Elif Tutuk: Hello everyone. Welcome to a data-driven podcast. I’m Elif Tutuk, Vice President of Product at AtScale. Today we have an awesome guest and an awesome topic. We will be talking about managing data like a product. And, I have Liat Ben-Zur, joining us, who’s a distinguished technology executive and corporate board member with a reputation for driving digital transformation, product innovation and strategic growth. Liat has an exceptional career, spending more than 27 years, where she has taken transformative leadership, positions at Microsoft Royal Flips and Qualcomm. during her experience as a Corporate Vice President at Microsoft, she transformed Microsoft’s consumer business, and also before that she was a Senior Vice President at Royal Flips, where she led the company through the digital transformation, launching over 30 connected iot and personal health products, and really, driving a data dream culture. to be honest, I couldn’t think anyone else, invited to this podcast, especially to talk about the topics about managing data, like a product, other than Liat. So Liat, welcome to the podcast.
Liat Liat Ben-Zur: Thanks, Elif. Happy to be here.
Elif Tutuk: Great. So Liat I’ve been, you know, hearing so much lately, from the prospect customers, the chief data officers that I’m talking, they all say that they do they want to treat data as a product. I think, many organizations inspired for their technology to go from being a cost to be, to being a differentiator. And now many organizations come to the realization that they have to do the same thing for data and data technologies. And from that, maybe we can take a step back and first, you can talk maybe about the product-led growth practices, because I know, you also coach many startups, and helping them to understand what does it mean, you know, to to to do product-led growth. And this is a very, trendy topic as well. So maybe you can talk about, what is product-led growth
Liat Ben-Zur: Yeah, sure. product-led growth really relies on making an amazing product that solves really user needs. And, and the goal, is to get users to adopt and see value in the product, and then using that engagement to drive growth. And if you contrast that with traditional sales led models, they tend to acquire customers through outbound sales and marketing. the product in that example is typically secondary to sales processes and in driving growth. So when I help companies adopt a PLG model, I, I tend to work them through a a nine step, what I call the PLG framework. And the first step in the framework is finding your ideal customer or customer profile. This is absolutely critical. You gotta identify the customers who are most likely to benefit from your product and become loyal advocates. by understanding who that ideal customer profiler.
Liat Ben-Zur: A lot of times you might hear ICP, you can tailor your messaging, you can target marketing efforts, and then you can build a product that meets very nuanced, specific urgent needs of those ideal customers. That’s step one. Step two, you gotta build an amazing product. This one seems kind of obvious, kind of silly, but you would be shocked at how many companies struggle with number two, because you need to solve a real urgent pain point or a real user need for that target customer. It can’t just be about I have a, you know, a kick ass new feature, a new technology, I wanna release it. You need to make sure it delivers meaningful value. I spend a lot, a lot of time on number two. Step number three is make it easy to get started. So that just means remove as much friction as possible for your new users to start using the product.
Liat Ben-Zur: you could focus, for example, on lightweight sign signup process, simple intuitive interfaces. Really the easier you, you, you make it to get started, the more users are gonna activate. That’s super critical. Step number four, optimize your customer journey. So you wanna guide users through the onboarding and get them to those value realization moments as fast as possible. And every product has a different value realization moment. So you gotta know what yours is. Then you work to build key retention hooks, keep the users coming back, things like gamification or notifications, habit forming designs. You wanna ensure that your user journey converts, engaged free users to paying customers through in-product guides, upgrade prompts, meaningful premium features. And you, you wanna make sure that you continue to delight your users with an amazing experience that minimizes churn and keeps them loyal. All of that is about optimizing your funnel, optimizing your customer journey.
