Hyper-aware customers are forcing brands to think critically about the personalized experiences they deliver to customers. How can brands navigate personalization in the world of the discerning customer to deliver the experiences consumers want and expect?
This panel will explore how marketers can effectively leverage machine learning and AI to create meaningful dialogues with customers and why proven channels like email marketing are having a resurgence.
Panelists:
- Oded Benyo, President, Email, Epsilon-Conversant
- Ginger Conlon, Editor-in-Chief, MediaVillage
- Nicole Perrin, Principal Analyst, eMarketer
- Dane Mathews, Senior Director, Marketing Activation and CRM, ConAgra
Watch the 40-minute panel discussion or read the transcript below.
Ginger Conlon: Everyone, thanks for joining us today to talk about personalization. It's so important to connect with our customers today. They are hyper aware, they're expecting more and more from us. Recent Forrester Research points out that even companies that are top in customer experience are having a harder and harder time achieving their rankings because customer expectations keep going up. So even if your experiences are on par, it's harder, and they're great, it's harder and harder for customers… to get customers to appreciate those experiences. And a lot of that's around personalization. That's what we're going to talk about today.
We've got three great panelists with us. Oded Benyo, who is president of email for Epsilon, Nicole Perrin, principal analyst at eMarketer and Dane Mathews, senior director of marketing activation and CRM at Conagra. Welcome.
So the first question is the easy one, which is to tell us all a little bit about yourselves and your role so everyone gets a sense of your perspective on the market. We've got a great mix of different perspectives of the view of what's happening in the market with our panelists today. So, Nicole.
Nicole Perrin: So I'm a principal analyst at eMarketer, which means that I spend most of my time talking to practitioners about what they're doing, what their problems are, what they're trying to achieve. And I synthesize that information along with all of the quantitative data that we aggregate at eMarketer and our forecast to write reports about what's going on in the market. And I cover mostly some really core topics in digital advertising and marketing, including email, including search and also including the advertising businesses of the biggest companies selling ads today in the digital marketplace.
Ginger Conlon: Excellent. Oded?
Oded Benyo: Yeah, I'm the president for the harmony business in Epsilon. It's actually email and our push messaging SMS. And we really help our clients communicate with their clients and do it in a meaningful way, and help them personalize, help them manage their communications across channels and know who they're talking to across devices.
Ginger Conlon: Excellent, thanks. Dane.
Dane Mathews: And I lead marketing activation and CRM at Conagra and I basically use the best practices, and I leverage the technologies to help bring our a hundred X businesses to life in the marketplace.
Ginger Conlon: Excellent. Oh, and quick reminder to use Slido if you have any questions, we'll save some time for Q&A at the end. So Oded, Epsilon recently conducted some research around personalizing the customer experience. So I thought maybe you could share a couple of data points to level set for the conversation.
Oded Benyo: Yeah. Well, I think that an interesting finding to some extent may be obvious, but research to back it up. That when we went and asked customers, what do you think about personalization, right? That does it improve your chances to actually buy something if the experience is more relevant and surprise, surprise it is. Right? But more importantly, when you ask them, how do you feel about that in the channel and is this experience? Are your experiences personalized and meaningful? Most people say they're still not.
So we as an industry still have a way to go and figure out how we make that communication better and how we become more relevant. We're still, I think, all discovering, there's a lot of awareness but there's not so much execution at scale still.
Ginger Conlon: Right? Absolutely. And so with that awareness, I think all of us in the room here are aware of all the opportunities that personalization presents, but maybe some of the challenges aren't as obvious. So I thought we could start, let's just dive right into that part. So, Nicole, what are you seeing in the market?
Nicole Perrin: Well, I'm definitely seeing some challenges. So I think the first one is whether the organization is approaching this in a strategic way, whether it involves the whole organization or just one team. Because if you are trying to personalize within just one channel as opposed to across all your touch points, it's going to be limited in terms of how personalized it actually is. Some of the other challenges include getting all of an organization's data in one place, clean, in a usable, actionable form.
A lot of marketers still struggle with that and it's the foundation really of all of this. And then there's the challenge of measurement because personalization ultimately depends on optimization. If marketers can't measure their results, again in a holistic way, how can you optimize toward those results that you want? I could go on but maybe we should give someone else a chance ...
