Uncovering Sales Growth Secrets: Data, AI, and Culture | Torben Rytt

Welcome and Introductions

Brendon Dennewill: Torben Rytt, thank you so much for being here with me today on the podcast. I know we've moved in similar circles for at least the last 13 years that I've been in Minneapolis, but we've never really had the opportunity to work together directly. So I'm really happy to have you on the show today.

Torben Rytt: Yeah, thanks for having me. And same here. I've been here 14 years, so we must have arrived around the same time.

Brendon Dennewill: Yes. We were living in Spain for five years before we moved to Minneapolis. What drove us here was ultimately love and my wife, since this is where her family was based. But it was also because the Spanish economy was really struggling after the 2008 crisis and just wasn't recovering, whereas in the US it was bouncing back pretty well. So it was a really good move for us.

Torben Rytt: Same story here. My wife was from here and she imported me. It ended up being a great business opportunity as well.

Brendon Dennewill: I was blown away by the business community here. Being in the middle of the US makes it so accessible to grow a business both nationally and internationally.

Torben Rytt: I definitely think it's a wildly underrated part of the country in terms of doing business. There are a lot of really good benefits, both in terms of access to affordable talent and connectivity across time zones. It's been very good to my career to be here in Minneapolis. I'm thankful.

From Siteimprove to Gamebox

Brendon Dennewill: The folks who know you know you from your success with Siteimprove, which was essentially your first gig after moving to the US. And then more recently, Gamebox is what you've been working on. How long has that been now?

Torben Rytt: About a year and a half, not quite two years.

Brendon Dennewill: Feel free to jump around between the two. I'm intrigued by Gamebox, especially now that I've noticed it integrates with HubSpot in addition to Salesforce. But thinking back to your previous roles, what were the key elements that separated the teams that scaled more efficiently from the ones that didn't?

The Power of Data-Driven Culture

Torben Rytt: One of the things we always did really well at Siteimprove, and that we were very intentional about, was collecting and using data. We were very fortunate that our founder had been an Oracle consultant. When he started the company back in 2003, he built the internal systems we used right up until we implemented Salesforce around 2018. The beautiful thing was that he had really thought through all the data collection. We had the same data sets going back to the very beginning of the business.

Of course, we expanded on them over time, but it all lived in the same database. It was genuinely a gift that we could go back and find data from the very first outreach attempts made in Denmark, then the UK, and so on, all in one place.

What we saw again and again was that the reps who were good at using the information they were given, and the sales leaders who were good at digesting it, were the ones who could then operationalize it. Those were the two main things: understanding how to get the data and analyze it to drive decisions, and then actually making it possible for teams to use it.

We had a real advantage in the earlier days of Siteimprove because we had a homegrown CRM. I would spend my evenings programming features for the reps, so when they came in the next day, they could immediately use the data we'd found to be impactful. That was a huge part of our agility. We could move very quickly and we did a lot of experimenting. We would try things, some worked and some didn't, but having a constant mode of experimentation was very helpful throughout our journey.

Brendon Dennewill: You're not the first person on the show to talk about the importance of data, but I love that you started there because it is such a critical component. We do know it's only one of the components, whether you look at it from a RevOps perspective or from a business operating system perspective. Data is always in there. Some businesses do a better job of first collecting the right data and then, more importantly, using it to make data-driven decisions. What do you think it was at Siteimprove that made you all so data-driven?

Torben Rytt: I think it was curiosity. It would drive me nuts, and luckily it drove the people around me nuts too. We did a really good job of getting curious people on the team. We would have these one-off successes: companies that would buy, upgrade, and do all these great things, but they didn't fit our ICP. Why were they so successful with the product? And then we'd have companies that looked just like them and they'd churn in the first year.

We had so many of those questions that we just wanted to answer. You'd look at the data at first glance and say, "There's no pattern here." And then you'd say, "Well, there has to be." Then you'd realize, "Maybe we don't have the right data." You have to go out and get new data. I think that's where a lot of companies get stuck. They stay in the box of what they already have, analyze it a hundred different ways, and can't figure it out.

What you have to do is stay curious and ask, what other data points are out there? Even if you can't find something definitive, can you find indicators that start hinting at what it might be? We would broaden our search for data, then narrow it back down, then broaden again. We kept doing that over and over until we started seeing patterns form.

Once we started understanding why certain companies were successful and others weren't, we had to figure out where to find more of the successful ones. The frustrating, and then ultimately relieving, realization was that they were spread across all kinds of different industries. I had to stop thinking about targets in terms of industries. We had to start looking at use cases. How are they using the product? Why are they using it? What do their customers have in common?

