Sep 24, 2019
Grubhub CRM lead and No-Shave November Co-Founder, Christine Hill gives us insight into, operationally, what it takes to have a successful IP Warming, survive a migration, and achieve high deliverability. Our very own Nicole Codd, previously at ReturnPath, adds her commentary from countless email setups.
PJ Bruno: Hi again. Welcome back to Braze for Impact, your MarTech industry discuss digest. I'm PJ Bruno, and I'm here today in Chicago in the Grubhub HQ. Joining us today, Christine Hill, who is basically a Renaissance woman in the marketing org here. How's it going? Thanks for being here.
Christine Hill: It's going well. Thanks for having me.
PJ Bruno: And rounding out the group today, I have Nicole Codd, who's also in town from our success org at Braze.
Nicole Codd: Hey, PJ.
PJ Bruno: Thanks so much for being here.
Nicole Codd: Course.
PJ Bruno: This episode is focusing on avoiding migration migraines, and also talking a little bit about IP warming. Christine has gone through a number of IP warmings, and so she's going to share some of her experience there. But before we jump in, Christine, correct me if I'm wrong, you were a part of kicking off No Shave November, is that right?
Christine Hill: That's correct.
PJ Bruno: How did that come about?
Christine Hill: That's a family effort. My siblings and I, we lost our father to cancer a number of years ago. No Shave November for us was a way that we could take that pain and grief and do something about it, knowing that a lot of people are affected by cancer in their lives, whether it's a close relative, a friend, a neighbor, we grow our hair for those that lose it. That's No Shave November. It was a way that anyone of any bit of means could donate something small or something large and raise awareness by growing out their hair and starting that conversation.
PJ Bruno: I love it. I'm sorry for your loss.
Christine Hill: Thank you.
PJ Bruno: I'll be growing my beard extremely long, big, wispy mustache. Nicole, you're going to take part.
Nicole Codd: I will not. I'm sorry.
PJ Bruno: Well, not the mustache on that, but you know, just letting it go.
Nicole Codd: I'll let my hair grow out. I'll probably-
Christine Hill: Skip your trim.
Nicole Codd: I'll skip my trim that month.
PJ Bruno: Beautiful. Were you involved in the marketing strategy around that as well? How did you divide up your team there?
Christine Hill: Sure. It's something that my siblings and I kind of divide and conquer. I'm one of eight. People joke about having a baseball team. We tell our mother that she had a business. So I'm a marketer, I have a brother who's an engineer, another one who's in finance, and you fill in all the spots and it's basically a business. We take all of the skills that we have in the backgrounds that we've established and we came together to do it.
PJ Bruno: That's unbelievable. You guys are dangerous. I got to assume there's an ongoing text thread. It's like, "When are we all going to just start the real company?" That's got to be ongoing conversation, right?
Christine Hill: That's been a background conversation probably since I was in my early teens, like, "All right, what are we going to do? Hill Enterprise. It could be something."
PJ Bruno: Oh my God, that's so exciting. Eight. God bless your mom. What were some of the lessons and techniques that you took with you from that experience which launched you onto this bigger mission you have here at Grubhub?
Christine Hill: Sure. With that one, definitely learning how to build an audience, knowing that there were people out there that could relate to our cause and our mission but how to have them hear of who we are, how to get that awareness out there. We started with maybe a hundred or so followers, and how do you grow that? We took to social and all the digital means, because that's where people are. And using that, we were able to kind of be explosive and target a lot of people with minimal effort and minimal resources and really get it out there. From there, it took off because people relate to what it is.
PJ Bruno: Well, if you guys out there looking to take part, go to no-shave.org to learn more info. We'd love for y'all to get involved. I know I will be. Moving on to IP warming academy, luckily both of you have a good amount of experience with this, some horror stories, I believe. First and foremost, IP warming is crucial to get right for your deliverability. For those of you who don't know what it is, it's sending first small amounts of email and then incrementally larger and larger amounts of email day after day so you can get the respect of your ISP and they know that you're not sending out blasts of email to people where it may not be super relevant. What do you need to be able to do that? Those are your first sends. What needs to be in order before you click the send button?
Christine Hill: First and foremost, you need a plan with IP warming. I think a lot of people want to jump straight into it, and it is something you can jump into and ramp up fairly quickly. But you should have a plan with calculated risk because if it goes wrong at this point, it's a lot harder to get out of the hole if you had instead set yourself up with a solid foundation.
Nicole Codd: And I think that's the main thing that people tend to forget about, is they don't necessarily think about the ramifications of if it doesn't go well. There's always these business needs that are like, "All right, we got to get this IP up and running. We have X deadline. Let's do everything possible to get it there." And because of that, they end up rushing through it and don't necessarily have a successful warming and they got to start from ground zero again.
