Lobcast Podcast: Marketing Optimizations & Old Fashioneds

STEPHANIE: Hello, And welcome to the Lobcast Podcast: Mixers and Marketing. I'm Stephanie Donelson, your hostess with the marketing mostess, and I'm the senior content marketing manager here at Lob. I'm thrilled to be joined by Mike Tuffley, VP Solutions Engineering. Tuffley, do you mind introducing yourself to our listeners?

TUFFLEY: Yeah, happy to. So again, my name is Mike Tuffley. I'm the head of solutions engineering. So effectively all things technical in our customer-facing side. So we I'm really the mad scientist of Lob, so all of the cool stuff that we get to see presented in demos like all of that futuristic technology that we're bringing to the world of direct mail. Oftentimes, that originates out of my head and the folks on my team.

STEPHANIE: Well, thank you so much for joining us today. And listeners, if you wanna make the complementary cocktail that goes with this episode, which is an old fashioned, you're going to need two ounces of bourbon or rye whiskey, a quarter ounce of simple syrup, two dashes of angostura bitters, and either an orange peel or a Luxardo cherry for garnish, I chose a cherry for mine, but cheers, Tuffley, and welcome to the show. All right, so today we're gonna be talking about a topic that many marketers like to focus on: optimization. So, Tuffley, you work pretty closely with our customers. What are some of the top challenges that you found that they're trying to solve by optimizing their direct mail?

TUFFLEY: Absolutely. Well, I mean, even when I hear like, oh, it's not about ROI, like let's be real. It's always about ROI. So they're trying to optimize for one of two things. How do we how do we maximize that return and how do we minimize that investment. So for we'll use marketers for example that, you know, they'll they're like, oh, how do we increase that average card size or what sort of use cases can we get to re engage folks in, you know, life cycle marketing use cases like hey, sixty day win backs or maybe an abandoned cart type of use case. How do we how do we maximize customer, LTV, like all of those all of those things that they're going to experience, they're gonna experiment with in terms of like the offers or the creative, things like that to really try and capture a higher response and a higher conversion rate, and then on the other hand it's going to be okay, how do I minimize that investment? A lot of folks especially the larger more mature organizations that have been using like leveraging direct mail as a channel for decades have completely bloated processes, the timelines are obscene, the amount of like human resources and energy and back and forth to get something shipped takes forever and then when it does get shipped there's still all of the fallibility of a data, not data informed approach to executing on these campaigns, how much of that mail gets returned and shows up in giant USPS tubs to your office, so you have to go clear out a PO box. So like contending with all of that that realized waste as well. So how do we make sure that offers aren't being forwarded you know, when a customer moves, they they file that change of address with the USPS. How do we make sure that if an offer is not valid there, we don't want to get it forward, have it forwarded and then it's a poor customer experience, not only that but maybe the household where the offer would have been valid doesn't receive that mail piece and we spent money so there's all this really bad experience and waste that's generated as a result. So those are the two things that like I see most folks trying to optimize for. I'm Like I said, I'm a scientist, I'm a mathematician, I always think of optimization in those those, that perspectives of the lens of like, okay, we have a multi variable function. We wanna find all of those maximums, all of the minimums, So like one axis is going to be, you know, like that return axis, like how do we have like a higher response rate? Yeah. And then Another access is gonna be, oh, the the cost associated where it's the hard costs of the the printing operations or some softer costs of the time and manpower that goes into getting these things out the door. And then where do I... And then a third one could be something entirely different and then where can I find those maximums and you can only do that through like continually experimenting and iterating, and that's that's optimization. It's a journey not a destination.

STEPHANIE: I love that. That's perfect. And I mean, I think both of our teams benefit from each other, right? Like you can think of the mathematical side, which No offense, but my brain does not do very well. I did okay in algebra, but I'm okay not dabbling in math anymore, but I love coming up with the creative side of things, especially doing those A/B tests, like what CTA options can we do? What kind of copy can we change? How much can swapping out an image of you know, a father and son for a mother and daughter. How much could that really improve conversions? And then sending the results over to you guys to be like, okay, so did this actually do anything or not?

TUFFLEY: I love it. And like that's that's exactly what I mean by like the absence of data informed is not doing any of that. Versus how do I leverage my first-party customer data and maybe it's been like also supplemented with some third party household data as well. How do I take all of that information how do I map that to a creative that creates a very compelling mailpiece where I know that I'm gonna drive response and I know that I'm going to be able to convert because I've put so much intention into all of this and then how do I measure that.

STEPHANIE: Yeah.