That was the biggest piece for us: really understanding how to find the next prospect that was likely to convert. And I think that's hugely underrated, which is ultimately why I founded Gamebox.

Why Gamebox Exists: Fixing the Top of the Funnel

Torben Rytt: We spend so much time optimizing our pipeline. We introduce new sales methodologies, do sales training, pipeline monitoring, deal desk, all of these things once a deal is already in the pipeline. But companies spend very, very few resources on making sure they're starting with the right accounts in the first place. Who are your BDRs or SDRs calling? Who are we trying to hit with our marketing efforts? Do we actually know what it means to be a good customer? And do we have data to back up those assumptions?

Usually, the ICP is just made on a whiteboard by a small team, put into a PowerPoint, and declared as the one to go after. I think that is very backwards.

Brendon Dennewill: So what you realized at Siteimprove was that you were learning psychographically more about your ideal customers. It was that curiosity, and the ICP customers shared that same curiosity. And now you're essentially building Gamebox with a very similar mentality.

Torben Rytt: Exactly. What I'm doing with Gamebox is building the tool I wish I had back then, when I'd spend long evenings and weekends trying to figure these things out. To do good data discovery, you need infrastructure that lets you do it. If you have to go into every system, export a CSV, build the model in Power BI, and then start all over again every time a new data point needs to be included, it's time-consuming and tedious. It means only a very small part of the business can do it, and the people you need might not have those technical skills. You might need more creative people who think differently than developers.

Brendon Dennewill: And it wouldn't be scalable either.

Torben Rytt: Not scalable at all. It's very inefficient and limiting, and it doesn't have to be. Plus, with everything happening with AI, we have new tools we've never had before. And there's certainly a lot of talk in the market about using AI to reach more people: personalized emails at scale, AI-powered SDRs, and so on.

I think that's just a race to the bottom. We're all going to start ignoring those channels because if something costs you less than a cent to do, it doesn't have any real value. At some point it all gets watered out. It's a shame that the first wave of AI tools is going in that direction.

What we're trying to say instead is: rather than blasting things out to more people, what if we narrow the scope? Who are the people that truly need our solution? The people most likely to care about what we do. Then let's have humans connect with them. Let's spend the time to say, you as a prospect are worth more than a fraction-of-a-cent email. I'm going to go out and make a human connection because I genuinely think you need what I have to show you. I think that's the future.

Brendon Dennewill: Yes. It comes back to the ICP and who you want to be a hero to. You cannot be a hero to everybody. You alluded to it: data without somewhere to manage it is unusable. And with AI being built into platforms like HubSpot and Salesforce, our clients are asking, "What do we need to do to be ready to use AI?" And when you do AI readiness assessments, the very first thing is: what does your data look like?

Infrastructure, AI Readiness, and the Role of the CRM

Torben Rytt: Absolutely. If you're not collecting it, it's the same problem we had before, just amplified. As we grow more reliant on AI and automation, the importance of data just goes up. And I think what I'm hoping will happen as the next wave is a shift away from enrichment sprawl.

When you bring a new client in, you look at their CRM and they have hundreds of fields: everything enriched over the years from ZoomInfo, Apollo, all the different providers, each giving slightly different data in slightly different formats. Instead of speeding things up, all that clutter has actually slowed people down. They have accounts in there that probably shouldn't be, some that went out of business, others that were never a real fit but were on a list imported five years ago.

One of the things Gamebox does is enable you to assign a value to every account in your CRM based on how interested you are in them as a business, not just how interested they are in you. Once you know which data points drive that value, you can start painting a picture: which accounts do we actually need, and which fields do we actually need to care about?

Then, moving forward, instead of having people create new fields every time a new data list comes in, you bring that into Gamebox instead. What I really like about that concept is it separates your workbench from your production environment. Gamebox is where you do all your data discovery. It's where you import the lists and figure out what's worth acting on. Once you've identified the accounts you actually want to work with, then you push those into your CRM.

We want to put the CRM up on a pedestal. It should be a pristine environment where salespeople can be productive. All the data discovery, which is very important but also very messy, should happen somewhere else.

Brendon Dennewill: That was actually going to be my next question: why are your current and future Gamebox customers integrating with it? What are the one, two, three driving reasons for connecting Gamebox to their CRM?