PJ Bruno: And this isn't a new problem.
Nicole Codd: No, no.
PJ Bruno: So why is it still happening? Why does it still happen?
Christine Hill: We live in a world where we want things done yesterday, and so we think with that we rush it. And second to that is that we don't think about all of the stuff that's foundational to ensuring successful IP warming. So setting up all the authentications on the backend, registering your IPs, verifying your records, splitting your audience, cleaning your list, I don't think a lot of people do that before they start IP warming or migration. If you do that, you're going to set yourself up for success because you'll be sending to people who do want your messages, versus becoming someone who is spamming everyone at once. And the email clients pick up on that, and they'll block you.
PJ Bruno: You mentioned list cleansing. Now, what exactly is that for people who don't know what that is?
Nicole Codd: Yeah, so at a base level it's essentially getting rid of email addresses that aren't valid email addresses. If I'm going to a website and I know I don't necessarily want to receive mail from that specific company but I want the deal or whatever they're offering, I'll enter a fake email address. We'll see email addresses that are just a string of letters or something completely unintelligible, or there's potentially typo errors. So, see a lot of those. Gmail.con, or Yahoo with one O. So list cleansing is the process of running your email list through some sort of service or something you have on your own backend to get rid of those fake email addresses, get rid of typo errors to ensure that you're sending to real humans or real email addresses as opposed to fake ones.
PJ Bruno: Gotcha. I forget who it was that we had the conversation about list cleansing, but we had a client who was adamant about not doing it because he said, "We don't give up on our users." I was like, "You're not quite getting it, though."
Nicole Codd: You've got to hang on to those strings of unintelligible letters. It's very important.
Christine Hill: It might turn into someone.
Nicole Codd: It's very important.
Christine Hill: To what Nicole said, removing the mal form users, unknown users, and those who haven't engaged in years, and I get wanting to hold on to them for business needs, but there's a certain cutoff point that it does not make sense to contact them through an email channel. There are other channels.
Nicole Codd: I think that's a big, big, big, big one. People are always sitting there being like, "But what if they come back? What if I send them this one thing and that's what revitalizes them and they want to order from us and engage with us?" The thing is, if they haven't opened a message from you in years, they're probably not going to.
Christine Hill: The likelihood is slim.
Nicole Codd: Right, and especially with most companies these days, marketing is everywhere and I'm not going to need a email to think about your brand. There's going to be other methods, which I'm like, "Oh! Oh yeah, I forgot about that. Let me go back."
PJ Bruno: All right, story time. Let's hear about your first IP warming. We don't need to name drop or anything, but I just want to hear a little bit about the first time you came up against it, some of those challenges.
Christine Hill: Oh man, there's been so many. The first, I have a feeling it didn't go according to plan. I'm sure we didn't really have a plan. It was one of those rush case scenarios, "Let's get it done in two weeks," when I think even in the best case an IP warming probably takes a solid 30 days, even with a small MCL. We probably went out the gate sending to all records and wanting to trigger a very large program, versus starting with a small send size, gradually increasing it, and in that hitting our most engaged before those who are less engaged. I do know that in that first IP warming, we hit blocks with Gmail, with Microsoft, with AOL. Once you hit blocks, it's a lot harder to get into the user's inbox and you are digging yourself out of a hole.
PJ Bruno: So blocking meaning the users blocked you, not an ISP.
Christine Hill: The ISP blocked you.
PJ Bruno: Got you. Okay.
Christine Hill: Because you were, in fact, spamming people.
PJ Bruno: Was there a comeback from that? What was the reaction from you and your team?
Christine Hill: Usually for that, you need to register it with the ISP and get de-listed and say, "I am a true sender. I'm going to take step back and send to you in a rate. Will you recognize me and like allow me to be authenticated to get into the user's inbox?" And Nicole, if you want to add to this from your return path experience?
Nicole Codd: No, that's exactly right. I think with mailbox providers you're kind of guilty until proven innocent, honestly. Their job is to protect their users, so the people signing up to use their inboxes, and because of that you're starting out with basically a reputation where they're going to treat you very carefully. So their initial inclination is going to be like, "Okay, you jumped up volumes too quickly. We're going to just shut it off because we don't trust that you're an actual good sender."
PJ Bruno: So what kind of metrics are you looking at as far as IP warming? Is it really just like open, so it's just getting through it alive?
Christine Hill: Yeah. Looking at sends, looking at your rate of deliverability, your bounces, hard bounces or soft bounces, and some of that would be your list cleaning, looking at engagement, so your opens and clicks, as well as the spam complaints that you're getting and unsub. So it's really looking at the full spectrum. What is happening positively, what's happening negatively, and what is our deliverability? What is the rate of sends landing into the user's inbox?
