In today’s episode of “Next Gen Builders,” Francois is joined by Gage Sonntag, Senior Director of Data at Super.com. Their conversation hits upon the intricacies of working in a data-driven organization, how roles are evolving in the industry, and the power of experimentation.
In today’s episode of “Next Gen Builders,” Francois is joined by Gage Sonntag, Senior Director of Data at Super.com. Their conversation hits upon the intricacies of working in a data-driven organization, how roles are evolving in the industry, and the power of experimentation.
Throughout the conversation Gage and Francois discuss creating a balance between data self-service and governance. Super.com has an innovative approach to embedding data analysts within mission-aligned teams to promote agility and deeper expertise in solving specific business problems while maintaining a centralized data engineering team to ensure robust data infrastructure. You’ll also hear about the inherent challenges and excitements in driving data quality and experimentation. Gage stresses the need for a culture that prioritizes learning and truth-seeking over just achieving expected outcomes, advocating that empowering individuals with the right tools and mindset can spur significant organizational growth.
For those looking to advance their careers in product management or data, Gage's insights offer actionable advice. He highlights the importance of creating a culture of experimentation and continuous improvement, emphasizing the need to iterate and learn from every opportunity, even failures. Overall, the interview encapsulates how having a robust data culture and a flexible, yet structured approach to data management can transform organizations and drive remarkable success.
—
Guest Bio
Gage Sonntag is the Sr. Director of Data at Super.com, where he leads initiatives across analytics, business intelligence, data science, and data engineering. With a robust background in data practices, Gage has significantly enhanced the company's data ecosystem, achieving remarkable results such as accelerating testing by 4X and improving customer retention by 90%. His leadership style emphasizes rapid experimentation and democratizing data access, empowering employees to engage with data tools and contribute to innovative solutions.
Before joining Super.com, Gage held key positions at Bond Brand Loyalty and LoyaltyOne, where he honed his expertise in analytics and engineering. He holds a Master’s Degree in Management Analytics from Queen's University and degrees in Chemical Engineering and Chemistry from the University of Saskatchewan. Gage's diverse career path also includes a brief stint as a Brewer, reflecting his multifaceted interests and commitment to blending creativity with data-driven decision-making.
—
Guest Quote
“What didn't work for us was to be a part of every single experiment you run. The more that someone can come up with an idea and then, as frictionless as possible, go from that idea to that in-market and getting data, was really powerful for us. We found it's not always the people at the top of the org chart that have the best ideas, but it's really about making it easy for anyone to take an idea from inception to, it's live.
A good colleague of mine kind of came to me one day and was like, ‘Gage, running experiments is really hard. There's so many buttons to configure. I'm not super sure if I'm doing it right.’
And after that conversation, I made it my mission to be like, ‘Okay, it being hard to run an experiment is not an acceptable barrier to testing something.’ Cause the alternative is you just roll out the things you think are best and likely not what your users think are best.’” – Gage Sonntag
—
Time Stamps
00:00 Episode Start
01:39 Gage's Background
03:12 Oil, beer, data
05:58 What is Super.com?
09:10 Balancing self-service with data governance
11:23 Experimentation at Super.com
14:54 Diving into data quality
16:44 Perfection vs. Speed
20:04 Gage's Oh Shit Moment
24:30 The people side of data
27:36 How to make data teams thrive in a smaller organization
31:02 When small tweaks have large impacts
—
Links
0:00:02.1 Gage Sonntag: I think the people-sided data is, like, the hardest part of it, and it's also the funnest part. Like, inherently, it really comes down to, like, organizational behavior and incentives you create. So, for us, we put a lot of time into celebrating truth-seeking and celebrating learning, rather than celebrating the outcome that I expected we'd get. We really tried to focus on if something didn't work how you expected, what is it about the user behavior that's different from your mental model, and how do we take that and iterate upon it?
0:00:39.5 Francois Ajenstat: This is Next Gen Builders, the show for the growth and product leaders of tomorrow. If you know me, you know I love data. Data is in my blood. And some people say that data is a new oil. Well, like oil, it's got to be refined. It has to be controlled, stored, and protected. And you better make sure there's no data spill. So, how do you do all that and still let people build with it and use it to drive impact? Well, joining us today to talk through building and running a data-driven organization is Gage Sontag, Senior Director of Data at super.com. Welcome, Gage.
