Navigating the Data Forest: Building Conviction in a Digital Age

Lena Waters: I think as leaders, we have to think about that kind of decision-making environment that we're creating for our teams so that they do have the right resources, but we're not overloading them with false expectations of we have to try everything. We would drown. And so, we're looking for, I think, clarity. We're not necessarily looking for everything.

Eric Weber: You're gonna have to have conviction and a hypothesis and a belief about what is true. And you should still use data to try to refute that or gather evidence against it. But you're gonna have many fewer situations where the data is going to be like, here's the path forward. Like, guess what? That's not gonna happen. And I think the more that we embrace that, we're gonna be more successful in one or two years when a lot of this visibility is actually gone.

Kailey Raymond: Hello, and welcome to Good Data, Better Marketing. I'm your host, Kailey Raymond. The partnership between marketing and data teams is complex and sometimes underappreciated. With the rise of performance marketing and data-driven everything in marketing, alignment between these teams is essential to ensure better decision-making and ROI tracking. But what might be even more essential than getting the data right is the human side of the equation, ensuring that data enhances rather than replaces sound business judgment and innovation, and teaching teams to lead with a strong hypothesis and conviction in their solution. Grammarly's Lena Waters and Eric Weber joined me to discuss their collaborative approach to aligning business and customer priorities, the implications of a cookie-less world on marketing attribution, and the balance of data insights and human judgment to drive meaningful customer experiences.

Kailey Raymond: Today, I am excited to be joined by two leaders from Grammarly, CMO Lena Waters and Head of Data, Eric Weber. Lena leads marketing strategy, focusing on delivering the brand promise to customers, driving demand for products, and growing revenue to scale the company. Eric leads teams of data scientists and software engineers who are responsible for improving how to measure and iterate on the product to create the best ever communication assistant. Lena and Eric, welcome to the show.

Lena Waters: Thanks, Kailey. Great to be here.

Kailey Raymond: Excited for you both to be here. It's a rarity that we have a marketer and a data person come together on the show.

Eric Weber: Kind of dangerous. Be careful what you wish for.

Kailey Raymond: Yes. It'll be a fun perspective to get you both chatting about how you all work together at Grammarly. So that's where I want to start. I do think that there's a very natural work and process that you all build together between the two teams, data and marketing together. But how do you all work together at Grammarly? Anything that you want to shout out in terms of your working relationship and some of the big projects you all work on? 

Lena Waters: Well, Kailey, when you asked me to come on, my first thought was, well, of course I have to have this conversation with Eric. We're great business partners. Our teams work together really well. It seems silly to me to think about creating a great customer experience without really looking at data. And of course, everything can't always involve data, but that's sort of the magic about it is making those decisions well. I think when Eric and I and our teams are talking, we're always trying to find ways to align around how do we think about putting not just business priorities first, but customer priorities first. And you don't really know what your customers' priorities are unless, A, you literally go out and talk to them, which we also do. But, B, you look at a lot of data about what your customers are already telling you, and there's always a really good trail to follow there. And so, we don't really look at data teams having a seat at the table with marketing. We really think about us coming together and using data like any one of the tools that we would use in any kind of marketing program.

Eric Weber: I think in my first week at Grammarly, I was like, I need to talk to Lena. And it's continued since then. I think a big piece of data being successful as a company is seeking out people and programs and areas of the business that will benefit from data. I think in companies where the relationship is not working as well, you would have someone in data join and wait to be tapped on the shoulder. I think in a successful version of partnership, you're coming in and being like, I'm not successful unless I go find my partners internally. And I think it's been a really productive back and forth in relationship. I think you'll probably see that as we chat about what we work on together here.

Kailey Raymond: I love it. Yes, it does seem like not only are you really aligned, but you're also aligned about who is at the center of attention at Grammarly, which is definitely the customer, making sure you're solving problems with them in mind. One of the other things that, of course, is impacting how we're making decisions besides customers perhaps are some of these more macro trends coming in, industry trends, consumer trends that are impacting the way that marketers and data teams make decisions. I want to hear from both of you, if you have any trends that you think are impacting your teams, what you're listening to, what you're caring about right now.