PJ Bruno: What about any big either misconceptions when you went into your first or second one that you're like, "Oh I didn't know that's how that worked," and/or biggest takeaways from one of your first two IP warmers where you're like, "Yes, this is going to stick with me forever because I recognize how valuable it is."
Christine Hill: One for me, looking at the metrics, looking at unsub versus spam, spam is worse to have a than unsub because unsub, you're telling me you don't want to receive my messages. I'll remove you. That's better than you marking me as spam, because when that happens then the email clients have that and it can push you off of landing in the inbox. Going through the first few, I realized that that's more important, and tied to that would be the list cleaning because if you had lists clean, you'd be sending to people who will likely be opening and not unsubscribing or marking you as spam.
PJ Bruno: Interesting question here. Have you seen strategies where people make that unsub link a little more visible? Just because obviously pretty much every time I've seen it, it's the same thing that you've all seen. It's super small, it's at the very bottom, there's no context around it. Then you click and you link out and you do the motions.
Christine Hill: So of the best practices is to put that unsub in the header. There are a lot of ESPs that make that very easy to do, or in Gmail if you have it a you're more likely to get through. And then those users who don't want to hear from you can click it versus marketing you as spam or filtering you to your spam folder.
PJ Bruno: Nice.
Nicole Codd: And also, even outside of that list unsubscribe header, something I've advised clients to do in the past is to move the little unsubscribe link at the bottom up towards the top, potentially. Because if I don't want your mail, my initial inclination is going to be to mark it as spam, not to scroll to the bottom of the email and click the unsubscribe link. That's going to take too much work and requires me going through the entire email.
Christine Hill: In some of the migrations that we've done, when you've been part of another brand that then became part of Grubhub, we've put it in the body of the email, which might sound crazy, but it is that way for the user to hit unsub versus spam.
PJ Bruno: And also they're making that transition, and so you guys are being super forthcoming with that opportunity. Well, I'm glad you mentioned migration. Why don't we migrate over to migration conversation? You've had a handful of those, and it's great. It means you found a better tool to support you and everything, but how can you be more agile and efficient beforehand? Because sometimes that eagerness can blind you to what it actually takes to get it done. So what does it take?
Christine Hill: Again, IP warming, having a plan and having a realistic timeline. This is something that people rush a lot, but before all of that, it really is data. Understanding the data you have, how to map the data over to the new platform, and data in, data out. So what you're starting with is what you'll be working with. A lot of people will rush it with unclean data and will fix it later. Then, again, you're putting yourself back behind the starting line. So it's really the time to take to do it right, if allowed.
Nicole Codd: And I think something people forget about is the time piece. So if you're on one platform and contract's coming up for renewal and you know you're migrating over to another platform, it's so important that you're putting that into motion well before you have to be off of that other platform, because mistakes happen. It always takes longer than you think it's going to, and it's way better to be fully moved over before you have to be off than to be trying to add months on and be like, "Oh my gosh, we need a few more months here before we can be fully on to this other platform." It'll just go much more smoothly if there's plenty of time in between.
PJ Bruno: Without getting too nitty gritty into the details, are there steps to a migration? Obviously, the data needs to go from here to here. We need to make sure there's enough time. What are the pieces in between all that?
Nicole Codd: The data mapping, ensuring that ... Something I see with a lot of clients is, like Christine said, it's kind of like a hodgepodge of data. And so ensuring you have something on your backend that is like, "This means X and this means Y," so that when you move it over it matches exactly. Because it's going to be a problem if one side says one thing, you move it over, and someone's like, "Oh, I thought this meant that." So ensuring you have proper data mapping on the backend.
PJ Bruno: How much time do you like to give for a migration? What's the timeline?
Christine Hill: It really depends on the amount of programs you're moving over and your list size. Part of it is that IP warming piece where you have to establish the IP and the domains and grow that up to be able to send, and second to that is being able to send your program, so your welcome stream, your reactivation, one offs, abandon cart, all of those things. Your list size could be 1 million, it could be 16 million. Those would take different lengths of time to migrate. The other thing that I've noticed a lot with migrations is people want to go out and launch a program that's easy to get up and running, so maybe like a one-off campaign, so a blast. That's not the way to do it. It's to do the gradual sends. So establish a program that isn't gated by a specific time. A welcome series could be a good one because you know that there's users just signed up and are engaged with your product or service, but in a way that are you okay with them receiving it in the first five days of becoming a user or the first one? Because if it's the first one and you have a large increase of users and you're not quite yet warmed, you won't be able to send to them.
PJ Bruno: Okay. Story time once more, and you can choose what the prompt is. It's either/or. You can tell us about your biggest success with migration or one thing you wish you could do over again.