Torben Rytt: Gamebox wouldn't work without that two-way connection. We need to understand what you've already tried on all your accounts, so we pull in all the activity data, the accounts, and everything you've already enriched. We bring that into Gamebox, and then we connect it with other data sources.

The one that's been really interesting is using AI to make sense of first-party data, specifically website content. How is a company talking about itself? Who are they marketing to? What kind of products do they have? All of these questions that used to require manual research, rep by rep, can now be answered by AI.

And because we have the CRM data, we can start back-testing against your historical performance in each segment. We can say: for the accounts that had this signal, what was your win rate compared to the ones that didn't? What's really important in that kind of analysis is being careful of large averages. Something might be true in one vertical and not in another. So a big part of our implementation process is helping customers map out all the different signals and understand whether they could mean different things in different parts of their market.

The long answer to your question is: we need that two-way street to get all the information out, add real-world data to it, analyze it, and then put all that complexity into a single output. A grade, for example, that reflects how good a fit an account is for a specific use case. That grade goes back into the CRM, and that's all a rep needs to know. When you balance territories, you just use that one field. When you want to understand your pipeline account mix, that's how you do it.

Complexity stays in Gamebox. The end result, the grade or whatever we produce, goes back into the CRM. That's all anyone has to care about there.

AI Research Agents and Real-World Use Cases

Brendon Dennewill: That sounds very attractive. Torben, is Gamebox using AI to enrich the data before pushing it back into the CRM, or are you using your customers' AI?

Torben Rytt: We have our own. We call them AI research agents. As soon as we connect to a CRM, we start crawling the websites of every account in there. Even before we get to the AI part, that tells us a ton: if a domain is redirecting, maybe the company was acquired or rebranded. Just doing that crawl is a really helpful exercise. But it also means we collect all this website content, and then we can spin up agents in Gamebox that go out and look for whatever specific information you want.

The fun part is you can get really, really specific. We do most of this with OpenAI because they have some really good and cost-effective models for this kind of task.

For example, we were talking to a company recently that produces laboratory equipment for dairies. They had a product where it was easy to switch between liquid and solid production, whereas with competitors that switch was really difficult. So they wanted to find all the dairies that produced milk, ice cream, and cream cheese. If a dairy had all three, that was the golden ticket. But how do you find all of those across the entire country? That's a task that now takes us about half an hour across hundreds of thousands of dairies.

Once you can answer those kinds of questions at scale, you can target very specifically and say to a prospect: you are a really good fit for our product because you have exactly this product mix. And that's a huge win.

Brendon Dennewill: I'm already thinking of multiple examples from conversations in just the last few months where that exact scenario came up: "If only we could find the combination of those two or three types of customers." So you've built your own research agents that go and do very specific tasks at scale.

Torben Rytt: And everything we've done is meant to be done at scale. We're not trying to be another list builder, not trying to compete with tools like Clay where you make lists. We're trying to map out entire markets. We can spin up 400 agents and plow through 100,000 accounts in about half an hour. That's the kind of speed you need to ask really specific questions and back-test them in real time. If you want creative exploration, you need speed. You can't spend a month on a research project every time.

Brendon Dennewill: Absolutely. Everything is going to come down to better and faster. If you can only do better but not faster, you won't be able to compete.

Gamebox's ICP and the Mid-Market Sweet Spot

Brendon Dennewill: What is your ICP for Gamebox?

Torben Rytt: The benefits multiply as the company grows. More reps means bigger impact. But beyond just team size, we really excel when companies hit a ceiling. Most businesses start out with good success in a couple of clearly defined niches. They know who they are, they go after them. But at some point those niches start to dry out and they need to move into a broader GTM motion. That's when it gets hard, and that's where we excel: helping them figure out where the next opportunity pockets are.

To answer your question directly, under 10 reps you probably have a pretty good handle on it already. North of 25 reps, it starts to get really interesting. So that 10 to 100 rep space is usually our best fit.

Brendon Dennewill: What do companies with 150, 200, or a thousand reps typically have in place?

Torben Rytt: Most of them have complex Power BI solutions and dedicated teams whose entire job is this work. Very expensive teams. We're a young company, and I think someday we can help them quite a bit. For now, we're focused on the mid-market where we can really help people who might not have the resources for a full-blown data science practice, but who do have a RevOps team that's incredibly busy.

Every single company we've talked to recognizes the problem, but the challenge is finding the time. It's been very time-consuming to do this kind of analysis the old way. That's again why Gamebox is such a help. Not only does the platform itself help, but we also come in as consultants. We don't ask "what would you like?" and hand it off. We own the project. We sit down, establish what we're trying to achieve, meet with all the different stakeholders, bring in the different departments, and make sure there's buy-in at the end. It's a really robust process.