Eric Weber: I don't think this trend is new. I think the trend of really big promises being made by new technologies is something that's always going to be happening. I think the question of how do you handle it, understanding what it can actually do for you versus what is just noise is really critical. Even now, as we think about, hey, we have AI models, we have all of these new things that we can do. A lot of the fundamentals about measurements and what parts of the funnel and aspects of the funnel you can actually measure remain fairly consistent. And so, the trend might feel new just given the types of technology that we're talking about, but the fundamentals are still pretty much the same. And I think we're finding our way through this, which is we have a whole bunch of opportunities to do things differently, but we still have to solve some very consistent and clear problems that have not shifted day to day or week over week.

Kailey Raymond: So, you're seeing consistency there in making sure that you still need to do the hard thing, which is making sure that we are measuring our success and understanding how things are coming together. Lena, anything to add? 

Lena Waters: Yeah. I'm with Eric in terms of when you think about what trend means, I think there's this tendency to think what's the new thing, what's the cool thing, what's the thing that everyone else is doing that I'm not doing. There's a lot of FOMO that's implied when you say the word trend. I think where Eric and I really see eye-to-eye is our tendency to zoom out and look at this bigger picture of, are we in the midst of a trend that has been going on for so long and is so pervasive that none of us realize that it's happening anymore? And how do we bring self-awareness and awareness to the company around what we're going to do that is going to impact that customer journey? I think for me on marketing side, the thing that I see us being in the midst of is buyers want to use digital channels instead of relying on more traditional channels. They want to learn for themselves. We all know that people are going to go online and they're going to talk to their friends much more than they engage with the company. And so, because of that, the customer journey has become more digitized. Then what happens? Then you have this trend, as you'd say, where people think, well, more data is better. And so then you move into people thinking that everything has to be knowable. And if something isn't knowable, any threat to not knowing something causes panic. And then organizations, you see how this snowballs, right? 

Kailey Raymond: It's a vicious cycle.

Lena Waters: Organizations, yeah. Then organizations feel this pressure to fill the gaps in all of the data. And so, we get to this point where we're so busy trying to figure out all the things that we need to know that we flip away from what we should really be doing, which is how do we ask the right questions? Did we actually go into this with a hypothesis? Do we understand what we want to measure? If we had that data, would it actually answer a useful question or would it just make us feel secure that we have a lot of data? So, I think there's a lot of proliferation of tools out there, marketing in particular, but it has led us to a little bit of an extreme over-reliance on tools where we're so busy looking at dashboards and trying to measure everything that I think sometimes we've forgotten how to ask good questions.

Lena Waters: I would say that the trend is turning to an over-reliance on data and technology, which are super important, but not at the expense of asking the really important question about how are we going to create a great customer experience that really adds value, that teaches the business something that they didn't know that we can come in and help them with? And if you focus there, it'll help you ask better questions about the data and it'll get us away from this trend of more is better.

Kailey Raymond: I love this concept, because I often think that folks can fall into the trap of saying, we just need to learn know everything. We need to collect everything. We have all of the data and what you're saying is, actually, maybe you don't need, "all of the data". You need the right data to solve the questions and the problems that you need solving. Eric, is this a refreshing point of view from a marketing leader, from your perspective to hear? I imagine you've been in rooms that have other stances, perhaps, in your career.

Eric Weber: Every week when I talk with Lena, it's something refreshing. And I think this is particularly true in this case. I'll put it this way, I think the idea that the answers exist in some place and we just need to work harder to measure them is really dangerous for a business. And if you... There's two versions of what could be true. You could go into a week and you could plan to look at all of the data and every dashboard and every metric and say, wow, what should we do now? You could also go into that week and say, here's what we should do now. You should then look at data and be like, does this track with what I believe? You should be to refute your hypothesis, stress test it against what's actually happening. And I think we get to the point where we operate very much in the former, where people just look at this massive amount of information and hope that the right direction is going to emerge from that. It's probably never gonna happen. You have to make bets and you have to have a strong conviction on what we should do. Of course, try to disprove your hypothesis.