Christine Hill: I have a biggest success one, I'd say. Grubhub, as we do, acquired another brand. We were given a very, very tight timeline to turn it around and migrate the users over into our full Grubhub experience, which would offer more restaurants, drivers, and a more powerful network for our users and our diners. There were a lot of records, and ones that typically we would scrub out, people who haven't been engaged in a very long time, that we wanted to hit, because as a business we believe that there could be some return. So there was a very big lesson. I was working with Braze's deliverability team. My initial thought at seeing this plan is, "There's no way we can hit these records. We will be blacklisted on every platform." I was pressed to do it and I said, "Okay, if I'm going to do this, I'm going to do this in the most calculated way that I can, a calculated risk, that ensures that we can maintain deliverability in some of the ESPs, like hopefully." It was a very, very old audience and I sliced the heck out of it to find the most engaged of this entire unengaged audience. I called it unfavorables, and I said that given the volume that we're sending to people who are engaged, who want to hear from Grubhub, we can only 10% of unfavorables of that total volume getting emails a day. And I started with the most engaged of this least engaged, stacked against Grubhub's very engaged. And day over day, and this was a 14-day migration that should have probably been 30 plus, I monitored it in the morning, in the mid day, in the evening, in the late evening. I knew that if any one thing went off, that I needed to be able to adjust it. I said, "I will do this, but I need the keys to be able to adjust it." We get through day one and day two, surprisingly not hitting any blocks, and I'm getting pressure to increase that 10% to a 20. Held my ground and did not and continued to say, "You know, we're prioritizing the most engaged of the unfavorables, and this percentage." We get through it and we're trickling through all the way through the list. And I think it came from Andrea on your delivery team was, "That's amazing. I'm not quite sure how you did that, but you were able to get through it." So that's my thought, too. Some deliverability god was looking down upon me. Again, as calculated of a risk as I could take, but knowing that what I was doing is not what I would choose to be doing, but for the business it made sense to push through. It was probably my most proud moment. I said, "If I can get through through this set of audience, I could migrate any audience."
PJ Bruno: Yeah, totally. So when it comes to migration, good rule of thumb is do it bit by bit, monitor all the way.
Christine Hill: And don't be afraid to adjust. If you do run into a block, it's better to adjust to take the time to do it then, versus to really put yourself back.
Nicole Codd: That's huge. That's something everyone always forgets, too. They set this plan in place and they're like, "All right, I'm just going to let it run and let it sit." Having the ability to be able to be like, "All right, we're going to bump back today, or maybe we stay at the same volume for another few days," that's huge and that's not something people always think about. It tends to be a set it and forget it exercise, which doesn't work.
PJ Bruno: It's a common thing we see with people taking automation in their hands. They think it's like, "Oh great, the job's done now." But I think the job's just beginning. You need to stay close to that.
Christine Hill: I think for us, especially moving more and more over to the Braze platform, we shifted our email and our push and we use your in-app. Looking at our marketing holistically, omni-channel messaging for our users and hitting them on the channel that they are most engaged. So understanding that it might not be email, it might be push, it could be in-app, and having that congruent flow for a user. If we push you, what are you then seeing in our app? What are you then getting in your email later? And it's really opening up the door for us to be able to do that and to do it in real time, which is something before this, and when I first joined two and a half years ago, we were on multiple platforms and nothing was talking to each other. We couldn't have a program like that.
PJ Bruno: Two and a half years is such a small amount of time. It's crazy. When you were telling me earlier that it was a team of two when you started, and now it's massive.
Christine Hill: That's just an explanation of the growth here. It truly is explosive, and with that, we're not afraid to be taking risks and innovating. That to me is exciting, and it's nice to have a larger team and that manpower to do it.
PJ Bruno: Love it. Also, you guys recently bought Tapingo, which is kind of Grubhub for college kids. Since you do have your hands in most things marketing here, is there something you guys are doing differently with your approach to that demographic?
Christine Hill: We are. Tapingo is becoming Grubhub campus dining, and it's a whole new instance of Grubhub tailored to students, things that are in and around their campus, identity experience fit for that demographic that they'll use in their time in college. And as they graduate out of college, they'll graduate onto full Grubhub. But something that is more personalized for them.
PJ Bruno: Now we've got to just do it for toddlers. We got to start them super young.
Christine Hill: If you get them young, you can keep them longer.
Nicole Codd: It's all about retention, you know?
PJ Bruno: That's what I'm saying. We don't give up on our users. Awesome. Christine Hill, thanks so much for giving us your time.
Christine Hill: Yeah, thank you.
PJ Bruno: Nicole, thank you so much for being here in Chicago.
Nicole Codd: Yeah, thanks for having me.
PJ Bruno: And thank you for being here. Bye.