0:01:20.6 Gage Sonntag: Thanks for having me, Francois. Excited to be here today and chat data.
0:01:24.4 Francois Ajenstat: Yeah, I love data. So, we talked to a good amount of people here on this podcast that don't have a traditional tech background. And I think you're one of them as well. Tell us a little bit about your background.
0:01:38.2 Gage Sonntag: Yeah, I probably have one of the most eclectic backgrounds of people I bump into. But my education is originally in chemistry and chemical engineering. I've had a lot of different jobs. I worked in scientific research a lot of the time I spent in university. I drilled for oil for a living. I made beer at a brewery. Eventually, most paths lead to data. But that really pushed me to go back to school and get my master's degree at Queen's University here in Toronto. But really, my undergrad was like, if you've read the book, Thinking and Systems by Donella Meadows, that is my degree. And so, it was actually really interesting reading that book years later and being like, oh, wow, everybody doesn't think this way. But really, my senior thesis was like studying a uranium mill in northern Saskatchewan and figure out how to optimize and de-bottleneck it to kind of have more throughput go through it. It turns out that's a very useful skill. And it's really the same skill for like, how does a data team generate more impactful insights faster?
0:02:41.7 Gage Sonntag: How do I build data pipelines faster? How do I run experiments more reliably, better, faster? Inevitably, these are all people, process, technology challenges. So, it's definitely a winding path on how I got to where I am today, but all of it's kind of useful at different points in time.
0:03:01.2 Francois Ajenstat: So, how do you end up drilling for oil or making beer and then becoming the data guy? How does that transition happen?
0:03:11.8 Gage Sonntag: It happens in a very confusing way. In maybe the early 2010s, I was working in Calgary. Oil was probably $90 or $100 a barrel. In that industry, every day, you're keeping a pretty close tabs on commodity prices. I spent a couple of years working in operations, so working relatively close to Alaska, I guess, for maybe the American listeners on your podcast. And so, after that, I spent about a year in a team that was focused on modeling and pipeline development.
0:03:45.7 Gage Sonntag: And it was incredibly data-intensive, incredibly mathematical, like a lot of modeling physics. The technological part of it was very interesting, very cutting-edge, very novel work. The business impact was a couple steps removed. Eventually, unfortunately, I lost my job as oil prices dropped 50%, 60%, 70% as happened to a lot of my colleagues. I took a winter off. I was a ski bum just outside Banff, if you're familiar, but really just enjoyed a bit of a winter. I thought that making beer was a hobby that I could turn into a career. It turns out that all the things I love in a hobby are not at all what it is as a career. And then it was time to go back to the drawing board and go back to school. And I was really looking for a place where I could be a bit closer to the business but also embrace a lot of the technical, mathematical things I enjoyed in my earlier schooling. And then, eventually, data fell into my lap.
0:04:43.8 Francois Ajenstat: That's great. So, really, part of the moral in all this is there's no direct path to becoming a data leader. And all your skills just really can come together and do really powerful things.
0:04:56.0 Gage Sonntag: Yeah, I think so. It's only maybe recently in the last couple of years that data as a subject as an undergrad or graduate degree has gotten a lot more popular. But a lot of it was the early pipelines were in from a lot of STEM careers. I found it's been really helpful as I picked up a lot of disparate attitudes, skills, beliefs. And I think it creates for really resilient teams when we kind of look for that. I think being a great leader in data, you have to speak a little bit of performance marketing and growth marketing, a little bit of FP&A, a little bit of ops. So, the role inherently is usually at the intersection of a lot of different functions within the company. So, the more experience you can kind of bring to bear, the more diversity you have, probably the better you'll do.
0:05:54.1 Francois Ajenstat: Yeah, data is at the heart of everything. Now, you're running the data team at super.com. For those of you that don't know about super.com, can you tell us a little bit about the company and what do you guys do?
0:05:55.3 Gage Sonntag: Yeah, for sure. Really, the vision for super.com is to provide access for everyone to experience more of what life has to offer, regardless of income or circumstance. We provide an all-in-one app that puts more money in your pocket. So, we enable users to save big on hotels, access cash advances, get cash back on purchases they might make with our MasterCard, boost their credit score, earning money playing games, and more. So, we're building quite an advanced ecosystem at this point that's really just centered around helping our users save and earn money.