Eric Weber: Try to refute that with information. But at some point, you're not gonna have a magical path that's laid out by a certain series of dashboards. You have to actually be willing to say, this is where we're going. And that takes a lot of guts, and honestly, I think the more data we have, the easier it becomes to hide behind it. And that is a pretty worrisome trend.

Kailey Raymond: There's a lot of discipline here in what you're sharing, which I think is something to take note of, is making sure that you're following through on this too. Because you're right, this kind of mountain of data, you can get lost in it really easily. Lena, you were pointing towards this trend of digitization with lots of new channels opening up and perhaps even the buying cycles of B2C creeping into B2B and the way that people are actually starting to think about their purchasing decisions. I'm wondering, how do you think marketers can avoid this trend of over digitization? 

Lena Waters: I think this, for marketers in particular, this over-reliance on tools to measure everything, I think people are very focused on, there's a customer journey and if only I could connect all of the touch points of that journey, and then if only I could connect all of my systems to track what happened, I would be able to discover the magical, perfect customer journey that would lead to more sales. And if I did that, then we could grow. I think it's led to this very false idea that if we can see everything, we know everything. I think the problem with that is that when people choose to work with your company, to purchase your solution, they're taking a risk and they are trying to gather information to make a decision, often within a group of people, which is the hardest way to make a decision. And they're trying to figure out how much risk they want to take and how much they want to de-risk their decision and how do they do a good job on their side. Very often, we are so busy trying to connect the content and experiences that marketing has given them that they've consumed and we've mistaken content consumption for a decision-making process and intent to buy.

Lena Waters: And it's really tricky because that's what all the systems are set up to tell us, who visited our website, who signed up for the webinar, who came to an event, who downloaded an asset? Very rarely does it say, what is the stage that somebody is in terms of their belief and understanding about what we're trying to tell them and how are we asking questions to demonstrate that that's what they are doing? It's largely a long string of content consumption together. And so, this goes back to the earlier point of digitization is great and it helps you understand a lot of things, but if you're not asking the right questions and you just consume the data that someone else has given you, you're not going to get any closer to understanding people's decisions or why they take risks. You're just gonna have a lot more data and you're not gonna know what to do with it.

Kailey Raymond: This is a really interesting point. I really like this concept of you don't want to conflate somebody's actions with just consuming content with this just perhaps absolute intent that they might have for your software or whatever else you're selling. I think that you're right. It's exactly what the attribution systems are built for and put in place for, but it's not always that linear for us. And Eric, I wonder how you think about helping to tell that story and piece together what Lena is sharing with us about this perspective of over-digitization and marketers.

Eric Weber: Taking a step back, if you look at the picture of what do we know about people and customers and where they behave and how they behave, we know very little if you look at the broader journey. We know a lot at the moment that they do something with us, potentially click an ad, visit a webpage, download a client or an app, but there's many things that we don't know about that person and what they did. And there's also many people who never got to a place where we could measure that we also know reasonably little about. And so, I think there is a selection bias that happens if you focus entirely on what can we measure because you're looking at a relatively small group of people compared to who you could have been talking to and making many decisions based on what that population did, which I think works when you are talking about product iteration and development and things like this, but it misses a big part of the conversation, whereas we could optimize and develop everything for the group that we have now, but it's gonna miss out on what are we not doing that could have affected this much larger group of people. The second thing that I think about here is this tends to get a lot more pressure when growth gets hard, when things are up and to the right and people are like, great, numbers are up. There's actually a relatively surprising lack of pressure on we need to see every portion of a customer journey.

Eric Weber: I think if you think about this, if we're gonna take the time to measure something, we should believe that it's a durable need for measurement, that it's not just in response to a particular problem or moment in time, but something that we think is worth knowing, go forward, regardless of if we're growing or having growth challenges. And I think that's a hard conversation to have because engineering problems are hard and measurement problems are hard, but I think those two things, one, the bigger piece of what are we not measuring is much bigger than what we can measure. Then two, if you're going to spend time actually digitizing something and measuring it, make sure that you have conviction that it's worth knowing, go forward, and it's not just in the moment decision.