TUFFLEY: How do I measure that accurately? And then how do I take the analysis of all of that data and feed it back into my model so that I can A/B test in the next cycle and do that quickly and not have to wait for the results and not have to ship several campaigns with a non-performing, say, creative before realizing it. And then I can course correct, so being able to iterate quickly is also like time is of the essence.

STEPHANIE: Oh, definitely. I think you hit the nail on the head, right? Like it's all about the investment portion. You don't wanna be wasting that, especially sending out mail that in the end you're not getting results on.

TUFFLEY: I could not agree more. And I, you know, I I love A/B testing because it's still testing. But when I really think about that is a great maybe phase two, so if we think about like a marketing marketing maturity scale, right? Like that's somewhere in the middle. But there's still kind of a time delay because I'm really only testing two elements. I still have to have a lot of controls in place to make sure like validity of my scientific approach is sound. I have to make sure that everything is measurable and that I do the analysis and I could run, you know, several A/B tests in parallel and I can A/B test, you know, quickly like ideally to get that into like maybe a sixty-day time frame, so I can still course correct within like a same quarter. But where I think that the real value is like multivariate testing, like how do I do this at at scale where I still have like the controls in place necessary to make sure that my that my experiment is statistically sound.

STEPHANIE: Yep.

TUFFLEY: The population size is statistically significant based on my historical data and then actually conduct a test where I could get an entire year's worth of A/B testing out the door shipped in one pass.

STEPHANIE: Oh, guys are a lot busier than I am.

TUFFLEY: It's a lot more planning upfront, but I think that the the downstream effects are where you're gonna see all of the return.

STEPHANIE: Oh, for sure and especially when you get to that analysis part which we will talk about later, But right now, I do kinda wanna return our conversation back to just optimizations in general, and, Tuffley, from your perspective, what are some of the different types of optimizations that our customers can work on, like workflow optimization, technology optimization, campaign optimization. What am I missing here?

TUFFLEY: Those are all that's a really good start. Workflow optimization that kind of goes back to the original point of how do I take this very time-consuming person human resource intensive process that has all of those little manual steps and the back and forths that when you actually like aggregate all of the time and energy to ship one campaign, like how do I reduce that? And what's amazing to see is we have we work with like a very high up on the fortune list telecom company where they started out with this very like bloated I mean, the workflow diagram will make you motion sick like it is that overwhelming. And when they walk you through each of those nodes and then what you don't even have an appreciation for because it's a static image are the little feedback loops when something doesn't go according to plan and that's where all of those incremental time costs can really start to balloon. And it was insane. So like on average it was taking like usually around two months, you know, maybe a month if they were lucky to ship like a large campaign of millions of mail pieces. And what was so cool is that when they adopted Lob and they really implemented all of the automation and were able to, like, really achieve those efficiencies at scale. They took this process that was two months and condensed it down to two weeks. Oh, wow. And we're able to eliminate so much time from the process that they relied less heavily on their outsourced marketing agency. So like there's all of those peripheral cost reductions that you also don't even realize are again going back and ballooning the overall like actual cost.

STEPHANIE: No, definitely. Well, I think we even had webinar not too long ago with our nonprofit partner, United Through Reading, and they talked about just the manual process and how much time it was taking to send these letters, donor acknowledgement letters, and just by automating that process alone, they were able to have people focus on what only a human can do and really applying their talents elsewhere. It was like, Yeah, you, yeah, you can calculate that cost out, but it's more just the, oh my gosh, we're getting so much more important stuff done. Because obviously the donor acknowledgement stuff is important too, but it was such a manual labor intensive process that being able to move that over there and then focus more on outreach, you're getting gains on both sides.

TUFFLEY: Absolutely. I mean, at the end of the day, we wanna give folks back more time to work on what's important, you know, like, and like male operations and logistics is not that thing that you wanna be spending your time on as like a marketer.

STEPHANIE: No, I'm not gonna sign up for that one, but speaking of marketing, marketing automation is obviously something we all rely on in this space, but how does that play nicely with optimization? Or I guess another way to think about it, what are some things that marketers should keep in min as they optimize their marketing technology?

TUFFLEY: That's a great question and I think that that also kind of again, goes back to the preceding question around what are the things that we're going to optimize for. So you mentioned workflow optimization, campaign optimization, I think that marketing automation does those two things really nicely. Right? Because when I think about again going back to a manual process that's time consuming It's error prone. It's costly. If you can have a marketing automation tool and there's a number of them, I know a large number of our customers use like several of the Adobe products, the Salesforce products, you know, some of your even like the so many of the CRMs and ESPs and CDPs are also starting to to borrow elements from the marketing automation to achieve something similar, but again at At the end of the day, we want to make sure that everything is event driven in real time we wanna make sure that it's like harmoniously executed in concert in a fully automated way with our other digital channels, so that I should be able to send a postcard as easy as an email and have those customer experiences and messaging be indistinguishable from one another. And that is, to be honest, that's what I like about Lob and I always use the analogy like it's not really direct mail, it's really more like getting an email in your mailbox. So, you know, you have that marketing automation. It's just running in the background. It's, again, it's like scheduled or event-driven.