Oded Benyo: I completely agree on the data front, which is the foundation, right? You have to have great data to be even able to talk about personalization, but then you have to be able to activate it too. And so a lot of people now have gotten to that, having that data, but it's very hard to actually act on it. I think another big challenge that we see a lot of our customers have is around content. And how do you manage all that content and make it relevant for each individual.
As well thinking to your point about the silos in the organization, thinking about it widely and thinking about personalization beyond just content. When do you communicate with people? How do you communicate? Right? So content is just one element. The frequency you communicate with them. Content is just one element of a whole slew of areas that people need to think about and it's not that easy.
Dane Mathews: I think all those are true in all organizations and even Conagra's at the beginning of its journey in terms of organizing its data and having some of the technological solutions to be able to understand, listen and then syndicate communications to consumers. But I think for us, the biggest challenge has been how do we adjust the culture and how do we change the culture? And so those tools and technologies only enable a great culture. And so in many ways, we've taken the word, we'll say personalization almost off the table.
We still know it's a vision, but it is, I think the word personalization can tend to mean I need to know your shoe size before I can say something about eating a healthy meal or something. Like you almost need to have this kind of really intense structured data against a consumer before you could really make something happen. And we've chosen a more of a progress over perfection. And the word that we like to use internally is really relevance and how we're being relevant.
It's not necessarily that it has to be personalized, but it has to be relevant. And when you say it's relevant, it actually introduces the idea of context. And I think that's really important. And so in order for us to get on the journey, we had to choose a different word and we had to get started. We had enough capabilities to get going, but we really emphasized, let's just get started on this journey and create the right behaviors and skillsets, and then when we put in the right technologies and capabilities, those things will bloom.
Oded Benyo: I think… oh, go ahead, sorry.
Nicole Perrin: I think it's a really good idea to get rid of that word. I know one thing that happens when I have conversations about it is I usually have to take a moment at the beginning and say, “Well, are we talking only about one-to-one individualization?” Some people only consider personalization if it's one-to-one. Other people want to talk about customization. Some people want to talk about micro segments. I think relevance is a great way of thinking about it.
And I also really love your point about making incremental progress. I have talked to folks about this for several years and marketers tend to add one channel at a time when they're doing some type of personalization or doing some type of customization. They're probably not going to have that executed on every single channel at once. And if you can just put two or three channels together, it's going to be better experience than nothing. Right?
Dane Mathews: Right. And the consumer context is different too. The expectation around, well say, personalization or relevance is very different for consumers depending on are they on your own channels, are they in a paid channel? And I think that's an important part that doesn't limit your ability to just to get started and get going, and find the places you can activate most relevantly the fastest.
Ginger Conlon: Right. Which brings me to the next question, contextual relevance as you said, it's so important, it can happen ideally or happens… it's most likely to happen in the best way when you have a view into multiple channels, right? Because someone might expect a different message if they're walking down the street versus if they're seeing it from their T.V. or what have you. So, but even in that regard, email is still a work horse, it's still a center of customer experience. There's a lot of expectations around email. Dane, how are you using email to help bring those channels together and get that insight towards personalization or relevance for you?
Dane Mathews: We think about email as a very different channel than when we're in, we'll say the paid sense. And so email has a way to bring to life the owned native experiences that we have. And so it's almost, we think of it as an extension of it needs to be connected to and as an extension of your behavior on our own sides. And so it should reflect that, which makes the job much easier. And it just helps us understand that email isn't… it's not about being personalized or relevant in email per se, it's actually about how does relevance in that experience drive you back to a fuller experience that allows you to do and engage more?
So email just lives directly right next to our own channels and it's almost an extension of those own channels. And we have to be careful as an organization of trying to be matchy, matchy, which is, you looked at this thing, so I'm going to show you another version of this thing. It's funny, the panel right before us, there was this dialogue around science or creativity and or science and art, and marketing, and a lot of the conversations around the science and it makes sense. But there is a place for innovation and creativity, and you have to leave that.
If you only follow the data, you will continue to optimize in the same pool. So email is a place where we're both trying to show things that are relevant, but we're also testing new, different types of content or content structures, or themes in our ability to learn a little bit more about how to get consumers primed and back to our own channels.
Ginger Conlon: All right. Can we have a slide for when… Oded, I want you to talk about what you're seeing among your customers in that regard too. And we've got a slide, a visual to go along with that.
Oded Benyo: I don't remember what the visual is, but regardless-
Ginger Conlon: The next one. The next one, sorry.