Kailey Raymond: This is a great insight. Ultimately, I think the short-term versus long-term that you're describing right now is something that is easy for your right organizations to perhaps glance over, especially in times when that metric isn't going up into the right because everything feels like a fire drill. So, I really like that concept of durability, making sure that the dashboard doesn't get looked at once. You want to make sure that it's something that can be leveraged time and again. I guess this is a metric that actually matters.

Lena Waters: Here's what's really interesting about what Eric's saying is that, he's really reflecting the critical problem with all growth businesses. When you think about the customer journey and you think about this upside-down triangle that we love to show everyone, the lots of people at the top, the mid-funnel, the bottom of the funnel, we've all seen this upside-down triangle. You think about the percentage of your questions and your data insights and your dashboards. They are highly skewed towards the smallest part of the audience down at the bottom at the tip of the triangle. And what Eric's so rightly calling out is that when a business is under pressure, everybody focuses on the bottom, on the people who are theoretically the highest down the journey, the most convinced, the most willing and about to convert, the highest propensity. And the problem with that is that repeated focus on a small group of people who, in all chances, were probably going to convert anyways because that's where they are. It leads to this complete neglect of the largest audience at the top of funnel who are in their most vulnerable stage because we all know everyone comes to the table with a consideration set and that's usually about three vendors. And what is it, 80% of buyers, B2B at least, who come to the table have three buyers in mind and something like 90% of people buy one of those three.

Lena Waters: And we know that, you think about your own personal purchases. When you go out to buy something, you don't think about 12 brands and create a pros and cons column. I mean, some people do, [chuckle] but most people have three things and then they pick one. And that's how the human brain works, right? There's a reason that we look at these smaller choices and small numbers. We can't deal with all of that. People on the B2B side aren't any different. And so, that moment where you have the chance to position your brand against others to make people believe and feel something, as soon as you start to get into that idea of brand or awareness or persuasion or positioning, none of those things get measured in a dashboard. You can measure them. There is data around them, but you don't really usually buy a SaaS tool for it. And it doesn't usually come up in the boardroom. And those concepts make most business people very uncomfortable because they can't measure them. They can't look at a line that goes up or down.

Lena Waters: And so the problem is that this over-reliance on this data and the digitization of the experience means that we are neglecting customers in their highest moment of need, where most of them live, because we think that the spot that you can use data to measure makes it therefore more important and that that is the most dangerous thing that the over-reliance on data means, and you can measure things at the top of the funnel. It's entirely possible. There are just different ways that you have to go about it and you have to step outside the box of refreshing a dashboard. And I think until people do that en masse, you're going to continue to see strong brands that do really well by winning hearts and minds and helping people understand the value that they bring. And then you're going to see everybody else chasing the same small amount of people by going after them with content consumption, and this has been happening for years, and I think there's just very few companies that have the alignment and the leadership in order to make that change and put the investments where they need to.

Lena Waters: And start putting more of their data insights equally out across the customer journey. Every executive should be asking themselves, how much of the customer journey can I actually see and understand, and am I asking for data on it, or am I accepting a PowerPoint when somebody comes to me about some bullet points and some ideas? Why aren't we asking for data at the top of the funnel the same way we're asking for it at the bottom of the funnel? It's a mystery.

Kailey Raymond: This insight is so, so important. I've seen some stats on this recently, which is really just showcasing the over-reliance and over-investment in performance marketing and perhaps people spending a ton of money on demo request ads and things that, to your point, are very, very bottom of the funnel. And what that's done is that's taken away a lot of money and time and dedication and resources to build your brand and build a customer experience that really matters and building this kind of trust with people. And one of these kind of connected trends that I think is related to what you're saying here, which is really around top of funnel attribution, it's not easy. It's something that can be done and it's something that we know we all inherently need to invest in. We have to have conviction to be able to do that, but it's this trend around cookie lists, I think, is what marketers are often very afraid of.

