Welcome to How to Create, Differentiate and Measure Your Ads. I’m Taylor Holladay the managing partner of Common Thread Collective. And today in Module 1, we’re going to be answering the question, “What makes good creative?” This is a question that we have worked really hard to answer across our internal agency as well alongside all of our clients.
And the most common answer that we get back is this quote from David Ogilvy. “If it doesn’t sell it isn’t creative. Good creative sells product”. And this is a fantastic answer – sort of. Our opinion of this answer, though, is that it is incomplete. It’s a lot like saying to you that if an ad isn’t working or if it isn’t selling.
It’s a lot like saying your car is broken. The reality is that cars don’t break any more than ads break parts of cars brake and parts of ads brake. So we began to try and understand what are the component parts or the signals of an ad that lead us to understand whether or not it will sell.
And this pursuit really began by trying to answer this question about whether or not a stimulus response model exists that would allow us to understand the leading indicators of ad performance that would lead to roles or return on ad spend in the future. The object of any stimulus response model
is to establish a map them out of goal function that describes the relationship between the stimulus and the expected value of the response. In other words, is there an input signal within Facebook adds manager or whatever platform we’re measuring ad success in. That would correlate or lead us to understand or to be able to predict the future of our ad
selling the return on the ultimate investment of the ad and we became obsessed with this idea after learning from a company called First MediaFirst media is a media property that focuses on creating socially native content. Here in Los Angeles, and they make some of the most viral videos on Facebook every single year. I’m sure you’ve seen a ton of them.
They make great content across platforms like, blossom. And so yummy and they’ve become masters at building socially viral content. Here’s an example first media is responsible for three
of the five most virally organically shared videos on Facebook in the last two years. And when I heard them speak about this content, they boiled down their entire creative strategy built around one single metric share rate they’re able to determine within the first 10,000 impressions whether a video will go viral or not based
on the rate at which the video is shared. Now across all of Facebook the average share rate on content is 0.4 percent. That means for every 1,000 impressions that a video receives four people will share it well blossom is able to get a share in it almost six times that by focusing exclusively on that leading indicator.
So they get a share rate on their content of over 2 and 1/2 percent. So as I discovered this. And I was learning I became obsessed with trying to understand is there a similar model that we could follow for ad creative in the e-commerce world. Is there a stimulus similar to share rate that lead to the output, which for us we know is sales that we could focus on to narrow down
the focus of what creative strategy entails. So we began by doing a statistical analysis of all of the ad creative that existed to try and find out. Are any of the common advertising metrics cost per click through rate cost per thousand impressions to any of them give a signal that
would allow us to focus on that would correlate or relate to relax. Ultimately on our ads unfortunately, what we found in looking at all of these individual metrics was that none of them individually correlated to Ross. If you see our data here you can look at C the r squared for Ross an upper funnel metric.
So when we say upper funnel. What we mean is early stage signals. If you think about the customer’s interaction with an ad it begins with an impression and then from there, the click and then, ultimately, the price of the traffic related to what the CPM on the ad is. And so what we wanted to understand was that in most media buying systems
the cost per click the cost of the traffic directly relates to the return on investment. Because if you’re bulk buying media if I’m buying a media placement of 1,000 random customers the price of that traffic is fixed across every individual user, regardless of their likelihood of conversion.
But the unique thing about Facebook that makes it so that these individual metrics actually have very little relationship with return is that Facebook prices, their media at the individual user level. So in some cases, you should be willing to pay way more on a CPM, or cost per thousand impression basis because that individual user is much higher
has a much higher likelihood of conversion, what this leads to is a bunch of random variation across these metrics like cost per click, click the rate and CPM. So in fact, there’s actually no relationship between those metrics and sales. Now this is a highly controversial statement. And we would encourage you to go and explore
this data for your self. But if we look at the R squared, which is measures the correlation between two different metrics and it’s measured on a scale from negative 1, meaning it’s perfectly inversely correlated meaning that as one metric goes up, the other would go down. That would indicate a perfect inverse relationship or a negative r squared or a positive correlation, which
would be a one r squared as one goes up, the other goes up or as one goes down, the other goes down. Those would indicate a relationship between two metrics. But when you look at CPC c tr and CPM they basically have no relationship CPC has a 0.08 are square CCR is a 0.1
and CPM is a negative 0.12 which doesn’t mean that they’re inversely related or positively related. It just means they have no relationship at all. So they are not good signals for whether or not your ad is going to generate sales. This is really important because I’m sure a ton of you have in your ads dashboard these metrics are making decisions
about your ads relative to these metrics and are also turning off creative relative to the performance of these metrics. The only metric that we saw that was upper funnel meaning above purchase in the customer journey that has any relationship with Ross is cost per unique add to cart. That’s the first signal that you get.
That can actually give you a relationship with the lower your cost per unique add to cart is, the higher your roads will become. While this makes it really challenging that we don’t have a single metric like share rate that we can use in the same way that first media uses for their creative variations. The good news is that we have found a set of metrics
that when compiled together. Do create a relationship. And that’s what we’re going to go over now in module.