STEPHANIE: Yep.

TUFFLEY: Every day you send out, you know, hey, your birthday postcards to your customers whose birthdays are coming up and that job just runs every single day. You don't ever have to think about it or you have something that's event driven. I somebody came to my site, they signed up, they showed all of those intent signals, maybe you had some other data and you're like, oh, yes, this is high propensity buyer, they correspond to like a great persona within the model I've built around my ideal customer and I want to then have them enter this particular funnel. We'll send them an email. If they don't open it, we'll send them another email. If they still don't open it or maybe they unsubscribe, we'll send them a postcard. We know when that postcard is delivered with Lob. We know when that QR code is scanned with Lob. And then we can use that like our webhook technology to return all of that data back to your marketing automation platform and then we know, "Oh, cool, we can move to the next sequence in that automated customer journey." I didn't have to put like an arbitrary fourteen day delay there where I just hope in a prayer that it had arrived in that time, that it had actually been seen you know, like I now have again conducting all of this automation informed and in real time again, goes back to, like, campaign optimization. I know exactly what creative was was being used in service of that particular campaign or even better like for that specific customer like maybe I we have that level of personalization. So I know, okay, great. They received it. They acted on it. I can make sure that I can do the proper attribution, and then I could do the analysis in the back end. So those are all of the things that we like to see and again, it's just a cycle where it's just shampoo, rinse, repeat.

STEPHANIE: Yep, I've talked on the podcast before at one of my previous organizations I really wish we had found a service like Lob because we were sending nurture emails, and you would start to notice when people were dropping off or worse hitting that unsubscribe button and it's like, we had had an opportunity to send them a postcard, break up the cadence, stretch that nurture journey out and across multiple channels, I think we probably would have seen a lot more success.

TUFFLEY: Could not agree more and I'm you know what's funny is I realized emails like quick and easy, it's cheap, but at the same time like and I'm guilty of this even with like one of our customers where I will did not act on any of the emails oftentimes they get relegated to that extra tab in the back on my Gmail.

STEPHANIE: Promotions!

TUFFLEY: Yeah. And So I didn't never saw it, never opened it, but I got a postcard and it was it's nice because, and this is not not even reflective of the fact that I work in the space, but I like to be reminded of something. So, like, I'm like, oh, yeah, cool. I do want to act on that at some point, so I'll just, like, leave it on my coffee table, you know, we're, you know, magnet it to your refrigerator.

STEPHANIE: Yep.

TUFFLEY: And then, you know, Saturday night rolls around after a couple of beers, you're like, yeah, I'm gonna pull the trigger on that. Okay. You know, and that was like

STEPHANIE: That coupon code!

TUFFLEY: Yeah. Exactly that was the experience that I had. I was like, oh, that's really cool. And I think about from a marketer's perspective like it's not even just about the fact that your open rates are really low that your click-through rates are even lower, but like now that we've had all of this implementation of like privacy intervention measures being done by the largest companies in, you know, in the world, Apple, Google, like we don't even have great insight and one of the cool things about the direct mail channel is there's literally no intervention that they can take that's going to distort any of my ability to measure the effectiveness of that campaign.

STEPHANIE: Yeah, alright, so earlier we kinda talked about waste and making sure that we're getting the most out of our investment, so before we move on to our next topic, I just kinda quickly wanna discuss optimizing direct mail campaigns through address verification. Tuffley, can you talk a little bit about that and how that can reduce ways and just streamline campaigns?