Oded Benyo: I think that what we see or what I think that email to some extent is one of the best channels for customer communication. It's a very intimate channel. We expect it to be right, to be personal, to be meaningful, to be about me. And to that end, it's a great channel for brands to have a conversation with consumers and carry that dialogue and really leverage all that data that they're collecting. Right? I mean, a lot of personalization is just can we not have the same conversation over and over again as if you don't know me? Right? But can we actually take some of that information and plug it into that conversation and drive that content?
And I think that a lot of what we do is we try to work with customers to really increase the amount of time that they're looking at and move from this notion of a very episodic channel, which is how a lot of our customers and prospects use email today of, “Hey, I have a promotion to tell you about, or I have some other event to tell you about,” and move it too much more of a, “This is a personal communication line with you, Mr. Customer and we're going to conversate, we're going to talk about what it is that we think is interesting for you.”
And there's a lot of ways for brands to manage that conversation by picking up a lot of the signals that consumers leave. Right? So the other part of that conversation is consumers being able to they're opening the email, they're clicking, they're going to the website, they're going to the store, making a purchase, right? All these are clues and hints that consumers are leaving that you want to use as you're shaping that journey. And I think that makes it for a lot more interesting channel and you see very few marketers, at least I have, that really leverage that piece of information to carry the conversation over time.
Ginger Conlon: Yeah, that's a great point. When we were prepping for this, you and I were talking about the Goldfish Principle, which Don Peppers and Martha Rogers coined years ago, which is basically imagine a goldfish in a bowl, right? They don't have very good short-term memory. So there could be a plant in the bowl and they swim around, and they see the plant, “Oh, a plant,” and they swim around again and, “Oh, a plant,” as if they've never seen that plant before.
And so imagine that customer experience, right? If you're a known customer and you go back someplace and it's as if you've never had that conversation despite the company having asked for information or gathered information about you and now you're expecting them to know you to some extent. Right? So what are some of the conversations, Nicole, that you're seeing internally around personalization and email?
Nicole Perrin: So I've been having a number of conversations recently about exactly the kind of thing that you just described and the question of whether marketers actually care that they're doing this to consumers. So I have a couple of stories. I've actually been telling one of them for over a year, but it's still true. There's a DDC apparel retailer that I've shopped with before that I liked what I bought. I wanted to buy a second item, added it to my cart, wasn't sure, left it, got my abandoned cart email with a coupon that was only good for first time customers.
So why did they have my email address if I had never shopped with them before and they didn't check that against anything. I test that now regularly to see if they're doing it and I can confirm that as of two weeks ago, they are still doing it. I also have a very popular food delivery company that I order delivery from pretty regularly. I've been known to get emails from them asking if I want to order dinner tonight after I've ordered dinner that night via that service. So but I haven't unsubscribed. Right? And I'm still ordering so they're not stopping doing that.
And when I've talked to people about why this still happens, I mean, you would think that it would be embarrassing at best, but if they're just looking at their results and they're still seeing good results, and they're not seeing a huge amount of churn, and they're still growing as a business, they're still doing this. So some folks have told me that people won't stop doing this until they really start to feel the pain from doing it to consumers and with new companies that are start-ups in a high growth phase, that could take a while. So that's one of the conversations is whether you even really want to do it or whether you want to blast out these performance oriented messages anyway.
Ginger Conlon: Right, right. What are you seeing inside your company, Dane, in terms of those kind of conversations and your strategy around that?
Dane Mathews: Now, we try to be careful and smart about how we use email. The context for consumers are different in that they've chosen to sign up for something and you said no. Right? And so that's a great point. So what does it mean to be a known customer? Does it mean you get the same old stuff that everybody else gets? You just get it in an email format. And how do you just be smart and contextual about this consumer? Because it's a really important one.
For us we think a lot about if you sign up for email, what do you want that's new and different than no one else could get? And it has to go back to consumer value and it has to go, “How are you using any channel, emails, and is our topic, but how are you using any channel to actually derive value?” And so is that to get a behind the scenes look at a product or the company, or some innovation, or some things they're thinking? Is that a way to solicit more information? But you have to provide a real strategy for email. It can't be a me-two channel strategy meant to just supplant what a website or an app is going to do.