0:06:32.6 Francois Ajenstat: That's super, hence the name of the company, super.com. But now, you run data at super.com. What does that really entail? Are you the central data team? Are you part of a business unit?
0:06:45.8 Gage Sonntag: Yeah. Really, the debate of centralization and decentralization is one of the larger ones when you talk to people who work in data. We're trying to get the best of both worlds. We operate in what we call mission-aligned teams. This is a lot like Amazon's single-threaded ownership model. So, we embed analytics into all of these different mats, we call them, and we found that really works best for our data analysts. So, it allows them to spend every day sitting next to a PM or a growth manager or some engineers and really build deep expertise in the business problem they're solving. So, our mats might be focused on only acquiring users to our fintech product or only working on retention for users for our super app. So, really, it helps our team just sit and deeply learn what business problems our users have. And then, really, there is an aspect of centralized reporting and coaching. So, your manager is also someone who works in data, and we're trying to get the best of both worlds of both the centralization and decentralization.
0:07:46.7 Gage Sonntag: But really, our data team is on a mission to try and do, really, solving the hardest problems here at Super with data, really working on a lot of strategic problems. And then, also, there's a component of enabling our users to do self-serve reporting, to run experiments themselves, really, to do as much as they can without involving the data team.
0:08:10.1 Francois Ajenstat: So, is the, how do you call it, the mats, the mission?
0:08:14.8 Gage Sonntag: Mission-aligned teams.
0:08:17.0 Francois Ajenstat: Mission-aligned teams. Are they embedded in the business units? And then, is there a different data service organization that provides the data infra?
0:08:27.1 Gage Sonntag: Yeah, that's a great question. We found that, like, within analytics, they function best when they're sitting embedded with PMs, embedded into all of those business functions or product functions. And then, a lot of, they're often supported by a centralized data engineering team. So, we have a data engineering mat that serves up a lot of the typical data pipelines, observability, reliability, and kind of keeping a lot of the data ship running.
0:08:58.8 Francois Ajenstat: Got it. So, one of the big debates that we always have on the data world is, I'll call it, self-service versus governance or control versus empowerment. How do you guys approach that internal debate?
0:09:07.5 Gage Sonntag: One aspect of my career I enjoy is I've had a chance to work with a lot of really big companies and a lot of really small companies. I'm probably on the smaller side right now versus some of the time I've spent working with, like, really big banks and other places. I think it really comes down to kind of a bit of company philosophy. At a certain size, I think maybe accuracy matters more than moving fast. And then, I think that's where you see a lot of, like, kind of, like, really hard investments in governance and centralization because the issues with being inaccurate outweigh the ability to kind of help people move faster. And what we've, we're certainly not that today. We've skewed, the pendulum is very far on the self-serve spectrum. And really just providing people the data themselves to do analysis and come to their own conclusions and then action those conclusions has been tremendously powerful for us.
0:10:04.1 Gage Sonntag: My thesis is really trying to remove the number of people in the loop. And I've seen organizations where it's like you have an idea, you pass it to some data person who does a bunch of analysis and eventually kind of comes back to you with it. That's a great model for some companies. What works for us is really just trying to give you the data up front and the tooling and the skills and the knowledge to do all that work yourself. And there's fewer handoffs between different people. And we found that allows us to move a lot faster. It might increase the chance that you make mistakes at times. And we really just coach our team on asking for help and making a lot of the ecosystem as frictionless as possible. And we're there as a phone a friend when you need it or when you're like, hey, these numbers don't really seem right. Or usually they seem too good to be true. But otherwise, I think our end user is just being empowered works the best for kind of our scale and our size.
0:11:02.1 Francois Ajenstat: I mean, I agree with that. And in many cases, having the debates about the data and being able to have a conversation about how you're thinking about it actually helps a ton. But can we make it a little bit maybe more practical? So I know you do a lot of experiments and experiments is a way of learning. It's a way of moving things forward quickly. Like, how does that work at Super.com?
0:11:22.7 Gage Sonntag: Yeah. Experimentation is a space we've invested a lot in over the last year. Every day in Q1, we launched a new experiment, which even when we got to the end of Q1 was really impressive. When we started out on our new experimentation platform, we invested a lot in kind of figuring out the best practices as a data team and trying to distribute those. So how you're designing the experiment, like standardizing around key metrics to track a lot of that kind of work. We try and measure centrally so that the best practices exist. But really what didn't work for us or what we didn't wanna do was, like, be part of every single experiment you run. And so the more that someone can come up with an idea and then as frictionless as possible go from that idea to that's in market, getting data, was really powerful for us. And really we found it's not always, like, the people at the top of the org chart that have the best ideas, but it's really about like getting it, making it easy for anyone to take an idea from inception to, like, it's live.