TUFFLEY: Yeah, that's a great one. So there's a couple of different things that I think about when I approach like, okay, I want to ensure the integrity of my data. Trash in, trash out, so the better we can clean this up prior to launching a campaign, the better. So when I pull all of the members of that audience that I want to target. I'm going to run it through AV for to just do that additional peace of mind. Again, we're talking about something in the real world that's somewhat disconnected digitally. So the time between when you last sent them something and they received, there's a great chance that they could have moved that that address no longer even exists, or maybe they've implemented some sort of solicitation, mail, you know, interception type of things. So when I run that list through AV, I get a number of different factors back like one, like is it just deliverable? Is this a real address? If it's not like, is it a real address, but maybe it's missing like a unit number, so there's still a high likelihood that it will get to where it's intended, Or is this just like joker data like 123 Butt Street probably isn't is probably a waste of that that mail cost, right? So the being able to just understand whether or not it's deliverable and then I look at maybe some of the other facets of the data that's returned from that address verification, things like knowing how it's zoned. If I'm sending out an offer for, I don't know, you know, pet services. I don't know that maybe sending it to a commercially zoned address is like it's going to get to where it was intended or if it could actually be acted upon so that I know whether or not residential or commercial I know whether or not the unit is currently vacant. I know whether or not Lob who, this is where we kind of set ourselves apart in the address verification space is we send a ton of mail effects to over, you know, half of the households in the US, so we can tell you hey, have we have we been able to successfully deliver mail there before? And if so, here's the level of confidence that we have in your ability to get that mail to its final, you know, ultimate destination. So that's just everything just upfront, like so I think about that funnel and then now I have my final address verified list that satisfies all of my requirements, and then I'm going to send that over to Lob and what's really cool is I have NCOA enabled on my log account, so when I send that request over, I get a response back and one of the parameters that is returned is recipient underscore moved. And if that value is true, you're going to see a redacted forwarding address that's going to give you the city, state and zip five and that's a constraint USPS. We're not just like suppressing data because we hate our customer. No. This is one of those funny USPS things. But you can actually export the data at the end of one week, again, funny USPS convention, but you can get that forwarding data back Okay. But the nice thing, yeah, but the nice thing about getting that response in real time is I can decision off of that. Yep. If I'm beholden to like compliance I don't want my mail forwarded and then now I'm in breach of some sort of regulation, right? The case that I had illustrated before where if somebody sends something and they want to go to a household where the valid is offer or the offer is valid but not having forwarded, I can I can act I can I can literally decision off of that? I can like, okay, cancel it to that person, then we're gonna resend it with current resident there so that we still hit the...

STEPHANIE: Hit somebody!

TUFFLEY: Yeah, that we're canvassing a neighborhood or something, right? So, yeah.

STEPHANIE: Okay, cool. So I think that tees up nicely to our next discussion topic, which is optimization in action, we already talked about it a little bit, but in our new optimization ebook, we talk a lot about RVA, Recency, Velocity, and Accuracy. Tuffley, can you explain that principle a little bit more?

TUFFLEY: I can, and you know what's so cool is I've never invented something that I've heard like referenced before.That was something that I had devised last year as like a part of like how we would grade or like evaluate somebody's like optimization because we kept hearing like "Oh, I'm optimized. Yep. I'm optimized. I'm fine." Okay. Well -- Are you -- can't be what's the last time you actually ran an experiment, recency. Right? Like what's the old philosopher saying you know a person never stands in the same river twice because they're not the same person and that's not the same river, like that's the that's the evolution. We live in a TikTok generation like you can't tell me that you ran an experiment on you know your eighteen to twenty-one demographic and still tell me that that is applicable today, like that's that is not recency. So That is, that's one of my, that's like what I use. It's a little bit hyperbolized, but I definitely think that it holds true. And I have a fifteen year old daughter, so I think I can attest to how quickly these things evolve, especially in an instantaneous digital world. And then the next thing is velocity is I'm sure that you are near optimized in your digital channels because you're able to quickly like iterate test and try again, like you just run that cycle very quickly, but if you're doing like a traditional print in mail, like that's just not possible. So if you're not moving quickly, And in fact, one of the things that I love to ask marketers when I first sit down at the table with them is like hey, what are you doing successfully today in your digital channels that you would like to map over to direct mail so that you could see even higher response and conversion rates, right? And then let's start running those experiments in parallel. So we have customers that are literally doing campaigns in email and then taking the exact same content and putting it on a postcard, same link but it's been shortened as a little pURL or a QR code, like everything is the exact same, same pair of pants that you left in your shopping cart last week, same pair of pants on that postcard. Everything is one to one except the difference is somebody will see this.

STEPHANIE: Then it's like the Facebook ad that's following you around for the rest of your life.

TUFFLEY: Yeah. Yeah. And that's yeah. And that's that's a whole another conversation around where we get our third party data but that's like, that's how we can take the velocity component of that, right? Where we're able to do that so rapidly in real time and we know like, hey, this person didn't respond and we're seeing this campaign not materialize within the first thirty to sixty days. Let's course correct and let's change up those different elements. And I love like your example earlier about, you know, even if it's something as trivial as like lifestyle imagery like what do we know about our customers? What can we do to map that to an experience? If I know that they're a teacher or a vet, I want some imagery of someone in fatigues like giving their kid a big hug, right? Like that's the type of compelling imagery that's going to resonate with my target audience.

STEPHANIE: You wanna see yourself reflected back in the marketing materials that are coming to you. And then the accuracy portion, can you talk about that?