Oded Benyo: Yeah. I think to some extent whether we like it or not, money talks and generally speaking, even though email is perhaps the best channel marketers have in terms of return, in terms of ROI, right? I mean, what they have to invest versus what they get back by just more than 30 to 1 sometimes. In the grand scheme of things, if you look at all the other channels marketers are spending on search and display, and T.V., and whatnot, email is actually not a mega expense. And as a result, it just gets less attention.
And so what it ends up being is it's about are we annoying people as opposed to really focusing on the customer experience. And this is a great challenge that people pay a lot of attention to, how can we actually enrich the customer experience as opposed to focus on I can get away with it and it's not very expensive, so just blast everybody. And I think that sort of notion has to change as we start thinking about consumers in a broader scale and what you were saying, right, across multiple channels and what kind of conversation we're having with them and how it all fits together.
And it's hard because it's hard to collect all those signals and synchronize them, and then do it in a way that makes sense. And particularly in channels like email that by and large are not expensive. I mean, I've talked to marketers who told me I have my pay channels and then I have email. Right? Which is crazy. We have to start thinking about it more from a consumer perspective and how are people experiencing the brand as opposed to what costs more.
Ginger Conlon: Right. And that's interesting what you were saying because if you think about marketers think a lot about response rates, right? And so if you have, I don't know, a 7% response rate to an email, are you thinking about the 93% response rate from the people who didn't open or didn't click through, or whatever happened, right? They're telling you something also and I think that gets a little bit lost when you get the positive outcomes from that 7%. Right? So Oded, you earlier mentioned using email as part of a conversation. Can you talk a little bit about what that means?
Oded Benyo: Well, I think that part of it, tying to what you say about response rate, it's changing how we look and how we measure the channel, right? And really focusing on, if you will, increasing the aperture of time, increasing the amount of time you're looking at to understand how people behave, how they're engaging with the brand, that completely changes how you would probably run programs and see people or consumers evolve within the proverbial funnel, if you will.
And so to that end, again, very few people actually think about it in that context, because we're so quarter driven and we got to get the revenue in, and whatever the case may be, but if you start looking at the relationship as a long-term relationship and measuring all the signals that you're getting from consumers over time, and how they're shifting in their purchase cycle, if you will, you end up getting very different results and probably, it will drive different behaviors.
And what we see as you start looking at the longer period of time, it changes the communication. I mean, an easy one is frequency, right? Where brands send or some brands, send a lot of email because every single email, I'm going to get it a little bit more response, right? It's hard. It's very tempting. But the flip side of it is it's very hard to measure what is it doing… how many people are getting those 93% or that you mentioned or however many it is, that they're turning off because I'm done, right? At some point it just becomes noise and calibrating that is very difficult if you look in a very short period of time.
Over time it becomes a lot easier. And one of the ways is to help marketers test some of these ideas with a smaller audience, if you will, to understand how that communication goes and that's one aspect, and the other is, of course, trying to coordinate across multiple channels.
Ginger Conlon: Right.
Nicole Perrin: I think in addition to looking at the wider time aperture, also looking at a wider funnel aperture or activity, I've talked to some people about these email KPIs that we're all also familiar with. Click rate, open rate, click to open rate, and whether those are still valuable or whether we should be thinking of those more as vanity metrics at this point, and looking to see what the next action, the second and third order actions are after someone looks at an email or doesn't look at an email.
I talked to someone earlier this summer who told me about how if they had looked at those KPIs, they would have optimized in the wrong direction because actually, the people who were most interested in the offer didn't open the email. They saw the subject line and went right to the website, and bought. So those KPIs might not mean what you think as you start looking at attribution in a more holistic way.
Dane Mathews: And I think that in order to be able to do that, you have to step back away from the metrics. You have to understand what is the intent that we're trying to do? What is the goal of this communication and the goal of email? And then you can understand, well, how do we measure that goal? And some of that measurement does happen in clicks, but with 93% to 99% of all the content, there aren't a lot of clicks. And so how do you look more broadly and look for other places to get clues that you're either on the right track or you're on the wrong track.
Ginger Conlon: Right. Right. And that email subject might also be a reminder from something that happened in another channel and the email comes in-
Nicole Perrin: Totally.
Oded Benyo: Reinforcement, right?
Ginger Conlon: Yeah, yeah. Complex. So there's no conversation about personalization and relevance that can't include talking about machine learning and AI. So we're going there now. Nicole, what are you seeing in the market in terms of how marketers are using that in the email channel and connected?