Kailey Raymond: Cookie lists, of course, is a little bit, you know, Google's playing with us a little bit as to how much it's going away. But thinking about measuring this effectiveness in a cookie-less world, I think, is kind of directly related to what you're saying. But overall, just this trend of measuring these top of funnel tactics and programs that you're running. I'm wondering if you could share some examples with how you think about that at Grammarly. How are you measuring the effectiveness of some of these more brand level things that you know you need to do, but are much more challenging to actually get down to the performance data that we're all used to and we know and love in that upside down triangle? 

Eric Weber: One is, I think every company believes that they should be different than every other company. And like, we will find a way to get past and by this cookie-less world because we really. Like, this is not accurate, right? I think the realistically, most people are playing a very similar game. And if you spend all of your time trying to be like, we're going to play a different game than everyone else, you can burn a lot of time, a lot of energy and ultimately come up in exactly the same place that everyone else is. So, whether Google decides to do this now or a year from now or whenever they get around to it, like we're going to end up in that place, some version of cookie lists. And I think we should just start operating like we're there, right? Like operate in the place, understanding that's the most likely outcome and start building solutions that take that as a condition of operating. And part of that is like you're going to lose visibility. And there's actually very few ways entirely around it.

Eric Weber: You can start measuring things at the cohort level instead of the individual level. Like that's a very, you know, reasonably straightforward thing to do, but also like understand you're going to be making decisions with less complete and precise information than you may have before and get used to it. And it's a, I think this idea that you're going to look at things the exact same way and we'll have the same ability to say, like, this is exactly what the trade-off is. It's just not going to be the case. And so, when I look at what our teams are doing on the data side, it's actually just like building for that type of world, which is you're not going to have the same level of precision. And this back to an earlier point, you're going to have to have conviction and a hypothesis and a belief about what is true.

Eric Weber: And you should still use data to try to refute that or gather evidence against it. But you're going to have many fewer situations where the data is going to be like, here's the path forward. Like, guess what? That's not going to happen. And I think the more that we embrace that, we're going to be more successful in one or two years when a lot of this visibility is actually gone.

Kailey Raymond: I really like this. And I'm wondering, Lena, how you make your teams get used to this discomfort as well? This is obviously like a pretty big shift for a lot of marketers to make and feeling not like they have the complete control over that or the perception of the complete control that they have over the customer journey that they may have used to thought they have. What are you doing to make sure that you're acknowledging some of those limitations that now exist and handling the emotion that might come with it? 

Lena Waters: I could be fooling myself, so we're going to have to go check in with the team after this, Eric, but I see a remarkable lack of agita about a cookie-less world. I mean, one, we've been living with this hanging over our heads for so long now that I think there's just an absence of emotion around it. And there's now just a endurance game of, well, we've had lots of time to figure it out. And I think there's just a practical reality here that says, hey, people bought things before cookies and people are going to buy things after cookies. And so, you just have to get down to these first principles, which say, if you can discover groups of people who have needs and you can position yourself as some of the best providers to solve those needs and you can articulate those, a lot of this is just getting back to business and marketing 101. And to Eric's point about cohort, cohorts are great and holdout groups are great and A/B testing is great.

Lena Waters: And you can do a lot of these things that still make sense. Coming back to the issue of control, which we know is an illusion, but it doesn't stop us from trying to get it. I think to us, Cookie says, I can tag a person and then know things about a person forever so that everything I do is somehow tied to this individual person. And that's not always going to be true, even when you have Cookies. So, if that's not true, then what's this thing that you're losing? And I'd argue, what are you doing now with the data that Cookies has given you? And is that the thing that's making your business fly? And all of a sudden it's not going to if you don't have that anymore? I think very few businesses are in that position. So, I think we have to get out of this knee-jerk hysteria, which goes back again to this theme of over-reliance on technology. Cookies are just another element of that. And it makes you think about what is wrong with the fundamentals of our program. If the loss of one such thing isn't there, we should all be building more robustness into our go-to-market strategies so that you don't have that kind of hysteria. But I don't think that's the kind of thing that we're experiencing. And I think we just have to take a very practical approach.