TUFFLEY: Oh, yes. Thank you. So I've keep going back to this parallel to email, right? And one of the really nice things about your email, you know, your social, your SEO, like all of those digital channels it's very easy to measure. I log into my email account. I have one that I sent emails out with so I see what my open rate is, my bounce rate, I see my click-through rate, all of that being monitored in real time and that's how I can gauge the accuracy. You don't really get that with traditional print and mail. Like, I think I think for the most part, it's primarily a black box like you just hope that it gets there and you kind of you look for that halo effect or some sort of a conversion that you think is within that twenty one day attribution window, like all of these very muddy types of measurement methodologies of you know, last year and how do we how do we do that, you know, take that digital how do we take that digital measurement experience? So that's where we talk about like accuracy is like, hey, How complex? How sophisticated is your attribution model? Are you just doing single touch, first touch, last touch, are you doing, like, a really cool multi-touch attribution model or something like or maybe like a time decay type of model, you know, we're doing a combination of a weighted first and last touch. So, you know, those are really easy to do with digital mail or sorry with digital channels. With direct mail, if you don't know when something's delivered, how do you know when to start the clock on even that simple twenty-one day attribution window, let alone your time decay curve of how much weight you're gonna give that in the overall model, right? Yeah. Because so many of our customers again going back to the marketing automation, this is so many touches across digital, direct mail that nobody's just sending I'm sending email and, like, or direct mail. Like, it's multi-touch, most of them are like omnichannel like our larger customers, so being able to know exactly how much weight to give that that mail piece and then that's how you defend the spend in a channel like like digital mail.

STEPHANIE: Yeah, I've even talked on the podcast before about how I'll get a piece of direct mail, and maybe it has a coupon code, that coupon code is gonna expire at some point most likely, I'll put it with the rest of my mail, it sits there, gets a little dusty, and then I'm cleaning it out, and then I come across it, the coupon code may have expired, but I still will be like, oh, maybe they're running a different sale, I can go to their website, I know I'm gonna get that pop up that's join our email list and get twenty percent off your first purchase. There you go, you still got my business, even though it may not have been from that first touch via direct mail.

TUFFLEY: Right? And what I love about that model is that historically has been something that's like constrained to a channel like email or SMS, right, or or in app notification. Well, we had a customer who's like, I don't I don't wanna do that? Like, I don't first of all, I don't want a generic coupon code to make it out into, like, the coupon ecosystem and then. And then on top of that, like, yeah, and then it just, again, I get this halo effect, right, or like whatever very muddy attribution waters. So being able to pass in like a unique one to one offer code, again, just improves the measurement we have customers that are passing a UTM parameter with a customer ID so they know when that exact customer visited the site because they can map it back to that UID.

STEPHANIE: No, that's awesome. Alright, since we're kinda talking about parameters and tracking, What kind of data should marketers use to determine what needs to be optimized first, potentially going back to our conversation around workflows, marketing tech, what are some warning signs that marketers should be whoa, okay, something needs to change here.

TUFFLEY: Yeah. That's that's a great question. And it's not an easy answer because it really is going to vary by the maturity and sophistication of marketing operations and the size of the team and like all of all of those things. And how much and what type of data that you have, you know, like we have customers that have never sent direct mail, you know, maybe they're like, hey, I'm an ecomm or a fintech company. I've only been around for like three years and we were just heard that direct mail is awesome, so we were to adopt it. They don't have a lot to go on.

STEPHANIE: Yeah.

TUFFLEY: We're going to what we're gonna do is say we're gonna go back to that same question of like, alright, well let's just start with what's working in email, right? Or any of your other digital channels, let's see what's working, so let's start the experiment there. And in terms of like the targets that we wanna set is like based on your historical response rate and then we'll try and adjust it for the direct mail channel so that way we actually have target metrics to be able to see like, hey, did we hit those KPIs in the success of our campaigns? Let's say they have sent mail before, but maybe it's been very limited hey, you know, what couple of times we send out like quarterly postcards or letters or something like that. Not a lot of sophistication there or like historical context of like where to kind of begin like with experiment because unless you've been running a bunch of A/B tests you don't have a lot of data to work off of. So what we would say is like, okay, let's see what did work. And again, we're gonna borrow from the digital channels let's use some contextual data from the other parts that are those experiments wrap in near real time, right? So like I don't know what the window is, but it's probably like a week rather than sixty days. So let's take those inputs and then I think about maybe doing something as simple as like a prospect experiment, you know, where we're buying like third party data And oftentimes third party data comes with all of that other enriched, you know, metadata for the household. So being able to see that way we can take like a very broad swath of folks and we can just run like a generic experiment, it's gonna be simple A/B, but at least like it's that one step forward where then we can begin the iterative process. And that first step is always going to be the biggest, the hardest hurdle to get over, and then once you're there, then we move up the step to again people who have been doing this historically and it's like, hey, how do we accelerate this? Yeah. How do we take the A/B testing and we move you, we push you up to multi variant testing where we're going to take some AI generated body copy to generate like ten headlines and ten subheads and we're going to see which tone and sentiment whether it was value messaging or urgency messaging that had the highest response rate. We're going to test to see like which permutation of the two or which unique wording or phraseology and uniquely won out. So like that's where we see the coolest stuff like I said, we have we have customers that are doing experiments where they're creating like hundreds of thousands of permutations to be able to test and my favorite anecdote when I was sitting down at the table with those brand folks who hadn't used AI like they had to manually do all of this by hand like a copywriter create like ten different, ten or twenty different like heads, subheads, then all of the actual body copy that went into this letter So it was very like time-consuming process, but my favorite part was they're like, hey, what was the most interesting and counterintuitive results of our experiment is that we put this really cheesy, really campy headline in like it was something like "Vroom vroom come get your car loan," like and that wound up being one of the highest performing body copy headlines --