0:12:28.1 Gage Sonntag: A good colleague of mine kind of came to me one day and was like you know Gage, running experiments is really hard. There's so many, like, buttons to configure. I'm not super sure if I'm doing it right. And after that conversation, I made it my mission to be like, okay, like, it being hard to run an experiment is not an acceptable barrier to like, testing something, seeing whether it works. 'cause the alternative is naturally you just roll out the things you think are best and they're may or likely not what your users think are best.
0:12:58.0 Francois Ajenstat: And so is it your team that runs the experiments or is it the business team that runs the experiments supported by your team or something completely different?
0:13:09.8 Gage Sonntag: Yeah, great question. It totally changes from kind of experiment to experiment, which is probably the worst answer you wanna hear. Our hope is that, like, most experiments you don't need us. The best practices are out there. Our platform has a lot of guardrails. The technology space has improved a lot for a lot of the experimentation platforms that are out there. I think there's still always a helpful space. The data team has a role to play here at super.com. The first is, like, anything that seems to touch an offline experiment is always, like, a bit more hard. So when you're interacting with a user like, through the call center or other different ways, they're just harder experiments to run. And when there's much more complicated optimization functions. So when we're trying to look at, like, tradeoffs of, like, conversion versus revenue per booking, when we're looking at doing different kinds of pricing experiments. It's not as simple as, like, experiments where you're like, hey, if more users convert, that's great. Anytime there's tradeoffs that's generally a space where we tend to be a bit more involved is, like, the understanding what success looks like isn't quite as straightforward.
0:14:22.3 Francois Ajenstat: Do you guys have that same approach for other things like reporting or data quality? 'cause one of the things I've seen in the past is when you can empower the business to answer their own question, sometimes they don't get the right answer. Or they get an answer that's not exactly accurate, I should say. Who owns the data quality side of things? Or is that an iterative process that you do as you're engaging with the business to truly understand what they're looking to do?
0:14:49.1 Gage Sonntag: Yeah, I think data quality is, like, a journey everybody's on. And I think everybody would say we kind of suck at it, no matter who you are. I think one of the favorite parts I have about working at super.com is, like, some of our business lines are very, very mature. Like, we started as an online travel agency seven, eight years ago. And then other parts of our business lines are less than 12 months old. So I think, like, that's kind of how I anchor my perception of data quality is, like, if we've been doing this for eight years, we should probably be in a pretty good place. Reporting is very mature. Our understanding of our KPIs is very mature. And our ability to, like, get an accurate answer is very good. I think for business lines that we're still testing that, like the product market model channel fit might not quite be there yet. Our confidence in all of the data is perfect, accurate, isn't really there. But until you know you kind of have some scale, it's also probably not the thing that's gonna 10x your business if you invest in it.
0:15:54.8 Francois Ajenstat: It's interesting. I mean, it sounds like you guys have both a data culture and a learning culture.
0:15:57.8 Gage Sonntag: Yeah. I am very fortunate that both leaders before myself as well as our co-founders have really embraced this. Our data is one of our core values. I think that's the same for many organizations. But I'm very fortunate kind of how it comes to life here is one where we are intensely looking to measure everything we do.
0:16:25.2 Francois Ajenstat: Amazing. Now, when everything is data, there's always this question about risk and how do you balance off that risk versus empowering others. Have you seen something that works well, better to just mitigate risk, manage risk, control risk?
0:16:45.1 Gage Sonntag: Yeah. When we started out in experimentation or trying to scale this, I spent a lot of time kind of reading what other people do. And it was really insightful kind of hearing from some leaders at like Microsoft, Airbnb, companies like that and hearing that, A probably 10% of experiments you launch probably ship with bugs in them. And then B 80% of your experiments probably don't work anyway. The users don't like them. So.
0:17:09.3 Francois Ajenstat: Like so its 10% have bugs and 80% fail? Is that?