Brendon Dennewill: So you're essentially more of a services company that happens to have a very powerful set of tools. Your clients aren't just customers using software, they're engaging you for the work itself.

Torben Rytt: Yes. We have customers who have logged into Gamebox maybe once or twice because we do all the exploration for them. They've hired us to do that work. They use the outputs every day, but they do it inside Salesforce, working with the grades and signals we send back to them.

Brendon Dennewill: I like that. You're empowering your team of consultants to provide more value to more clients because you now have this robust tech platform with the agents doing the heavy lifting. Your team then interprets and customizes for each client how they should use that data to grow their business.

Torben Rytt: Exactly.

Rethinking the Tech Stack and Territory Planning

Brendon Dennewill: So we've talked about data and infrastructure, but we haven't talked much about the broader tech stack. I think what you're saying is that the tech stack is a critical piece of making data exceptionally useful, and fast, so that leaders can make the best decisions. And it sounds like you're focused very much on the sales side of the business.

Torben Rytt: Correct. Our customers are and will be companies worried about sales efficiency in their outbound motion. We help them be intelligent about how they target. You can use this for ABM as well.

Brendon Dennewill: It almost sounds like a replacement for ABM, but I'm glad you're saying it's really just a new and better way of doing ABM.

Torben Rytt: It's a way to do intelligent targeting for ABM. The way I see it: today, the traditional approach is to set up a target list by filtering on industry, company size, maybe some technographic criteria. The problem is that when a rep starts working through those accounts, the first thing they do is click into the company website and decide, "Do I think this is a good account?" It becomes very subjective and time-consuming.

That wouldn't be so bad if the list was great and nine out of ten accounts got a yes. But when it's two out of ten and the rep is saying no, no, no, no, we've wasted a ton of time. And when we don't know whether it's two or four out of ten, and we don't even know how those accounts are being judged, territory planning becomes really, really difficult.

By using Gamebox to do all that research upfront and at scale, we can push a smaller, higher-quality pool of companies to individual reps or to ABM campaigns. The likelihood that whatever they touch is a good fit goes way up, increasing sales efficiency. And all of a sudden it's the business that's in control, not the individual rep making subjective decisions. You learn as a team, grow as a team, and actually know the quality of what you're putting into the pipeline.

Brendon Dennewill: Which makes me think about how the ideal rep will look different in the future. They won't have to do the same amount of preliminary research just to know who to contact next. All they need to do is sit down, trust the data being presented, and focus.

Torben Rytt: I actually think what's going to happen is that you free them up to do more strategic thinking. I've never had success with salespeople who just call off a list like a robot. What I'd love to see is that reps use this to level up. Now I don't have to decide if an account is good or bad. What I have to do is make a solid game plan and understand how my product fits into their world and why they would care. I think what we're doing is freeing that time up for strategic thinking, not just eliminating thinking altogether.

Scoring Models, Avoiding Oversimplification, and Account Grading

Brendon Dennewill: Back to Gamebox's ICP: it's B2B, correct? And have you narrowed your target beyond B2B companies with 10 to 100 reps?

Torben Rytt: We're going through the exact same exercise we help all our customers with, and we don't have a lot of data yet. But where we're excelling is with companies that have hard-to-find niches. When it's hard for them to define their ICP, that's when we become that much more helpful. If you can already sell to everybody in manufacturing with more than 500 employees, we're probably not going to help you a whole lot. But if you need to find the manufacturing companies that make a specific kind of product or target a specific audience, that's where we shine.

We're working with a learning management system right now where they want to find all the companies that offer very specific types of educational offerings: certifications, continued learning, that kind of thing. You can't go buy that list anywhere. Using Gamebox, we've been able to map out the entire market for them.

We can also layer in values-based signals from company websites: is curiosity important to this company? Is there a culture of learning? These are often good indicators that someone would invest in education for their employees or customers. We've looked at hundreds of different cuts with that client, tried them out, and narrowed it down to a handful of signals that actually have strong correlation. Those are the ones we use to model.

Brendon Dennewill: For a company with 25 or 50 reps using HubSpot as their CRM, was there anything beyond what you've already shared that would drive them to bring in Gamebox?

Torben Rytt: I think everybody we've talked to has tried some version of this and failed, or at least not seen the success they hoped for. ICP is a decades-old concept. Everybody has tried different ways to define it. A lot of those projects didn't really succeed, and it usually comes down to a few logical reasons.

