Meet The Mailers: Learn First, Test Smarter

I chatted with Scott Harvey, Senior Vice President Direct Marketing at Quad. Quad is a print and marketing services company and is headquartered in Sussex, WI.

We talked about how using data-driven pre-market testing can accurately predict consumer response and shape better creative strategies faster and earlier in a campaign.

Among the many topics we covered:

• Scott’s background in marketing
• Quick rundown of Quad’s services
• New learn and test model for mail campaigns
• How the new model speeds up revenue
• Role of AI in testing
• What kind of mail get his attention

Here are some questions and answers from our conversation (edited for clarity and space):

• So, you know, this is a tall order, given the size and the breadth of Quad, right? But can you give kind of a quick rundown of all the company’s services for people who maybe aren’t familiar with Quad?
  

Yeah, well, we turned 55 this year, so it’s pretty remarkable. Great organization and a great culture.

We like to refer to ourselves as a marketing experience company. And sometimes people think, you’re an event marketer. No, that’s really not the way that we interpret that. It’s really about focusing on the marketer. They’re focused on the consumer experience. So we want to be focused on the marketer’s experience. And we’re rooted with our production roots in print.

Like I said, we’re 55 years old back in the day when Harry Quadracci started our organization, producing publications. And then that graduated into retail inserts and catalogs. And since then, maybe the last 10 or 15 years, we’ve done a lot of different acquisitions. We’ve consolidated the print industry down, but we’ve extended into a lot of marketing services. And so today we’ve got a full media organization under the umbrella of Rise. We’ve got a tremendous creative team under the umbrella of Betty.

We’ve got an amazing data stack that we just launched this past year. And of course, we root everything with technology and obviously the emergence of artificial intelligence and AI. And with that, right, we can really reach the consumer through the screens in your house, the screen in your pocket, the billboards that you drive by on the highway, the in-store displays that you see inside of brick and mortar, the screens that you might now see through our in-store media network, and then ultimately through the mailbox, and back to the business that I get the opportunity to lead here at Quad, which is the direct marketing segment of the business.

Just recently we started up a direct marketing agency that we refer to as Direct. And it’s really encompassing all the elements that you need in order to be a world-class performing direct marketer.

And so it starts with tremendous strategy. It feeds into audience, it ties into creative, of course, execution, and then ultimately back to measurement. You have to kind of ground that with world-class account management. And then you have to have this quest for always wanting to learn. Today, we refer to that as test and learn, but as I’m sure you and I will get into, we’re going to try to displace that methodology with more of a learn and test mindset.

• OK, well, let’s really get into it. The reason I wanted to have you on the program is because at the National Postal Forum in May, there was a workshop that you ran with Sean O’Neill of OneMain Financial talking about an innovative way to predict winning campaigns. So why don’t you talk about some of the details of this new model, what you and OneMain did together and how you overcame the limitations of older testing?
  

Yeah, well, first of all, it was an absolute blast to present with Sean and he’s great people and OneMain is just a tremendous organization. And them allowing us to tell the story in front of that audience at the NPF is just tremendous because it’s really supportive of our industry.

Getting into what Sean and I talked about, which is the pain points that today’s direct marketer feels is really predicated on the cost structure that has evolved here in the last 10 years. If you go back 10 years ago, postage was probably minimally half of maybe 60, more than that it’s gone up in the last 10 years.

And, at that conference, we talked about mail equals math. It’s just a math equation. I’m going to mail X number of pieces. My predictive models are going to tell me I’m going to get Y responses, and that’s going to prove itself out at an ROI model.

And eventually, the cost structure goes up. I can’t afford to mail as deep into that file. And out there at the NPF, I talked about 10 years ago, you might have had a mailer mailing 20 million pieces a month. And good direct marketing tactics would tell you should be testing 20% of your mail file.

Well, that would be 4 million pieces to test in a given month. And if your test cells 400,000, that’s 10 test cells a month. Maybe four audience, maybe three or four for creative, and two maybe for offers. All of a sudden you’ve got 10 great. And if you’re winning one out of 10, you’re probably getting an increase in your control each month on one of those 10 test cells.

Well, today, if prices have gone up more than 50%, that 20 is now 10. So we’re only mailing 10 million. And if we’re still true diehards to the industry, we’re still mailing 20% of our 10 million. That’s 2 million. And so now instead of having 10 test cells, I have 5 test cells.

And so they’re precious. In today’s world, those cells oftentimes still go to a lot of audience solutions as they should. That’s still the biggest impact in the piece.