0:17:13.0 Gage Sonntag: I think the 10% is probably in the 80%. Like the experiment fails whether it's a bad idea or a bug. So you're kind of left with like most of your shots at bat aren't gonna work. They're not gonna get the results that you want. And I think that was also a really helpful learning for us internally because I think when we got started, the expectation would be everything's really easy. You buy a new platform. We were migrating off of like a homegrown solution. And we saw it was actually a lot more obvious to see where we kind of tripped and stumbled. But realizing that even companies at scale have a lot of issues that go out the door as well. 10% is a high number when you think we shipped 90 last quarter. It kind of helped, like, keep our own expectations very reasonable.
0:18:04.7 Gage Sonntag: And I think our culture is one where we're trying to move very fast. And then it kind of means that you have to have a reasonable tolerance for things being a little bit off things breaking. That's kind of the cost in some ways of moving quickly. And as long as a company, you're both unified in the belief that that's better than moving slowly, I think you're kind of in an okay place. But I think every company has to kind of figure out their own culture and their own kind of risk tolerance for where they sit on that spectrum of, like, perfection versus speed. And I think in a lot of things with data, there probably is a little bit of a tradeoff to make.
0:18:46.0 Francois Ajenstat: Right. And not everything requires the same level of trust, security oversight, numbers that you might report from a financial standpoint, probably definitely do but maybe the number of people that visit a page has a less gravity to getting it not perfect.
0:19:07.9 Gage Sonntag: Exactly, yeah. I think a lot of product analytics inherently is like, you might have issues with ad blockers, you might have issues with users across different devices. It's not a perfect science like at the best at times, and so you just have to kinda condition all of that through, again, either the business function you're working with or the maturity of that kind of business problem you're facing.
0:19:32.0 Francois Ajenstat: So, you do a lot of experiments and of course, things fail. One of my favorite questions to ask people is to tell me about one of your oh shit moments where something happened could have been extremely positive, extremely bad but it happens and you're like, oh shit, what happened? How did you guys handle it? What did you learn from it?
0:19:53.0 Gage Sonntag: Yeah. I guess there's two halves of that, one is like, how do you think about failure and then the second is like the moments your heart skips a beat. A good mentor of mine kinda frames success as like the ratio of learning versus effort or learning divided by effort. So really, if you're not learning, I think that's kinda how you kinda frame failure. And so, what that came to life for us sometimes is like you have an experiment, you try and launch it, the results don't make any sense, there's a bug somewhere, you try and fix the bug, you re-launch it, it doesn't really work again. You're like, there's still something wrong. I think that that's kind of how I think about the lack of learning is really like how I kinda frame failure in some of these contexts. I think my favorite oh shit moment is we got our experimentation platform, we connected it to everything, we tried to have one place that all of our users would go, whether they're testing something in CRM or within credit risk or just within other parts of our products. And so, we were testing some changes to our cash advance underwriting like how we give our users money, and our team had shipped an experiment and it was their first couple of experiments, so I kind of expected it to be... Maybe have some speed bumps.
0:21:10.8 Gage Sonntag: And so over the weekend, I kinda checked in on how the experiment was going and saw we had given out much more money than I thought we had or I thought we would, which gave me a little bit of a heart attack. And so, I ride back to work Monday morning to learn from the team that the telemetry we had in our experiments platform wasn't like how many dollars we gave out but how many cents. And so, what I saw over the weekend was like 100 times the number I thought it would be and then that was definitely my moment of like, "Oh wow! Shit. We're doing real work here." It led me down a really good process of learning from our team what kind of guardrails we had in place. Unfortunately, I think we would have been notified before we completely lost our shirt, but it definitely gave me a bit of a newfound respect for like some of the impacts experiments can have when maybe they don't go as planned and I got that lesson without needing to experience it myself.
0:22:08.5 Francois Ajenstat: I felt I got an oh shit moment as well when you described it.
[laughter]
0:22:14.0 Francois Ajenstat: You mentioned guardrails, how do you think through the guardrails to, again, protect when experiments go wrong or when data goes wrong?
0:22:24.9 Gage Sonntag: Yeah, I think you kinda build guardrails for different things, and so you're either... Like I kind of frame it as like they're either... You're either trying to control for people problems, process problems or technology problems. And so, as it relates to people problems, we've invested a lot in training, whether that's leveraging things from different kinds of third parties and also eventually just building up our own best practices within our company and then saying, "Hey, before you run any experiments, here's all the course work you have to do, here's all the stuff you have to do to get up to speed and get comfortable." That's been fairly successful for us, and then also we've tackled a lot of process problems. For a small period of time, I have the data team review every single experiment as even with a bunch of documentation, there were still things we missed or like just like... It's like a peer review for engineering code. It was just like a helpful conversation that took place. And then within experiments, we try and think about what kind of like what downward effects we might have like increasing user conversion rates might increase cancellation rates. And so, just trying to think through as we build experiments like what are the other things to keep an eye on as well, what are the second order or third order effects that you might create as a result of some of the positive changes you have.