Kailey Raymond: I'm glad to hear that. For me, what that says is you're building a culture around this. This is something that you're distilling into your teams through your words and your actions every day to be able to take what marketers used to probably think was a risk and investing in something that was unmeasurable and making it part of kind of a daily action. I'm wondering how you bring that up from your teams to people, perhaps like the board, who are also looking at this upside-down triangle and probably asking you a ton of questions like that. How are you addressing it at every level of the organization that have been relying on this data abundance and these dashboards forever and for always? Are there different strategies in making sure that you're articulating this across the organization? 

Eric Weber: I think if you're looking at different levels of an organization, even the top level of a company is going to wanna understand, are we spending our money wisely? Are we investing our next dollar in the place that makes the most sense? Dashboards, metrics, all of these things can tell part of that story. But ultimately, we need to be able to have conviction that, yes, we are working with state-of-the-art, and state-of-the-art does not mean the greatest technology out there or the fanciest models, but it means that we have pretty strong conviction that we can say we're spending our incremental dollar in the right way. And so, if we can say that that is true, I don't think there's as much concern about like, oh, you don't have cookies or you don't have the like all these other things can get abstracted away. And so success for me is we're able to say yes to that. Are we spending our incremental dollar in the right way? And in many cases, abstracting away the details behind like how we know that.

Eric Weber: In some cases, it's important, like we can prove it. But in many cases, it's having that confidence to be able to say that. It is an interesting message when you're talking to like, I operate like an engineering org where data is part of engineering. And there, like there's always a focus on the technology and the latest innovations. Here, it's almost creating an openness to say like, it's okay not to keep up with everything. We need to be able to answer this question of like, are we investing our money in the right way? But ultimately, that doesn't mean that we need to change what we're doing every single week as a new method or a new technology or a new vendor comes out.

Lena Waters: And I think Eric's really speaking to a culture of leadership that is driven by a strategy, not driven by the latest shiny object. You look at the last couple of decades and how many different solutions and how many different technologies and trends and ideas have come out, and we all have a litany of those in our heads of what they are. And a small percentage of those things have changed our businesses and help us grow. And so, the emotions of a group of people working and collaborating together in making decisions on a day-to-day basis, you do have to ask yourself, how much of those decisions are actually informed by data? A lot of them are, but there's a lot of them that aren't. And so, this conversation around what we're doing with data needs to be balanced out with what are we doing with people and having that atmosphere of accountability and leadership alignment and, frankly, a sense of political safety and the willingness of leadership to create environments where people are not just allowed to ask hard questions, but are expected to.

Lena Waters: Those are the things that create great questions and then drive great outcomes. All the data in the world is not going to save us. And if we lose cookies, well, everyone's going to lose cookies. So you're really just talking about leveling the playing field and then we're going to have to go back and become more innovative again. And if it's not us losing cookies, it's going to be something else that comes up. So, I think the more useless energy that we put into lamenting something that may or may not happen in the future, there's lots of things that are going to go on in the course of our business. And I think it's the attitude about how we approach those things as a whole, not necessarily how we respond to one or the other. And to your point about how do you answer questions to the board and how do you think about cookies versus somewhere in the funnel? 

Lena Waters: You have to bring a point of view about what you think is important about what you did and how that informs the thing that you're going to do. And so, Eric's point about having a strong point of view about what we're going to do with the next incremental dollar or how we're going to drive the next incremental dollar of revenue? That is the big question that we have to ask ourselves. And what's the minimum amount of insight and conviction that we need in order to do so? Because you could throw a lot of solutions and technology and data at that question, and sometimes more is not better. So, getting down to the essential piece of do we have a strong point of view about how to grow the business and what's going to get us there? That does also always have to be the question. And piling more things on, that actually becomes the reverse. That's a dangerous proposition to think about too much.

Lena Waters: There's only so many things that people and systems can handle. And so paring things down to, I think, the bare elements of how to answer a question, that's really hard for a lot of groups to do that want to bring in more and more new things. So, it's a tricky balance. And I think as leaders, we have to think about that kind of decision making environment that we're creating for our teams so that they do have the right resources, but we're not overloading them with false expectations of we have to try everything. We would drown. And so, we're looking for, I think, clarity. We're not necessarily looking for everything.