Nicole Perrin: So I'm seeing that email is often the first or one of the first channels where marketers are applying those types of tools for personalization. Probably because it is that we're a course, because it's something that they're very comfortable with and an audience that they're pretty familiar with, and they tend to add more channels from there, again, in that machine learning sense. And basically, you just can do things with machine learning that you can't do any other way. Both in terms of the insights about the data that you get, the clustering that you see, the segments that created, but also in terms of figuring out what content to feed those people and literally how to do it at scale.
Yeah. Dane, are you using machine learning in your marketing?
Dane Mathews: Very, very early, very early tests with machine learnings. The first, our first step was really automation to pull the humans out of some of the hands on keyboard activity so they could really do strategy and they could talk a little bit of a more, but they could create better platforms and communication strategies for our consumers. We're starting to use machine learning in terms of giving us a better sense of doubling down on response rates and it's just an evolution for us. But I think the thing that is really interesting for us that actually gobbles up a lot of conversation around the table is the efficacy of content.
And we're actually looking at leveraging machine learning and more science-based ways to understand system one versus system two thinking and then some of the hedonic cognitive research that's out there. But how do you communicate to consumers when you really do have maybe less than a second to get something across? And starting to help put machines in that space. Only because we all have a personal stake in the game on the content that we think is either super cute or really resonates with us, or whatever it is. But what we found is much of that is irrelevant and that a lot of some of the technology we see today can help us get out of the game of nibbling around the content and help get content out the door, and get it in market.
Nicole Perrin: Okay. So for those of you who don't or who aren't familiar, can you explain system one versus system two thinking? It's good to know.
Dane Mathews: All right. System one thinking is where we are most of the time and a lot of what Conagra does is we sell food items that are in the supermarket. Much of that information is you don't analyze it all. And so system one thinking really just talks about it, it's just in the background. You're absorbing information and it doesn't really hit your subconscious. And system two thinking is where consumers are actively thinking and evaluating. It's good because sometimes we'll get into this little pickle around where we need to show everything in this ad unit.
We need to show how much fat it has, how many calories, we need to show it in context of the kitchen. The kitchen needs to have white cabinets versus dark cabinets and the ad needs to be blue because the packaging is blue, and there's all this stuff. And then all this comes in and we just get around the axle on creative. So by understanding that, we pull ourselves out of and we use a little bit more science where there was really a ton of creativity.
Oded Benyo: I think one of the points Dane made earlier about you got to know what your objectives are and at least what I see out there a lot is machine learning is like the buzzword du jour, right? Everybody needs some. I mean, you literally have people out there, they just need to buy some machine learning. They don't really afford but they need to buy it because somebody told them that it's cool. And to that end, you have to start with the end in mind. Why do you want to communicate and how is machine learning is going to help you and it's much easier to do when you think that way.
And then I think machine learning is really the way to drive scale, right? You just can't make millions of decisions with humans. But the other side of that coin, which is what I think is the hardest part for most marketers honestly, is about giving up control, right? Because if you want the machine to make decisions, it means that you can't or you're going to have a limited amount of decisions that you can make about the overall look and feel of the email, the message, whatever it is that you let the machine the on. And a lot of brands struggle with it because they are so used to controlling every… particularly if it's brand messaging, right? They're used to controlling every aspect of the ad, right? If it's blue or yellow in the background relative to the packaging.
And sometimes the machine will get to counter intuitive decisions, right? Because some people like it yellow in the background, even though it makes sense that it should be blue and how come you don't know that? But that battle in every organization usually takes a while. And what you see is people are adopting machine learning in a very small part of the creative, whatever the creative execution is and then running a test. And if you change one element in the campaign, you're going to see a change, but it's not going to be mind-blowing.
And then I go, “Okay, we tried, it didn't work. We're going to go back to the old way,” versus going, “Okay, this is a learning experience. We are going to try a little bit and there's going to be a little bit of impact, and I'm going to continue so that we're educated about it and we eventually get to wholesale changes in how we serve creative, how we communicate with consumers.” So there's a long way to go and a lot of it, I think, has to be about marketing organizations getting to the realization, “Hey, I got to give up some control in order to get the results that I want and have the courage to do it.” And it's hard because there's risk.