Liat Ben-Zur: that is a very difficult thing and a lot of companies don’t spend enough time sweating those details, but is incredible the impact that that will have, which takes us to number five. Step five, if you do all those things, the next thing you can do is drive viral growth. This is really about encouraging the satisfied customers that you have who’ve gone through the whole funnel, are really engaging, maybe even converted, get them to share their positive experience, maybe even give them benefits for referring others. And that’s often, you know, might be like a discount for telling friends, about a feature or a product. it might be points or, or access to more features. And then step number six is land and experiment. Now this is really interesting. For larger companies that have more than just one product, they tend to have a portfolio of products.
Liat Ben-Zur: And if one product is doing well, there’s not enough cross sell upsell that user that that, that the businesses are doing inside of the product. So upsell, your expanded usage, additional features, new products to the current happy customers, cross-sell adjacent products to existing customers, and all of those things will increase your arpu really critical, often overlooked. And, and by the way you look, the reason that that’s often overlooked is because in most companies, especially as they get larger, all these products are done by individual teams. And those teams work in silos. So they spend months and months and months optimizing the user journey and the features within their product. And they don’t think about how to cross-sell how to upsell. So that’s another key part of product-led growth. step seven, sales and support. So just because you’re doing product led growth doesn’t mean you ignore sales and customer service.
Liat Ben-Zur: You just use them more efficiently. They step in to help customers who are already interested and are already using the product rather than trying to get interest or, or garner interest from scratch. And this kind of high touch personal support to your top customers can also be really critical for retention and customer satisfaction. So very, very critical. Step number eight is track your metrics. Keep an eye on how people are using your product. What features do they love Where are they getting stuck Listen to user feedback and usage data to continuously improve the product experience. And use analytics tools to track metrics like acquisition, conversion, retention. All those are really critical. I always say you, you know, if you can’t measure it, you won’t improve it. Measure everything. And then step nine, build, build a growth flywheel. So as your product and your momentum grows, the flywheel starts to spin faster, more users activate more, share the product, providing more data and more feedback that you could use to drive improvements. Your improvements will fuel more viral growth and so on. And you can just kind of go back to step one and you repeat the, the process. so the bottom line, nine steps, and PLG is just this different way to think about growth. It’s really about making your product so good that it sells itself, and then using really smart strategies to keep that growth going.
Elif Tutuk: I just love that. thank you so much that I think it’s so important to, you know, go through those steps and you know, you have a great framework that you have been talking, and that’s why I decided like we decided to have you in the podcast. But I think all of those steps, you know, should be told when, you know, for the audience, if you’re thinking about creating a data product, not that I’m pretty sure you are already building data products or analytics products because your organizations are using data for decision making. so maybe what we can do, just, I want to double click on some of these steps and then, you know, help the audience understand maybe going some, a little bit more deep dive, in terms of, applying those to, creating the analytics and data products.
Elif Tutuk: So I think, you know, as you have laid, said it, the second step where you know, you know, the first step, like finding the right, users and understanding what is the need of that user, what’s the problem in my career when I talk with customers, I’ve seen many cases where they may be creating a dashboard and alters products thinking that, well, if you build, they will come. that doesn’t work. So maybe, can you help, talking to your experiences in terms of how, organizations should be thinking about finding the right problem to solve with data
Liat Ben-Zur: Yeah. So when it comes to treating data as a product, it doesn’t change the framework. I would still argue, first you wanna identify who’s your customer. it may be that your customers are internal, or external users who are gonna rely on the data that you’re gonna provide to make a decision. It may be that your customers are someone who needs, to use it to build other products or run analyses. So understand who your customer is, understand kind of the, fidelity frequency of data that they need and start there. Next, we gotta define the value proposition. And in the, in the lens of, of a data product, really that means understanding what makes your data valuable. Is it the quality of the data Is it the speed at which that data is available Is it the unique insight that that data provides
Liat Ben-Zur: Or is it just a very specific data that’s very hard to get and only you have access You don’t need to update it that often, but people are willing to. So really understanding what is the value of your data, and then relatedly, prioritizing, the data. What I mean by that is you wanna determine kind of the urgent needs and map that into a minimal viable piece of data that is gonna start to show value within a week. versus what I see a lot of times with data teams, is the need this feeling like they need to boil the ocean, where customers have so much data, so much data in their whatever, in their enterprise or in their, they feel that they need to start to cleanse and organize and categorize all this data, before they pull value. That can take like six or eight months, right To get, to get all that on-prem data ready to be used, whatever it is. But maybe you don’t need to wait for that. Like what’s the highest subset value, that you can pull out And can you just focus on cleansing and, and preparing that data first, get it to the customers learn. you know, a lot of this is very product similar, but it’s, it’s just data focused. Yeah.