Kailey Raymond: It's interesting, too, I think that one of the things that really strikes me in this conversation is that the cultural point of view that you're building, it really has to do with making sure that everybody in the business feels like they can contribute towards the business goals. And by that, I mean, sometimes it can feel like each individual team marketing we're going after, and I'm sorry, I'm going to use this four letter word right now, but we're going after MQLs, you know, and like everybody has their own little kingdom and metrics that they're looking to drive. But really what you're fostering is if everybody is an owner of the business, everybody understands the business context, that's what's going to drive the conviction to be able to make a suggestion and hold it strongly to carry through. And so, this is a really beautiful kind of culture that I think you're kind of building at Grammarly.

Kailey Raymond: And one of the things that I wanna hear your perspective on, which is shockingly something that really we haven't talked about yet today, which is this concept of AI. And I know we're talking a lot about humans making sure that they have the conviction to make decisions. There's probably a world in which in the future, AIs are going to be starting to make some decisions as well. I guess I'm wondering both of your perspectives as to how this trend is intersecting with some of the conversations that we're having today, how you're using AI, machine learning, predictive analytics at Grammarly today. How do you think it intersects with some of the trends that we've already been talking about? 

Eric Weber: I think there's two things that are important to talk about. One is how AI and ML are part of what we do in our product and then also how they are part of how we operate the business. I think if you looked at Grammarly six years ago, we were at the forefront of just using natural language processing, which in many ways is exactly what LLMs were intended to make more efficient. And so, a lot of people may look at Grammarly and say LLMs were a huge threat to the business. In a lot of ways, they made many things that we do easier in our product and allow us to actually create a better user experience more consistently. So, if you think about using Grammarly and the suggestions that we serve to you, we think you should do this or rephrase something this way. You are seeing AI in action. I think as someone who's in the data space, I still cringe when I say AI because it's such a loaded term.

Eric Weber: It's so loaded. Every part of it, I'm not sure if I'm ever going to feel comfortable using it, but we're getting closer to the point where I'm comfortable saying we actually use AI in our product. When it comes to thinking about how do we create the right next action for a user, say like, hey, this is what we think you should do. We have the ability to create a great user experience. And I think that's the value add for AI when it comes to Grammarly more broadly. As far as AI and ML and decision making, I still am a huge skeptic when it comes to pulling people out of processes. I think ultimately, if you think about AI as agents that helps you get certain information that you need to help make broader decisions more efficiently, then I think that I can buy that. We build a lot of models internally that help provide inputs for a decision. As far as like giving up control of that decision and saying like, hey, we're just going to like automate this away.

Eric Weber: I have a really hard time seeing that happen. And again, I could be completely wrong here, but I think there is a balance between model building and using ML for product versus actually like operating the business in this way. And I wanna make sure that we're very clear about what we do in these different situations. It's hard, though. It's really hard. And I think we're figuring it out. We're figuring out how do we trust certain output that becomes input to other decisions? And how do we empower people to use this information when they may not have been used to doing that before? Like all these things are TBD. And I would just like to present that it's much messier than anyone would like you to think it is. And we should be comfortable in that mess.

Kailey Raymond: What do you think of the mess, Lena? 

Lena Waters: Well, I think we all think we're going to be saved by another acronym, which is always hilarious to me. We've moved from this trend to that trend to this acronym. It is funny, a year or two ago, I mean, we all were saying Gen AI and then we magically started saying AI. I guess we just decided to, even though most of the things are still Gen AI. So I mean, every time I have a hard day, I go back and I look at the hype cycle and go, oh, right, this is where we are on our emotional journey. And we just all have to keep in mind that this is how this goes. And anyone who's been around for a bit has seen this happen over and over. You know, the fundamental value proposition of Grammarly is that every time you interact with it, what does it do? It gives you a choice. So, it will say, sure, this spelling could be better, this grammar could be better.