Ginger Conlon: So are you seeing within your customers, different types of marketers who are more likely to be comfortable taking those risks, personality, age, whatever, anything about them that they love technology and gadgets, so they're more willing to take the risk or they're just generally risk averse. Like any trends that you're seeing in-
Oded Benyo: Well, certainly, you have different organizations that are… culturally are more risk seeking or risk averse. Some organizations like to brand themselves as ones who try new technology and constantly try new things. And those are the, if you will, early adopters that are usually happy to test these kinds of things. But a lot of times, like you say, it boils down to a person and the personality, and some people like to play it safe and will follow the trend after it's been proven, and they want to see all the case studies and all the information, and others are willing to take a risk.
So we've seen both and it's about adapting to whatever a client wants to do. And I think for me it's more about getting them to sign up for the journey. Like this is not a binary decision. “Okay, tomorrow we have machine learning. Life is great, right?” It's going to be a journey over time to figure out how you're going to use that.
Dane Mathews: That's absolutely right. And that journey, I think, we think about it as the future of work, right? So like what is your role in the future where there is automation or there is machine learning? It doesn't mean that people go away necessarily. You can have the big red override button that you can hit. But it's really about, well, so then what tasks do you do and how do we allow you to do the tasks that quite honestly just a ton of tasks that machine learning cannot provide.
And it isn't necessarily creative. It works within its dataset and it tries to find patterns within the data set. So work beyond the data set and continue to fuel that machine with new and different iterations that help make it smarter. And that's it's you have to say machine learning, but then you have to put in context how people's work lives will adjust and change along this journey, just like you said.
Oded Benyo: Yeah.
Ginger Conlon: Right. Right. Interesting stuff. So as consumers, we often feel inundated with communications, whether it's email or just what we're seeing online or walking down the street. But companies do have control over much of that, obviously. How do you make those decisions internally, Dane, when you're thinking about, okay, we want to reach out, we have a lot to talk about, but we also don't want to inundate our customers. And how much does the relevance and personalization come into play in that regard?
Dane Mathews: We've built our advertising, we'll say, we'll build our communication philosophy around relevance. And this idea that from email all the way up to a social or CTV ad unit, we actually are looking for consumers who are displaying signals, behaviors. So it really all starts there. And we've worked with our consumer insights team to really organize and codify that from a behavioral science perspective. We take that information and we go to all of our favorite partners, the Pinterests, the Facebooks, the Google searches of the world, and we actually work with them to identify signals.
When we're responding, the consumers were already relevant. We've already past the first gate. And so we look for places where consumers are already acknowledging themselves and what they want to say, and we're using that as the bedrock for now with our response to that. Instead of trying to make it up along the way and try to say, “We're trying personalization because we have an AB test that's out there floating.” It's really about what is that connected to? And for us, it's really connected to what's the consumer saying in this context and how can we leverage that along with our system one versus system two thinking to really craft the right kind of communication cadence and then measurement plan.
Ginger Conlon: All right. So Nicole, you were saying earlier that it's tempting to just fire messages out there, but what are you seeing in the market in terms of where brands are trying to be more reserved and more tactical in terms of the relevance?
Nicole Perrin: That's a really good question. I think different brands have really different ideas about what an appropriate frequency is. And one thing that I found interesting as I was researching email over the past couple of months is that they even come to those different ideas based on data. Now, I didn't get to see their data to see exactly how they made those decisions and whether I thought they were the right decisions, but I would talk to one brand that would say, “We decided based on data that one email a week is the right number,” and then the next person says, “We decided based on data that two emails per day was the right number.”
Which is like a huge difference, and from my consumer point of view, two emails a day from any marketer, and this was an apparel retailer, that just seems like a ton, right? And huge brands, some huge department store brands send 40, 50, 70, 80 emails a month, others send five. There is just huge variation. And so it's hard to understand how some of those decisions do get made. But I think it often depends on whether someone really cares about that churn rate, that unsubscribe rate, and probably about whether they're looking to see about those ultimate actions driven by the email or just looking to see like, “Did this raise my conversion rate by half a percentage point?”
I often talk to people about personalization who say like what we're talking about in terms of effectiveness is like a one or two point difference, which for marketers is huge. But when you think about that from a consumer perspective, there is no wonder why consumers don't feel like stuff is personalized. You wouldn't notice if something was 1% more likely to make you interested in a product. Right? That feels pretty much the same as a consumer even though aggregated for a marketer, it turns into real business results.
Ginger Conlon: Right. That's great. So let's talk a little bit more about metrics. There are so many ways to measure success clearly and it's unique to your organization, but any favorite metrics that each of you have? Oded, what are you seeing in terms of across your customers?