Creative testing, offer placement, it tends to get kicked down a little bit, but it can yet still have a 30-50% impact on what you’re doing. And so putting those test cells into market, they’re just so precious. Why not go to market with something you think and know is going to win versus just what your gut tells you is going to win creatively?

And that’s what we developed here at Quad. We released Accelerated Marketing Insights back in 2018. Since then, we’ve done a ton of pre-market testing studies with our customers. And it’s a disruptive mindset instead of actually going in market to test, pre-market through our technology and process through online surveys, run our conjoint analysis, and then identify the winners before you actually go to market.

And so it’s that test and learn. And what Sean and team at OneMain identified was, yes, let’s learn and then test.

• In that example with OneMain, predictive modeling basically replaced much of the traditional testing process. So how confident can marketers be in a model’s predictions compared to results from a live mail test?

Yeah, you know, I think the root of that question is probably are marketers skeptical that if I do an online survey for creative, will those results actually hold up when I mail somebody a piece of mail at their mailbox?

And the math would tell us that today when marketers do creative testing, they win at about one out of five, one out of 10 rate. So a 10-20% hit rate on getting creative to increase response above their control. When we do an AMI study, we’re winning about 8, 8 1/2 times out of 10. So we’re displacing that control 85% of the time.

[Y]ou know what? Every marketer that we introduce this concept to, they’re skeptical at first. Of course, I was skeptical when we built it. Who wouldn’t be? Who’s going to think that what we survey online people with would actually translate into performance through the mailbox?

And the eight, nine years we’ve been doing this, we’ve calibrated the study so well with the kind of the art and the science of this and our conjoint analysis that we want. That’s really the backbone of our statistics to the survey panels that we put out there.

It’s just been phenomenal to see. We’ve got it dialed in pretty good. So if we, in Sean’s case for OneMain, we predicted …I think it was a 6 1/2 percent increase and it came in at like 7.1. So it’s pretty refined today.

• So when a predictive model identifies a likely winner before a campaign launches, what happens to the role for creative testing? Does it disappear or does it simply kind of move into a different stage of the process?

Yeah, well, remember earlier as I was talking about those test cells, right? And we were doing that fictitious company that went from 20 million to 10 million. So now they’re only testing 2 million. Creative testing has always been very, very critical for getting the brands to connect with the consumer. That the right creative to the right consumer will respond better.

And really, that’s what Accelerated Marketing Insights is all about. It’s how do we placate that creative. Where should the offer go in that package? Should it go to the upper right? Should it be called out in the middle? Should it be on the left-hand banner? What headline should you use? Should you use a teaser on your OE? What we do is we deconstruct creative. We reconstruct it through our surveys and serve that up.

We’re testing over 1,800 different combinations in one study. And it would basically take you almost two years of in market testing to get the same amount of creative knowledge and know-how that we do literally in one study.

So that’s the role of creative is critical. And it was time for the industry to think beyond our yesteryear of learning.

• So what impact does predictive audience selection have on the speed of a campaign? Can marketers realistically get into the mail stream faster with less investment than under traditional testing models?

If you just take maybe one of the financial segments, like let’s take life insurance for an example. You talk to any direct marketer in life insurance and they will tell you from the time that they drop tests, to the time that they get a read and they’re back in market, it’s probably every bit of 9 to 12 months.

And so just think of what, again, we talk about what it takes to create content. Well, you’ve got to have a strategy of what kind of, we’re using that life insurance. How are we going to talk about the product? Who are we targeting? Then we’ve got to generate that copy and that content. Oh, don’t forget, that’s got to get approved through your legal and compliance teams. You’ve got to apply all the state versioning to that.

So there’s this massive heavy lift. You do all that work to hope that you win one out of 10 times or maybe 2 out of 10 if you’re best of class. Wouldn’t you rather go to market with confidence in that new creative that, hey, I got an 80% chance that this creative is going to win?

And so the time frame in which we do a study, we’ll conduct studies in about 30 to 45 days. And that literally is from start to finish.

That’s the kickoff meeting with the customer, establishing our KPIs, identifying our control package, for us to generate all the creative versioning that needs to be done that goes into the study, identifying our audience and who we are going to send the survey universe out to, compiling our results through our conjoint analysis, and then doing the readout with the customer.
So from there, at that 30-45-day stage, we then say, okay, Creative one, two, and three, not four through 10, are our most likely winners. And they say they’re going to get a 12%, a 10%, and an 8% lift. Let’s build those out through final creative.