Lena Waters: But what it's really doing is saying the tonality and the meaning of what you were trying to convey to somebody else to accomplish an outcome. There's something in here where you could go many different ways and it could be tone, you could be formal or informal or concise or confiding or confident, or you could be briefer or you could word it in a different way or you could take a different approach. But everything is pointing towards the choice that you as a user has to make for yourself, because how you express yourself to the world says everything about who you are and what you want. And so, what's interesting about this and why it's complicated, as Eric says, is that there are so many different inputs, the natural language processing, the machine learning and the Gen AI. But also, I mean, we have a full team of analytical linguists, actual humans who apply linguistic experience and research to understand how humans use language in the context of Grammarly's product.

Lena Waters: And all of these things have to come together to present you with a simple choice that says, how do you wanna be understood? How do you express yourself as the best version of yourself? And we're here in the background across every single surface that you work in to help you do that. But at the end of the day, it still comes down to the human choice.

Kailey Raymond: Yeah, making sure that we're never going to lose sight, that there likely needs to be some sort of intervention, eyeballs, human eyeballs on the prize at the end of the day. It's funny, in doing this show now and asking AI questions over, I think, three years, there's been like this absolute peak hype cycle. And then this real, I would say like trough of disillusionment is how I would describe what we're currently living in, because you're seeing it every single day is there's some really great things that have practical AI applications that are going to build efficiency in your team and make sure that we can work better together. But there still probably needs to be a lot of caution or at the very least, a whole lot of controls in place to make sure that it's doing exactly what you want it to do and not losing sight of those business goals in the meantime.

Kailey Raymond: My last question for you both is, if you each had any steps or recommendations, if somebody was coming to you and asking how they could uplevel their marketing and data strategies, what would you each say? 

Eric Weber: This is a good one. I would tell them not to start with a dashboard and not to start with data. And I think about the irony of that and also think about my job security and that. But I would tell them to start with trying to understand what marketers do on a day-to-day basis, like what decisions are they trying to make? What type of ambiguity and uncertainty are they dealing with? Because if we're going to enable or unlock problems with data, we have to meet people where they are. And I think this is reinforced almost on a daily basis, which is if you build something and you expect people to make a gigantic leap to suddenly be able to use the thing that you build, you're never going to get anyone to touch it. And so, I think start with understanding the domain area and the practice so that you can understand what works and what doesn't, what kind of ambiguity exists, because then you're much more likely to, one, get buy in from the people who you're trying to create solutions for, and two, you're going to spend a lot less time building stuff that doesn't make sense. Because that's a huge amount of time if you don't start with reasonably good guardrails and hypotheses about what you should be doing.

Kailey Raymond: Lena? 

Lena Waters: My recommendation is you're not allowed to be a data tourist. And by that, I mean, people use the word data as a single word answer when you ask them how they're going to figure something out, which drives me nuts. You have to have a point of view on what you want to achieve and a hypothesis about what you think you might want to be able to accomplish. And then you go and look at the data to help validate that and to gain understanding, again, about the customer. But I think if you sort of wander into a dashboard or a database and look around for things without having that point of view, you may uncover like a few nuggets here and there and you may find things that are interesting. But I think in order to create motions that are truly useful for the business, you do have to go back to first principles. You have to think about creating value at different stages of the journey and you have to have some understanding before you go in there about what it is.

Lena Waters: And then you're using data to help you understand for whom and how and when and where? Maybe not whether or not to do it at all, because you should already have that good understanding. So, I'd say people spend a lot of time touring the data to help them give it the answers. And part of the responsibility that you have to own as any kind of a professional is to understand the outcome that you wanna accomplish and have a hypothesis that you're then trying to prove. And then data becomes very helpful to you. And otherwise, you're sort of wandering around in a data forest lost and wondering why the dashboard isn't presenting you the answers to your job. And I think that's where people are going to get into trouble.

Kailey Raymond: Step out of the data forest and have some conviction. I really like this conversation, this framework you're kind of putting in place today and making sure that folks are starting with a hypothesis first and then making sure that data can do work for them. I appreciate both your insights today. Eric, Lena, thank you so much.

Lena Waters: Thanks so much, Kailey.

Eric Weber: Thank you so much.