Elif Tutuk: I just want to double click on that, Leah, because I’ve seen so many customers where, as you said, they try to boil the ocean and it takes them, you know, six months, nine months, a year to move data from on-prem to cloud, for example. But if they can, as you said, like tying the first point of understanding the problem, like what are the business questions that are, you know, high priority that needs to be answered and then map those to the data sources and then focus on that part of the data first to make it an Alteryx radio or creation of that product. I think that is such an important thing for, you know, organizations to get value out of data right away. You know, start small, understand and prioritize. So that is great.
Liat Ben-Zur: A hundred percent.
Elif Tutuk: So, and then like as you kind of been talking about, you know, how you drive viral growth, and in my mind I tie that to data, adoption or analytics adoption. So although, you know, organizations are making so much investment in the data and the platforms to, you know, enable the users to make data-driven decision use analytics, the analytics adoption in organizations are still around 30%. And I really believe that it’s all about coming. You know, once you create that analytics product, how you create a community around that, how you make it viral, maybe you can share some of the experiences in terms of, you know, product led growth framework, in terms of making a product viral and the audience can start thinking about maybe applying those things to the, you know, end metrics and analytics.
Liat Ben-Zur: Well, I think, I think in terms of creating a little bit of advocacy or community, around data, there is, there, there are opportunities there for, for example, user advocacy. if you could identify internal champions who are advocates for the data or the data product you’re delivering within the organization, then they can help onboard new users. They can provide feedback and improvement. you know, I think that’s sort of community and advocacy is one way to think about virality of a data, of a data product. the other way to think about that is also through your sales and support team. So having a dedicated team to assist in like a much more complex data needs or data queries, they can, they can almost be the concierge of a data product and then they will help offer like specialized support that your self-help portal, may not be ready to serve for those who require it.
Elif Tutuk: Yeah, these are great, points. And I think maybe also being able to share the success, like, not only that, you know, what is that data or analytics product is, you know, what type of business questions you can answer to help them, the literacy of that. But then once you start driving the business value out of that, like sharing those successes, will help the viral adaption as well, right
Liat Ben-Zur: That is probably one of the most important points. Thanks, Elif. You know, we need to remember that data in and of itself is, is a means to an end. And you’re only collecting that data because you’re trying to solve some problem, right So if you can show that because of this data product, because of this dashboard, because of these analytics, you are able to move your retention, improve your retention, you’re able to improve your acquisition, you are able to reduce your, whatever it is you’re trying to reduce, everyone’s gonna follow suite. Everyone’s gonna start to ask you, what’s your data pipeline You know, where are you getting different, how can I apply that to my product So a hundred percent show the results and connect those results to, the data that you, you leveraged.
Elif Tutuk: Yeah, great. as we’re talking about like metrics and measuring, you know, it’s all about data-driven decisions, but I, I think as you go through this journey, whatever product you’re building, right, you have mentioned like tracking metrics, and, and then, and maybe starting with an MVP and then the other metrics measure it, come back. I think there’s great value. If you can maybe talk about what type of metrics, as part of the PLG framework you have been suggesting the products team to use and how we can maybe then talk about apply those to data products.