Oded Benyo: I think that you still see a large amount of customers, particularly ones who are selling product, whether it's hospitality or retail, or whatnot, that are very focused right on transactions and return from that perspective. And oftentimes, it's very direct response oriented. Like I sent you the email, what did I see in response where we've been working with a lot of our clients and of shifting that perspective to, as I mentioned earlier, looking over time and starting to look more at engagement as opposed to necessarily, did this email right now cause you to take an action such as purchase?
Because you may be in a different place in your personal cycle. And so some of that is understanding where people are in their purchase cycle and how do we communicate with them accordingly, and move them along. And sometimes it's just reminding them, “Hey, we're here,” other times it's you want to be in that consideration set right when they're about to make a decisions, and sometimes they're about to make a decision soon. And customers leave all kinds of clues along the way about their intent and what they're trying to do, and some of it you have to guess.
That's where actually machine learning and modeling is phenomenal because you can learn from the behavior of others to make some educated guesses. You're not going to hit it 100% of the time, but it at least allows you to calibrate the communication a lot more. And so we've been able to look… The key element here is to really increase the amount of time that you're looking at.
Ginger Conlon: Yeah. Dane, what are you using and what are your favorite metrics?
Dane Mathews: We use the same, but I think really increasing your aperture or your time allows you to be just a little smarter. We also in partnership with our customers, we'll say the Kroger's, the Walmarts, the Targets of the world where we sell to the consumers. We also think a lot about what's the role of a Kroger email program versus a Conagra email program? What are they meant to do? And so and working with those partners allows us to potentially have a better sense of we'll say conversion, if you will. And so a lot of our time is really spent with them understanding how they think about email, also increasing the aperture and how do we think about how do we grow our business over time, not stay tied to open rates, click through rates and click to session ratios.
Ginger Conlon: All right. Thanks. All right, so let's jump into a couple of audience questions. Thanks for your questions. Data privacy, that's a good topic. How should data privacy developments play into how we think about levers for personalization? Nicole you want to grab that one?
Nicole Perrin: Yeah, sure. So, I mean, obviously it's going to have big implications, right? We would be talking about things like ITP, things like GDPR. So first of all, email is a permission to channel, but that doesn't mean that you can just do anything with email or put anything in an email and still be GDPR compliant. So when we're talking about personalization, you're bringing in data signals from other places. You need to have the proper consent to use that data, process it for this purpose. So again, the things that are going into the email tend to be affected by that.
I've talked to a lot of people about ITP over the past few months and it's putting a renewed emphasis on email because email is a universal identifier and we don't have a lot of those left that work that well anymore. So folks are both leaning on email in terms of marketing. Also in-email advertising is becoming a little bit more popular because, again, that's a way for advertisers, essentially in like a programmatic display channel to still reach them, one, with that universal identifier. It also means marketers are even more eager to gather your email address so they can find you again later.
Ginger Conlon: All right. Oded, how do you see machine learning changing the personalization of creative content for consumers across ad channels?
Oded Benyo: So I think that there's two aspects to it. One is machine learning or having really big data, right, allows us to scale of making decisions across multiple channels, which is still a pretty hard thing to do if you need to do it at scale. And so we're definitely starting to use the technology to make a lot of decisions and to determine what content we're going to share with people depending on the channels. And we can also anticipate or predict what channels we're going to see them in. Right?
I know that I'm going to see you X amount of times this week on display. I know that I'm going to send you emails and you're likely to open them at certain times. And so using that to actually predict how you communicate with people better and become more effective. The other element which I think is super interesting is around content. I think we're just scratching the surface right now when it comes to content. For the most part, implementations out there are about picking from a list of choices, right?
So there's going to be, whatever, five images and the machine is going to decide that you get an image one and she gets image two. But I think where we're going is as we get more courage around this, is actually having the machines help us determine what content we want to display. How does it come together? It's like one way to think about it is right now we, by and large, make decisions in a single dimension, right? And we're going to transition to a multidimensional type decision where you're looking at three or four parameters when you're determining content. And it's much harder to predict how the end result's going to look, but I think it makes it a lot more interesting.
Ginger Conlon: Absolutely. Well, thank you so much, Dane, Oded and Nicole for all the great insight today. And thanks again for joining us. Have a great rest of the conference.