0:23:46.1 Francois Ajenstat: Interesting. There's a lot in what you're describing that makes me think that at the end of the day, there's data problems and then there's people problems where... And I say that lovingly, just to be clear. But, when you think about the people problems, it's people misinterpreting data, people getting their own view of how they want the results to look, people not, say, understanding what the data is, how do you think about the people side of data?
0:24:17.5 Gage Sonntag: Yeah, I think that the people side of data is the hardest part of it and it's also the funnest part. Inherently, it really comes down to organizational behavior and incentives you create. If you take people and you're like, "Hey, your job is to drive this metric," you'll see people intentionally or unintentionally be like, "Hey, this experiment, maybe it didn't drive this thing I wanted but it drove this other thing that I think is really good." And so, I think a lot of them are just what you'd expect kind of like as you create the system around it. So, for us, we put a lot of time into celebrating truth-seeking and celebrating learning rather than celebrating the outcome that I expected we get. We eventually built an internal sportsbook that allowed users to bet on different experiment outcomes. And rather than celebrating who has the most points at the end of it, we really tried to focus on if something didn't work how you expected, what is it about the user behavior that's different from your mental model? And then, how do we take that and kind of iterate upon that? So, by focusing really on truth-seeking, I think we got a lot of better outcomes or try to help control but really that human behavior of my bonus or my salary increase, my performance is coupled to how some metrics do, like there's a natural perversion of incentives that you always risk creating.
0:25:52.0 Francois Ajenstat: Interesting. I don't think I've ever heard the idea of a sportsbook for internal data teams. Is that like gamification, essentially?
0:26:00.8 Gage Sonntag: Yeah, yeah. We gamified the experiment process. It also does a lot of really helpful things like nag people to close experiments that have been open too long, but it's also been a great way to just celebrate the velocity, to recognize kinda some of the different changes we make. And then also, just engage people more especially across teams with things that are happening within our company.
0:26:21.8 Francois Ajenstat: That's fantastic. And then as part of that, do you rank people according to their success or do you rank the experiments based on their impact?
0:26:33.3 Gage Sonntag: There is a leaderboard, I think we're very careful not to put a lot of validity into the leaderboard as it risks kinda like encouraging the wrong thing, but where I think it's been helpful has been the cases where experiments launch and there's a lot of conviction around a certain idea and something else kinda the dark horse kinda comes out of nowhere as the winner and it's a great time for our team to really recognize a bit of group think that we might have or bias that we might be introducing to how we think about our users.
0:27:08.6 Francois Ajenstat: Fantastic. I might have to steal that idea on our next project. I love your perspective on big company versus small company. When you think about data problems, you would assume the bigger the company, the bigger the data problems but how do you think about setting up data teams, data cultures regardless of maybe the size of the company or the situation?
0:27:34.2 Gage Sonntag: Yeah, that's a great question. I think probably culture plays the biggest tone in it. I think in smaller companies, it's easier to kinda evangelize what the right kind of data practices look like. And it's often set from kinda the top down from the CEO on downwards how you think about data, how you talk about data, how you use it to measure your business results, to drive your business results. It kinda starts from there. I think a lot of people maybe pay lip service to the idea, but I think really everybody is just trying to kinda move a little bit further up that spectrum. Really, it requires a lot of pragmatism. The part I enjoy is like, it's a super messy space like understanding what to work on, what to focus on as a team is not very clear. It really starts with being close to the business, you always have a lot more projects than you do kinda time in the day and you have to get used to saying no to a lot of stuff to being really focused and really trying to create outsized impact. And then from that, it creates a flywheel of leaning more into data, finding more opportunities.
0:28:44.9 Gage Sonntag: I think the differences for a smaller company, we kinda have to focus a lot on 10X opportunities because for a lot of parts of our business, like a 3% lift in some metric or conversion rate changing by 1% might not be super interesting on a business vertical that we're still kind of figuring out the fit of.