Liat Ben-Zur: Yeah, yeah, absolutely. Anyone who’s ever worked with me knows I kind of live and die on this hill when it comes to, measuring, and improving performance. So when it comes to metrics, there’s like multiple categories or what I, what I think of as buckets that you can start to think of. There’s user engagement metrics, I can give you some examples of that. There’s value, or business impact metrics. Those tend to like inspire people. Did I save cost Did I, you know, generate revenue Was I able to to make better business decisions Those are kind of value and impact metrics. There’s product health metrics. so that’s really a little bit more like the quality of your data, the data latency, query performance error rates, things like that. You have end user satisfaction metrics. Those tend to be probably the hardest. But things like NPS, customer satisfaction scores, customer effort, and then you have like your adoption funnel metrics, right
Liat Ben-Zur: Those are your typical KPIs, like the dashboard of, of the car, right So you kind of see every part of the funnel. user engagement metrics tend to be where I always start, because at the end of the day, user engagement metrics tell me if I have product market fit, are people using my product, are they sticking around Are they leveraging the features that I spent so much time and energy and cost building, or not So that could be user adoption rate, conversion rates, churn rates, frequency of use, you know, ow, things like that. Feature usage, retention. So I in terms of those buckets,
Elif Tutuk: I think like if we, as we, as we are talking about the user retention, like when I think about, let’s say that an organization created a new dashboard and analytics product, looking at, okay, who are the users using how frequently you are com they are coming, as you said, what, metrics and dimensions they are using, what visualizations are, are, are being used the most, or what business questions are being answered the most. I think those are also helping will help the organizations to retire what is not being used. And I think when you think about the product, like, you know, as product managers, that’s something always we do. Like if there are any features that, I mean not being used, can we retired then because it’s cost to maintain, it’s just like the same for the analytics products, right So, if they want to kind of, retire some of the metrics or dashboards that are not being used, they should definitely use metrics. Do you have any suggestions t in terms of how they can instrument the products to be able to capture different, like telemetry from the products
Liat Ben-Zur: well, I think, I think every product has to have telemetry built in from day one. I used to, I used to not let products launch or software updates launch if all of the telemetry wasn’t there because no matter how great the new features were gonna be, again, if I can’t measure it, I can’t improve it and I will have no visibility. It’ll just be a black box. So for me, that was a non-starter. I, I really encourage product managers to, to consider that as, as a requirement, there should be no new feature, no new software update, no new product launch unless it is, it is engineered to be able to track like what are the key, parts of the engagement or usage or adoption or conversion that you care about. And of course, in every product, depending on the stage of the product, there might be different things that are critical to track.
Liat Ben-Zur: And you don’t need to track everything all the time, but you need to know what matters at that stage, and you need to be, and you need to be really leveraging it. And I would also say kind of to your other point, Eli, about people maybe not using features, this will connect back up to really understanding who your ideal customer is and who your advocates are. What I have seen sometimes is there are some features that end up being kind of niche and not a lot of people use ’em, but then it turns out that those who do use ’em are your loudest advocates. and so they’re kind of like these extreme hardcore users of your product and they use, and if you upset them, it will have larger ramifications. And so it’s not always easy to just look at, you know, usage and say, oh, these seven features have the least usage, so let’s cut ’em.
Liat Ben-Zur: You really have to understand who’s using ’em and then look at cohorts. If those cohorts who happen to use the least used features are the cohorts who convert the most, your job is not to remove those features. Your job is to figure out why the hell are more users not to discovering those features and using those features. So you’ve gotta build in your product better on-ramps to those features, better discovery, better awareness. So when it comes to feature usage, you know, there’s a lot of causality and you need to be very mindful and thoughtful as a product manager on, on how to, how to interpret.
Elif Tutuk: So as you’re going through those things, la in my mind, I’m just like thinking about the analytics products. Like, one innovation that we have done within AtScale as being the semantic layer, universal semantic layer platform, like AtScale feeds all of the BI and AI consumption. So AtScale knows all of the business questions or the queries that are run in an organization. So recently we did an innovation and we have now an out of box dashboard within the product that shows who are the most active users running most number of queries. So that’s such a valuable information as you point out. Those are your data champions, those are your analytics champions, and probably like by looking at, okay, who are the other people around them that you can create a cohort, to kind of expand that data literacy, analytics, literacy.