STEPHANIE: Yep. I can see that!

TUFFLEY: That they had. And they were like, they were blown away and like that's the cool part about being able to experiment. As you can challenge those preconceived notions and then iterate quickly. So they had a very involved like data analysis team that had created this really cool multi arm bandit like data model so that the higher performing versions or permutations would kinda propagate to the top and then they would continually iterating on the higher performers.

STEPHANIE: I mean, you have just been speaking my language. That was one of jobs at a previous organization, I literally would come up with the headlines, here's what we're testing, and it's so true, it is really crazy how much your customers can surprise you with, like, what they end up reacting to because I've done social media posts where I'm like, kind of a filler post, we're just getting something out there, and it goes practically viral, and it's, what? Like it's just your customers will never stop surprising you, or kind of what you said earlier with like the TikTok generation. Your customer base is always evolving and changing, so you have to make sure your campaigns evolve and change with them and what they're looking for.

TUFFLEY: Absolutely. Einstein said, you know, all the experiments in the world will never prove me right but a single experiment will prove me wrong.

STEPHANIE: Definitely. And that's what you're trying to prove, right, like making sure that the same result keeps happening.

TUFFLEY: Yep.

STEPHANIE: Since you were just kinda we're talking about A/B testing, I'm really curious, in your opinion, what is something that you think marketers should start with first in A/B testing when they're beginning to optimize their direct mail, like what would you recommend they start with form factor, the design, the copy, you kind of alluded to that a little bit, CTAs, what would you recommend?

TUFFLEY: Yeah, so it's if there's no precedent for testing, we're just like, hey, this is net new. Yeah, I mean I'm in the KISS model, right? So just like like simplicity is best. So I think the example that it just gave where you could do a couple of different versions of some headlines and some headlines just to see like where your market is at, you know, like the fact that you could run these tests, get the results, measure them, analyze, and go back to the drawing board, in sixty days as opposed to like six months, allows you to run that follow-up baby test where like, okay, we think we have a good winner for like the body copy, let's start testing lifestyle imagery, and let's see if we can't map it to what we know about the consumers in that segment. You know? Like I said, we have we have a bunch of people that are doing that with, like, the third-party data especially. There's a politician who knows whether or not you're a cat household or a dog household, so it's him holding a cat or a dog accordingly. That shows up on your mailpiece.

STEPHANIE: What if you're a bird person?

TUFFLEY: Not not who I want my constituency.

STEPHANIE: Alright, Tuffley, I've got a great question for you. In your opinion, what are the most commonly used triggers for direct mail? And what questions could marketers ask themselves to optimize those touch points with customers or prospects?

TUFFLEY: Yeah, that's a great question. So I think that on a scheduled basis one of the more common ones that we would see is going to be like a daily or weekly job that runs where I get it. Just queries your data and says, okay, show me the folks who haven't interacted with us in the last sixty days.

STEPHANIE: Mhmm.

TUFFLEY: Right? Let's send them some sort of win back campaign with like a personalized offer, and then let's try and map in all the information we know about like maybe their last purchase or or maybe the reason why they canceled their subscription or whatever that could be to really target them at the individual level. So those are great ones. They typically it's the set and forget they just run periodically. One of our, some of our largest customers have a lot of them running in parallel. So they'll have their win back campaign, they'll have their remarketing campaign, they'll have their cross sell campaign all running in parallel on varying cadences, daily, semi,-weekly, weekly, and the mail just goes out. And then like I don't know when I think about more of like triggered use case, something that's like event driven and done in real time would be the would be another one. That could be something like the abandoned cart.