0:29:07.8 Francois Ajenstat: Interesting. You guys are lucky that you've got a top-down support for data culture. What advice would you give to folks where maybe that's not the case?
0:29:19.0 Gage Sonntag: That's a good question. I think really to get good buy-in, you kinda have to show the opportunity and the value of data. I think that starts inevitably with understanding your business. At the top of the call, I touched on being a little bit of an FP&A analyst, a little bit of a growth marketer or like being able to see across that and see the bigger picture allows you to both kinda A, filter out the not super great opportunities to invest into with data but as you can find opportunities that you can tie the data team's output directly to, "Hey, we made some more money here, we became more efficient on a unit economic basis," that creates a flywheel of being like, "Okay, where else can we apply data?" But it really requires a brutal prioritization exercise to be like, where is the highest opportunities, how can we really focus on those and not get stuck becoming a team that's like building dashboards all day, every day or like other... That people may or may not use, but really focused on directly attributable revenue.
0:30:32.8 Francois Ajenstat: Don't build dashboards, build impact.
0:30:35.0 Gage Sonntag: Yeah, exactly.
0:30:35.5 Francois Ajenstat: The dashboard may be a means to get there.
0:30:37.0 Gage Sonntag: Sometimes, the dashboards are a means to get there but...
0:30:40.9 Francois Ajenstat: Exactly.
0:30:41.1 Gage Sonntag: May or may not be the case.
0:30:43.8 Francois Ajenstat: Thinking about that, do you have an aha moment or an unexpected learning that you saw this in action where you drove a process change and then all of a sudden the light bulb comes out, everybody else starts aligning?
0:31:00.2 Gage Sonntag: Yeah. Around the time I had gotten hired here at Super, we were in the middle of scaling the company as a whole and one of the areas I've always enjoyed throughout my career is working in pricing problems. And it's always been a really interesting space that's both like a cool optimization plus also really hits my love of behavioral economics. And so, as I got started, I started learning more about how we do pricing here at Super and felt like there was a larger opportunity we could lean into. At the time, we were kind of just pricing all of our hotels the same of kind of the costs that we had plus some markup, and we'd spent a lot of time in this space with previous iterations of the data team and couldn't quite kinda like crack the nut. We're able to find a couple early wins of kinda cutting down certain segments for users who are arriving a couple of days from now and saw a huge lift in conversion.
0:32:00.0 Gage Sonntag: And that was kind of big aha moment for our team as we realized the unit economics of the business might become very different but as we thought about pricing differently, we could actually drive really strong top line results and really kicked off a really larger investment we made in pricing and complexity around pricing, and really was just driven by a few really key wins early on but also just like a belief that there is a smarter way to this as we get bigger as a company. And especially for us, these might be small incremental gains but they drove really meaningful impacts to our financials and that was really because we'd gotten to the point in growing that business where smaller optimizations mattered. In years before, it just didn't drive the same kind of outcome or impact that other things could. So, both kind of finding the right pockets of opportunity where the data team was best suited to play, really thinking about trading off conversion versus the revenue we might get for each conversion allowed us to find a really nice sweet spot and drive a lot of results.
0:33:12.2 Francois Ajenstat: I think it's great. And as somebody, myself, who really believes in the power of pricing, I actually love pricing, I love kind of both the behavioral side, psychological side and mathematical side, I think there's just a lot of opportunities of leaning in, learning, testing and figuring out what actually makes sense for the business you have and for the customers you're trying to reach.
0:33:36.9 Gage Sonntag: Yeah. And ultimately, as much as you can kinda try and centrally design it all, until you put it in front of users, you're not really gonna know what works out well or works out not. So, it's a great use case for experimentation but also just like a business domain that's very easy to tie to tangible outcomes.
0:33:58.1 Francois Ajenstat: I love it. Well, Gage, as a fellow data person, thank you so much for joining the podcast. I'll definitely be thinking about how to apply the idea of the mission-aligned teams to really build better connections and again, the idea of having a learning mindset, building a data culture is just so fundamental to transforming companies and driving better outcomes.
0:34:23.7 Gage Sonntag: Cool, thanks for having me today, Francois.
0:34:25.8 Francois Ajenstat: Well, Gage, thank you for joining us and thank you all for listening to Next Gen Builders. Look out for our next episode wherever you get your podcast and please don't forget to subscribe.