Elif Tutuk: So that is great highlight. The other thing is, you know, with that dashboard, just being able to see what are the metrics that are being used most and what are the metrics and dimension, combinations or the business questions that are being asked most. so I think as you said, like just being able to like have the telemetry in place for every product is a must. And then, for the audience, thinking about your analytics consumption, having a mechanism where you can actually see what users are running, what queries or what business questions they’re asking is a valuable set to have.
Liat Ben-Zur: Yeah. And, and, and sorry to interrupt you, but you just triggered something for me that I, I would love to see in a, in, in data products, and maybe maybe you already deliver this, but I struggled getting this information. you know, even at, I mean at Microsoft, and that is understanding your cohorts. So if AtScale could automatically pull out, and identify and, describe certain cohorts and, and pick out based on those cohorts, these are the cohorts that retain the most, these are the cohorts that churn the most, these are the cohorts that convert the most. Now tell me a little bit about what makes each of those unique. If we could identify what are the most common traits of the cohorts that churn the most, then I have something actionable that I need to go work on to, you know, address that.
Liat Ben-Zur: If I can identify the, behaviors and common traits of the cohorts that retain the most, then my job is to go make all the other cohorts look more like my retention cohorts. That might, may mean they need to use, I need to, encourage them to use certain features. It may mean, you know, I need to make sure that they, convert at a, per a certain point, whatever it is. But every single SaaS company today is struggling with retention. And in every conversation that I have, we always talk about, do you deeply understand your cohorts Do you deeply understand what the churn folks look like versus your retention And if your products can pull that out, I think that’s extremely valuable.
Elif Tutuk: That’s, that’s a great, I think these are great points for the audience. Like literally you can be inspired from those points and then start thinking about data literacy programs in your organizations and, you know, think about creating the cohorts or having visible on the cohorts and the user profiles that you have. great inputs. Thank you. so if you kind of like, think about more around the, you know, obviously creating any product is a complex, initiative. and there are, there could be many challenges, you know, down the road. Maybe if we can share from your experiences that in terms of what are those challenges and then key lessons that you either yourself learned or you have seen, you know, other product teams are doing as they’re creating the products.
Liat Ben-Zur: I would say key, key lessons that I, that I see again and again, number one is get clear on your user and their pain points. First things first, don’t do anything until you can, you know, stand up and in any party and articulate that. don’t start with the data, don’t start with the tech, empathize with your users in order to determine the problem to be solved. I would also say involve users early and often. So the more you iterate based on continuous user feedback, rather than go back into your office or your, you know, with all these smart people and then come out the eight months later with a big reveal, don’t do that. Iterate. You’re not smarter than the rest of the world. and your big reveal will probably fall flat. So iterate often, don’t boil the ocean, right Don’t try to be everything for everybody.
Liat Ben-Zur: Instead, solve high value use cases first. That sounds simple, but you would be shocked at how rarely, product and business and engineering leaders look at a sh a set of opportunities or a set of problems and quantify their best guess for impact. You wanna be able to figure out how to quantify your best bet for impact. Impact could be revenue generation impact, could be whatever, whatever it is. But once you identify what your biggest high value, then go after that. and then maybe my last kind of learning would be focus on delivering business value, especially in today’s environment where, you know, growth at all costs is no longer the mantra. like the utility of your data product needs to tie directly to key business goals. And so work really closely to your p and l leaders, to your business leaders, as well as your product leaders. Sometimes there’s some incredible data insights that we can collect and it’s gonna be very interesting for some feature team, but it, even with all that, it’s not gonna ladder up to moving a business needle. Maybe you don’t prioritize that.
