What Should Your Ads Look Like? Diving into the Data on What Drives Performance

Think about the last ad you saw an ad that had an image of someone’s living room. What color was the throw pillow on the couch? Was there a dog present? How many people were pictured? You might not know the answer to those questions, but the marketers who put that creative together sure do, and you better believe that they thought long and hard about each of those aspects and dove deep into some data to decide exactly what to include in the ad

Those minuscule details may seem frivolous to the naked eye, but the data proves that every decision you make in your creative process has an impact on the bottom line, and the ecommerce businesses that pay attention to the data while still putting their own spin on the creative aspect are the ones that are rising to the top.

R.J. Talyor is the founder and CEO of Pattern89, which uses A.I. and machine learning to analyze advertising and guide marketers toward the performance metrics that matter. On this episode of Up Next in Commerce, R.J. takes us through the most recent trends report that Pattern89 put out, which includes some important information about why businesses should be paying particular attention to copy and hashtags. Plus, R.J. gives tips on how to avoid the trap of following “best practices” and why creativity will always win.

Main Takeaways:

• Big-Time Creative, Small Time Dimensions: Marketing creative is made up of tens of thousands of dimensions, all of which can be measured and distilled into performance metrics. And when those metrics are measured, you can utilize that data to inform the decisions you make about what elements are more beneficial to include in your ads (like an orange pillow versus a blue pillow and a cat versus a dog).
• #WorthIt?: Including hashtags in your marketing copy is a method to improve performance. The flip side of that argument is that hashtags have a cost, particularly if you overlay a hashtag on an image or video. 
• The Same… But Different: When developing creative, marketers are often looking at the same data, trends and industry reports, and they are following the same “best practices.” The trick is to take the information and make it your own in order to develop creative that stands out rather than just gets lost in the sea of “what works.”

For an in-depth look at this episode, check out the full transcript below. Quotes have been edited for clarity and length.

---

Up Next in Commerce is brought to you by Salesforce Commerce Cloud. Respond quickly to changing customer needs with flexible Ecommerce connected to marketing, sales, and service. Deliver intelligent commerce experiences your customers can trust, across every channel. Together, we’re ready for what’s next in commerce. Learn more at salesforce.com/commerce

---

Transcript:

Stephanie:

Welcome back to Up Next in Commerce. This is your host, Stephanie Postles. Joining us today is R.J. Talyor, the CEO and Founder at Pattern89. R.J., welcome.

R.J.:

Hey, thanks for having me. Appreciate it.

Stephanie:

Yeah, I'm excited to have you on the show. I just downloaded your report, as you saw, right before the interview and I feel like there's some good juicy stuff and the things that you guys are doing and we'll have a lot to talk about today.

R.J.:

Yeah. Great.

Stephanie:

So tell me a bit, what is Pattern89 and how did you go about creating that?

R.J.:

Pattern89, what we do is we predict creative performance with machine learning and AI, and I've worked in digital for almost 20 years now. I worked at ExactTarget and then at Salesforce and then at another social startup, and learned quickly that marketers are spending so much time creating journeys and figuring out who their audience is, but still use a lot of their gut on what creative to put in front of those audiences, or in the places in the journey. While data is being used to figure out what and who, it really is on the creative side, still kind of a gut decision. And so set out about four years ago to help marketers use AI and machine learning, to predict creative performance, and now we're a 25 person strong scale up and serving customers all over the globe.

Stephanie:

That's awesome. I was reading a bit about your data co-op and how many brands you have that you get to look through all the data, so give me a little bit of like the behind the scenes of what kind of brands do you have in there, if you can share, how many do you actually have, and how do you go about looking at all of their data?

R.J.:

Well, when we started the business, we knew we needed a giant data set in order for the machines to find patterns and outliers that would actually help creatives, and so we started offering a free scorecard or creative assessment in exchange for joining the co-op. We're almost at 2000 brands who-

Stephanie:

Wow.

R.J.:

... have connected their accounts to Pattern89. And our machines, they don't actually look at anything that's identifiable. Instead, they're looking at 49,000 different creative dimensions of a piece of creative and tying it to a performance metric. As an example, it might look at a picture of a living room and say, "All right, there's a couch, plants, window. Is there a human or not? What color is the couch? Is there a blanket on the couch? Is there a dog in the picture?"

R.J.:

Then associate that with a performance metric, like, did it drive X purchases or Y video views or X views, so then we can understand in aggregate what are the patterns and outliers that we're seeing across massive amounts of data, so that we know living rooms actually perform better when there's an orange throw pillow on the couch, for example, like the one small detail. So massive amounts of brands, massive amounts of data, and really granular creative insights.

Stephanie:

Very cool. How do you go about, if you're looking at that kind of anonymized data and you see something doing well, how do you know it's doing well? Maybe because it's a Nordstrom where it's like, well, of course they have brand recognition, it's very easy to buy from them, versus a brand new D2C company where maybe one of their ads is doing well, or it's not doing well and it's kind of getting skewed because they don't have that brand recognition.

R.J.:

Yeah. That is a big question. The first thing to know is that our data science team uses a set of statistical tests to kind of ferret out what is actually driving the performance or not driving the performance, so that we're not chasing after kind of red herrings. The other thing is from an audience perspective, the history of that audience use, so is it a brand new audience or is it a legacy audience? Is it a customer loyalty audience or something else like that? Those are all the components that are also taken into account so that we can distinguish between maybe the brand legacy of a Nordstrom versus an up-and-coming D2C startup and still provide statistical significance at over 95% competence.

Stephanie:

Got it. What things do you think are going to work in 2021? I mean, I was just going through the report that I mentioned earlier and I mean, it was cool because it's all the way down to "Here's certain emojis and colors and imagery is up, but maybe with multiple people or one person and video's down." So tell me a little bit about what you guys are seeing and predicting to work well when it comes to advertising and copy in 2021.

R.J.:

Yeah. Well, the first thing that it's important to notice, the creative lifecycle has shrunk and rom a creative life cycle perspective, it's just a lot shorter than it's ever been. In addition to the creative life cycle shortening, what is performance in those creatives is changing as well, so the best advertisers have a 2:1 image to video ratio. While video performs a lot better in general, image to video 2:1 is the best performer for top performers in the dataset, and we're seeing more and more emphasis on video and video performance, but videos tend to be more expensive, so 2:1 is a good framework.

R.J.:

Then taking a look at what's trending and what's forecasted to be a performant in 2021, we see things like TVs and electronics really improving cost per clicks, also spas and massages, those types of things. And then maybe not surprising, we're seeing a lot of isolated people, so one face in creatives and those people are reading or studying or doing something by themselves, which kind of reflects the world that we're living in, especially as we head into 2021.

R.J.:

One of the surprising things, and sorry for all the cat lovers out there, is that pets typically do very well, dogs and cats, but we're actually seeing cats driving performance down. So it's more expensive to advertise with cats in 2021, so just keep that in mind as you're here.

Stephanie:

That's funny. Don't put a cat in your ad.

R.J.:

Yes, exactly. Yeah. But consistent performers, things like images or videos of those people exercising, phones, images of cities, those types of things are really staying flat. This is all going to change by audience and by bra