Let’s take those through our legal and compliance teams and through our full copywriting team to make sure we get all of our state versioning and, you know, the Is and the proverbial Ts crossed correctly. And then let’s put those into market. [I]f you’re a creative winner, just take a calendar year. If you don’t get your first creative winner until June, right? You drop in June. You’re not dropping that creative until probably at the earliest, the tail end of Q1 the following year if we do our study in January and you’re testing in March. You’re dropping in Q4 for that same calendar year. So you’re accelerating the revenue by almost six months. [I]’s about revenue acceleration for the marketer, right? That’s what they’re ultimately trying to do.

• [L]ooking ahead three to five years, do you think we’ll reach a point where predictive models handle most audiences and offer decisions before a single piece of mail is produced?

Well, we’re on this cusp in the next three to five years with AI. And rarely do I have a conversation with a customer where we don’t start talking and engaging on how AI is impacting all the different aspects of what it takes to reach the consumer through the mailbox. Look, in three to five years, if I knew exactly where the market would be, I would be putting some bets in the stock market.

But I think we have a pretty good vibe as to where that’s going to be. AI is not going, it’s not a replacement. I had an accountant once tell me, AI is like Microsoft Excel. It replaced the calculator with spreadsheets. And what AI is able to do is incredible.

It’s going to streamline a lot of the processes, the data analysis, the predictive modeling that you alluded to, it’s going to have a massive impact on all of that. But what it’s not going to be able to do is ultimately tell us how is the consumer going to fall in love with my brand. Am I going to take action? And it might be able to help us refine our predictive modeling, but we still need humans that are going to make those decisions at the end of the day.

We are incorporating AI into our Accelerated Marketing Insights. We’d be foolish not to. Four or five years ago, it used to take us 90 days to do a study. Well, now it’s 30 to 45. And really that swag of 15 is just how quickly we get things done with the customer in terms of sign off and approvals, etc. Much of that is because AI now is doing a lot of our conjoined analysis and reporting. So our creative is the first generation of that. We’re generating that through AI, and then we’re refining it with our creative team. So it’s coming. The next three to five years is going to be really, really interesting to see the impacts that it continues to have on not only our industries, but the adjacent marketing industries in which Quad serves.

The one thing I’ll say about AI, it’s made what I get the opportunity to do actually kind of grow in relevance with the consumer. I heard this phrase and I use it all the time now. Direct mail creates a moment of focus in a world that’s built to distract.

That’s what’s so fascinating about it. We get the opportunity to take a brand, find a consumer. Help them create the right content and offer and package, and then try to bring those two together to ultimately help that marketer grow incrementality in terms of revenue and their earnings. In simplistic terms, it does that. And with AI, touching so many things that we do on screens, this return to touch. The consumer really enjoys interacting with brands in which they touch things. And I think in store, you’re seeing more elaborate displays these days than maybe we’ve ever seen in the past.

But this return to touch is real and especially for the younger generation. They’ve seen themselves on a screen 1000 times, a million times probably if you’re a Snapchat person… Well, today, that next generation, when they see their name on a direct mail piece and it’s the right brand and it’s the right offer, I think we’ve got something really special for the industry as we go forward.

• Yeah, to me, it’s the combination of those personalized touches when you’re using VDP in conjunction with soft touch and other embellishments and finishes, which go back to that touch, that tactile quality that you can’t get on a screen, you can’t get on your phone. Everything looks the same on your phone.
  

So I love getting mail, for many years. When I first got into this industry and you tell people, what do you do? [T]he most simplified way I would say that is, “all that stuff you get in your mailbox, I make all that stuff.” And they’re like, “oh, I got you to blame.” I’m like, “wait, tell me about one of the ones that you responded to.” And rarely have I ever had anybody say, “well, I don’t respond to anything”. Because if we do our jobs, right? And we do them right. We can find the right consumer with the right brand, with the right offer at the right time, and they will respond.

Here is our conversation (with all questions and answers). We’ve added timecodes for your convenience.

Thank you very much, Scott, for sharing your perspective and your expertise! To learn more about Quad, visit their website at Quad.com.

Your comments and ideas are very important to us in making your Who’s Mailing What! experience even better for you. Through these engaging talks, we hope you’ll take away practical tips, insights, and personal stories to inspire and build your own success.

If you have any feedback — or are interested in sharing your story and viewpoint with our wide and diverse audience on “Meet the Mailers” — please reach out to me. I’d love to hear from you!