Elif Tutuk: This is, this is great. I’m also, the other thing that I’m hearing a lot from the, chief data officers lately is, is really managing the data and analytics program like a business. And that kind of comes back to every new project should tie to the, you know, a business metric and a business goal that is, you know, abroad for, for the business. so great, great inputs in those things. And just overall, yeah, like this whole conversation start to make, make me thinking about, you know, traditionally you have those power users, business analysts who are creating those, you know, using the dashboards and or getting the requirements to create an analytics and data product. And I think they should start thinking themself as a product managers and having the practices of product management, in place that we have in talking. So I just, you know, there’s a thought that just came to my mind in terms of how the personas in the BI and data space is changing, and I think maybe the, you know, data product manager is a kind of a new, persona and a title that we should be, thinking in organizations.
Elif Tutuk: Great. I
Liat Ben-Zur: Like, I like it a lot.
Elif Tutuk: so Mike, maybe just as we have been thinking about, about data and, in Alteryx adaption, I think of course, like, you know, the most common, topic lately is, AI and generative generative ai, you know, AI become more tangible, more human. but you know, we all know that the AI adoption is also low at organizations even though there’s so much, talk and then money that’s going investing investment right now in enterprises. Maybe if you can share from your experiences in terms of why the AI adoption is low and how some of those things that we have been talking should be also told about, within the AI context.
Liat Ben-Zur: Yeah, I mean, obviously like every single leader I speak to these days is feeling pressure on what’s my AI strategy. I’m, my board is asking me what’s my ai, I gotta, you know, what we’re, we’re trying to figure out how to use ai. but I think, I think the first like fundamental problem that they need to solve is lack of strategic clarity. There needs to be a really clear business goal, and alignment to core priorities for ai. Of all the things that you can solve, of all, I mean, AI can do so many cool things. What is the biggest problem keeping you up at night that has nothing to do with ai, pre generative ai What is that big Is it operational costs Is it retention Is it, you know, is it, is it a revenue, whatever it is, security. figure that out and then you can ask the question, can AI help me address that business problem or that product problem or that technical problem faster, cheaper, more efficiently
Liat Ben-Zur: so that strategic clarity is of utmost importance to really have successful AI adoption. And then once you have that strategic clarity, and I think this is why the work you guys are doing is so important at, AtScale, there is no AI strategy without a clear data strategy. And there’s so many data challenges that I think prevent companies today from executing on their beautiful AI vision. It might be that they don’t have, you know, enough data, insufficient data. It might be that their data is siloed or poor data quality. too much work is needed to aggregate, to clean, to label whatever it is.
Liat Ben-Zur: Getting that data strategy right is the foundation to ensure that your a AI strategy, makes sense and perhaps is differentiated. There’s some sort of mode. ’cause usually data is one of the primary moats. otherwise everyone’s gonna use, you know, one of the top five LLM, same LLMs. So, I think it all comes back down to data, but we need to remember that data is in the service of, in this case, your AI strategy. And your AI strategy is also not the end. It’s just a means to one end is in the service of solving a business problem. So looking at things end to end, always measuring along the way and keeping yourself honest with your North star. Did I, did I or did I not move the needle in the direction I’m supposed to with all of this work I think that’s what sets, movers and shakers in, in today’s world apart.
Elif Tutuk: That’s, that’s a great way of, summing up at the end. I think you are right on spot Elliott. In terms of, in AI strategy or, AI is, should start with data and then even if it’s AI or data, just you need to think about those like nine steps of the PLG framework. ’cause it’s all about starting with the problem, what’s going to move the needle from the on at the business side. and then understanding your users and continuously improving, collecting the data, to make iterations on that, minimum lovable product that you have built. great. Well, yeah, thank you. I don’t know if you have any, you know, last comments, but this was a, you know, a great conversations to kind of help understand, how you can apply, the product-led thinking into creation of the data products.
Liat Ben-Zur: Thank you very much for having me. I appreciate it.
Elif Tutuk: Thank you.