STEPHANIE: Mhmm.

TUFFLEY: It could be like a new mover. Yeah. Well, I'm actually gonna I'll take that back. I'll put new movers into the weekly sense because you'll usually get that data from a third party, so just kinda run it all at once. But we'll go back. Anonymous website visitor and we've got a pixel, we do a postal append so we know who visited with what degree of certainty and maybe some other like persona data about them, and then let's send them an offer like, hey, and then that's when you get that creepy like, oh my gosh, is my Alexa listening to me.

STEPHANIE: She always is.

TUFFLEY: I shouldn't have said that word because all of a sudden she started talking to me in my headphones. But but exactly, you're like, oh my gosh, this is why are all these targeted ads appearing in so many different areas like those are the great use case for direct mail as well.

STEPHANIE: No, I think one of the examples I used it on our blog when we were talking about holiday marketing, and it was actually a mailer my husband got And on it, it specifically said, Santa's been watching you at, and then it had our address data. I thought it was hilarious. My husband thought it was very creepy, but I just I was like, At least you're being honest. So, like, come on, that's pretty clever. I like that.

TUFFLEY: Oh, I have! I'll give you a couple of examples that I thought were like borderline creepy know too much about you. One was it was a real estate customer that I mean, this is, I love the use case because of how technology brilliant it was. They had a picture of the front of your house from Google Maps Street view. So they knew, as if here's your house, so they hit the Google Maps API, then they hit the Zillow API and said, hey, this is how much your home may be worth, are you interested in selling? And it's like that, I mean that's one of those things where it's like wow, that is a lot of data and what's great is that's not even like they didn't have to buy any third party data. They knew the address and all they had to do was hit a couple public API endpoints and present that. So I thought that was a really good one. Also a really funny one was like a customer that they guilt neighbors into voting like they basically sent you a letter that says, hey, here are all the people in your neighborhood who didn't vote and it kind of tries to like shame you into voting, which I think is, it's controversial but I think, hey, we should all be participating in our democracy.

STEPHANIE: That is interesting. Oh, that's yeah, there are some unique things you can do with the data that's available out there.

TUFFLEY: Oh, yeah, we have a solar company that has a picture of your roof with Google Earth, superimposed with solar panels like this is how cool your home could look with solar panels.

STEPHANIE: Haven't gotten that one yet. I've definitely gotten the, "Here's how much your house is worth" and I'm like, "Nah, I don't believe these numbers," but. Alright. So, before we wrap up the podcast, I would love to quickly talk about measuring the optimization results. We kind of talked about it a little bit earlier, but let's do a deeper dive. So, what metrics should marketers be tracking as they work on optimizing the direct mail marketing campaigns.

TUFFLEY: Great question. And I I love this because this is like one of those things that really differentiates us from anyone else in the space. Like I know that I know that there are some certain like mail tracking tools out there like how robust their functionality is I I don't know, but it is also segregated from how they're executing their campaign. So like having an all in one solution and one of the coolest things that we've introduced recently that they're not going to have is the ability to measure click through on QR codes. Like I know in real time, I'll give you a fun story. We did a really cool event at Taylor Swift this past weekend.

STEPHANIE: I bet Kim was very jealous.

TUFFLEY: She was very jealous, but not as jealous as my fifteen-year-old daughter that I wasn't able to take. Story for another time. But I've made a mobile app that allowed people to take like selfies and send them this cool, Taylor Swift, souvenir postcard. And it's got the picture of their self and they could send the message to their family or friends or whatever. But the cool thing is it had a QR code that when somebody's scanned, I get a message in my Slack, I get a message in my email, they get an email saying, hey, thanks for scanning this QR code, here's to learn more about it. So like like having that ability to not just measure but integrate into your other marketing automations is key value there. But most importantly, I know their IP address, I know how many times they scan that QR code. I know when each of those times occurred, and again, that's all information that's feeding back into my system so that I know, hey, does this person like QR codes or did they interact with the pURL? Or maybe not as much resolution measurability maybe they were part of the folks that called and then I can go and see like does their phone match my record. So there's less but it's a great way to know exactly when somebody interacts with one of those CTAs on your postcard and we now measure QR code click-through.

STEPHANIE: That's awesome.

