the problem of getting rid of incumbents for no good reasons–such as a small change in policy, shipping price, dates, and things.
So ad customizers may in the future help us more and more swap out components of the ads without placing the ads and killing your quality score on that ad. That’s a sort of separate issue, but I think we have to bear in mind that Google whether it’s on the organic side or on the paid side, really loves incumbents. They love data that is very significant. So the incumbent first of all, may continue to win for a long time. So it’s probably safe anyway.
But obviously its impression share is going to diminish. There’s no great way to fix that. If you just duplicate that ad, unfortunately, Google doesn’t really know what you’re up to there. It’s starting in with blank slate on quality score so it’s actually not a workaround. That ad may not perform the same as the incumbent, the data can be very confusing. So I would say though, DO let your upstarts have a lot more rope because of that quality score, they’re virgins with no quality score, they’re going to require more runway than you might think.
What advice would you give to somebody who’s in house who needs to hire an agency. What are a few things they should look for?
I think you do now start off with the certifications. We’re a Google Premier Partner, a Bing Ads Elite Partner, those are a place to start, they didn’t use to be, but they’ve gotten a little more stringent at the higher levels. Beyond that I think, “fit”. So in the sense of what types of accounts do they work on? Whether it’s small and local, whether they have experience in B2B, whether they have experience in shopping and e-commerce, so that’d be second.
And I think from there, certainly depth and bandwidth. Size doesn’t matter if it’s 300 people, but I really think it might matter if you have a two person agency where you’d really need to know the two owners of that agency are going to work on your account. If they get busy and they get swamped, that’s an issue. But how to tell whether someone really can move the needle for you, I think you have to engage in it. A bit of a dance there with maybe an audit, maybe something to figure out how they would problem solve your account without trying to steal ideas from agencies. To get at least the sense of what kind of problem solving they might bring to the table.
Thank you Andrew Goodman of Page Zero, for helping us out with digital marketing best practices!