One, they looked too broad. They used whatever filters were available in their systems and it just didn't make any real difference. Or they went way too narrow: now you only have a couple hundred accounts that match your criteria and you still have no idea how to scale from there. What made these exercises so simplified was that it's what was technically feasible at the time. If you had to do everything in Power BI, you'd end up with a complex file that had all the answers but no practical way to use them. How do you run every new account through that model? How do you operationalize it?

And often it wasn't tested against historical data. So many ideas sound great until you look at the data and realize it made no difference, or actually showed the opposite effect. The ability to back-test is something people either haven't had or found extremely time-consuming.

And then there's the issue of oversimplification in scoring. If anybody tries to tell you that you should rank your accounts on a scale from zero to a hundred, go talk to someone else. That's not the reality. If you have an SMB team, a mid-market team, and an enterprise team, a single percentage score tells you nothing about which accounts go to which team.

One of the things we do a lot in Gamebox is what we call a size and fit matrix. What are the parameters that tell us something about the size of opportunity with this account? That's one dimension. And then separately: what are the parameters that tell us how good a fit this account is for our specific use case? You might have an account that really needs your product but it's going to be a smaller deal. You might have another that's a great fit and an enterprise deal. Being more nuanced in how you look at it is what makes these systems actually work.

Complexity lives in Gamebox. The grade, or whatever output we produce, goes back into the CRM. That's all anyone has to care about there.

Brendon Dennewill: Especially with a sales team that has SMB, mid-market, and enterprise reps. They're going to need different information, different data, different intel.

Torben Rytt: And maybe different products, different audiences. There's all kinds of complexity that can go into it.

Who Champions Gamebox Internally

Brendon Dennewill: So the folks you're working with on the customer side are strategic leaders, even though the true beneficiaries are the reps whose daily work becomes more interesting. Who are the people you're working with most closely on the customer side?

Torben Rytt: Our champions are usually RevOps. They're the ones who've been tasked with finding a solution. The CRO has said, "We need to improve our sales efficiency," and RevOps owns the project. But once we're up and running, it's actually product marketing that often has really valuable input. They know who the product was built for, how they think about the ICP for different parts of the product, and what signals might be interesting.

What's really cool is that product marketers typically don't have access to Power BI or the data warehouse. Suddenly they have a way to come and say, "Hey, could we try looking at this? Could we run through these ideas?" That's how we love to work. We want sessions where people bring ideas, we try them in real time during the meeting, and then decide if something should move into a scoring module.

Brendon Dennewill: Very cool. Torben, I know we've blown through the time here. Because you're so humble, you didn't talk much about what comes before. You started with data and technology, but I'm guessing the precursor to all of this is strong leadership, right?

Leadership, Ownership, and Curiosity

Torben Rytt: Yeah, that's true.

Brendon Dennewill: I just wanted to make sure we mentioned it. Maybe leadership is table stakes if you want to scale better than your competitors.

Torben Rytt: It is. One of the most important things about leadership is taking ownership. If you want to succeed, you have to take ownership of every part of your business, including this. Ownership and curiosity together are a powerful combination. When you have both, you start questioning things, see a problem, and go figure out the solution. The leaders I've worked with who have done really well all do that. They take ownership, they're curious, and they keep looking for better ways to grow.

Advice for RevOps Leaders and Closing Thoughts

Brendon Dennewill: To wrap up, Torben, what advice would you give to a RevOps leader today, knowing everything you're working on and what's coming from a data and technology perspective?

Torben Rytt: We're in a really exciting time. There are a lot of products coming out, Gamebox being one of them. I think we have to be open-minded as GTM leaders and be willing to explore. But I also think it's going to be really important that we understand our brand and the values of the company, and use that to guide how we use all this new technology.

Are we okay with using it to reach more people, potentially creating more noise in everybody's inboxes and on their phones? That's something most brands doing outbound already struggle with. How much is too much? Now all of a sudden everyone has access to enormous firepower, and every brand has to figure out how to harness it responsibly. There's not going to be a universal right or wrong answer. It's up to each individual brand to figure out what makes the most sense for them and the way they want to reach their audience.

Brendon Dennewill: Which ultimately comes back to leadership. Leadership is about making decisions, and what you're saying is: make the decisions that are right for the future of your business, which also means the future of your customers.

Torben, thanks again so much. I definitely look forward to following Gamebox's success much more closely now that I know more about it.

Torben Rytt: I appreciate it. Thanks for having me.