TUFFLEY: But the most important and I think way more impressive is our deliberate effect so we historically had this process for delivery event. It told you like, hey, this it's loaded on the truck, it'll be in the mailbox in the next like twelve hours. We now know exactly like when it's delivered within an hour because they're like USPS on trucks now that tell you when they're out at their stop. So again going back to some of those more complex attribution models where I want to know when to start that twenty one day attribution window I want to know when to start the time decay so that I can wait the touch appropriately within the scheme of other touches. Even if I just want to do simple match back knowing whether or not something was actually delivered so I can normalize my data to actually get a more accurate ROI calculation and then really like having that delivered event so that I could make a compelling case when I'm doing my ROAS calculation whether or not I could truly with confidence attribute that conversion, that order that was placed, whatever that incremental add on lines with conviction say, yes, this was due to this person reading the postcard. So that's what I like and like again, the more of those personalized components, even some like breathing like new life into old technologies, we have a customer that does one to one barcodes because they have people come into their stores and redeem the coupons in person. So, and they know when that bar code gets scanned exactly whose postcard that was.

STEPHANIE: Oh, I love that. I mean, like, yeah, you really do wanna be able to get that granular sometimes on your campaigns, especially if it is an omnichannel one, so then you can prove like no no, direct mail is a channel that we need to continue investing in.

TUFFLEY: Exactly. I I love the halo effect because, like, people just accept it, but I've seen more scrutiny especially in like recent months like this past year where there's like more more eyes are on that--

STEPHANIE: Yep.  

TUFFLEY: The marketing spend. Like how do we defend that spend and saying that something has a halo effect is great, but with lack of if you're not able to measure and demonstrate with like conviction and accuracy, then you have a very soft case for the channel even if you have a lot of confidence and faith historically in it.

or STEPHANIE: Yeah, no, since you just talked about the halo effect, that actually leads really nicely to the last question I have for you. How do you often tell our customers to give their campaigns extra time after making these optimizations that we've already discussed to determine if an optimization has worked. And if so, how much time?

TUFFLEY: Absolutely. So like I said, I've used kind of the twenty-one days that seems to be like the convention that most of our marketing customers have adopted. That's like the window that we're going to entertain the notion that this mail piece was received was viewed, was viewed, and if they had acted on that touch, it would have happened within this time.

STEPHANIE: Okay.

TUFFLEY: Knowing exactly when it's delivered, that that also kind of, like, helps to, like, again, to to to wait, you know, how much importance you would wanna put on that particular mail piece, what we typically see and it will be very like I said one of our customers has like an army of data scientists that could turn these things around like in like thirty to sixty days or something like that. But we typically see that I think both folks like after they do like match back is a really common attribution methodology. You know, it's usually like around sixty days, right? The more the more robust your like your integration with Lob is where you're getting not just like the creating mail but also receiving all the data back to your systems and you have that feeding into your BI tool for pre generated reports and you're just watching your cost per incremental ad go down every day or any of the other KPIs that you're keeping an eye on. Those are the things that we can really optimize for where we would like to see that to be closer to like thirty days. But again, realistically like I think sixty to ninety days is like the industry standard, but I think that I think we could get it down with greater evolution of like our technologies and our automation.

STEPHANIE: That's perfect. Alright. Tuffley, are there any final thoughts that you want share or is there anything you think we didn't get to today during our time together?

TUFFLEY: No, not at all. I'm really excited to see like I think that AI is going to be like that's gonna be my prediction. I'll give a shout out to our technology partners and friends at copy.ai who we've partnered closely with on this front to be able to demonstrate how customers can even in an automated way incorporate like that testing at scale for short form body copy. So that's and we'll give it a few months but then AI imagery I think is gonna be the next one. So that way you're not having to license stock photography, you're just like, hey, I know I know this person's like these household characteristics give me a great ad that will appeal to them And my hope is that that will just spit out, you know, like a family watching TV or, you know, like a dad unloading groceries from a mini you know, based on, you know, all of those inputs and prompts.

STEPHANIE: Yeah. Just wait for like the dad to be like having something weird on his head or something just based on the keywords that you put in there.

TUFFLEY: Right? Like the thirty fingers on a single hand with like a weird row of like a billion teeth. Yeah, that's That's why I said give it a few months like our AI technology is accelerating really rapidly. So I think that that's going to be where the next step in optimization where GPT is just gonna design your mail automation for you, and I'll be out of a job, but that's okay.

STEPHANIE: Just try and do, like, a commission on sending people over to copy.ai or something.

TUFFLEY: Yeah. Exactly.

STEPHANIE: Thank you so much for being our guest to talk about optimizations today. This is really fantastic. I learned a lot. I hope our listeners did too. And to our listeners, thank you so much for joining us for mixers and marketing. If you want to dive deeper into the topic of optimization, please download your complimentary copy of our ebook, Optimizing Direct Mail for Maximum Results at tinyurl.com/optimizingdm. That's tinyurl.com/optimizingdm. As always, you can browse our library of episodes over at lobdemoo.co/lobcast. But thanks for listening, and that's all folks.