“Building Data-Driven Revenue Operations”—with Peter Eigenschink (WU Vienna)

Recorded in German—English translation of the original conversation.

Intro

Maria-Lena Matysik: Welcome to Analytics Sprechstunde!

Today I’m interviewing a guest with a somewhat different perspective on marketing. Peter Eigenschink doesn’t come from “classic” online marketing; instead, he researches quantitative marketing at WU Vienna University of Economics and Business and earned his PhD in marketing analytics.

Alongside his research, though, he also works with companies to bring his knowledge into practice, into the private sector, to apply it and, of course, to challenge it to some extent. So he always looks at data and data-driven decisions at this very exciting intersection between research and practice.

His perspective is half academic: What value can data create in an optimal world—under optimal conditions, without friction, with clean testing and experimentation conditions? Through his practical experience, however, he also knows that you can quickly hit a wall with pure theory and that, in the end, a workable solution has to be found to turn the added value from the data into business value.

Peter has worked on a range of topics, such as pricing questions or forecasting and inventory optimization.

For this episode, however, we set out to do something different. It’s about the question of how we can use data in marketing when customer journeys are long and complex and are managed and optimized across teams. At the center is the topic of data-driven Revenue Operations: How does Revenue Operations actually work? What does it involve?

A particular focus is on a question that keeps coming up in the conversation: Why is existing customer data so often underestimated—even though that’s often exactly where the most valuable clues lie? We talk about how product usage data, support interactions, churn signals, or expansion potential can be used to derive not only initiatives for Customer Success but also important learnings for targeting, prospecting, and the ideal customer profile.

And we also look at what this means for companies in very practical terms: What role is needed internally for a topic like this to make any progress at all? How do you create value for the teams that are supposed to work with it? And where might a sensible first step lie if you want to build Revenue Operations in a more data-driven way?

Interview

Maria-Lena Matysik: Hi Peter, welcome to Analytics Sprechstunde. I’m very happy that you’re here today.

Peter Eigenschink: Hi Maria, thank you very much for the invitation. I’m also happy that we can talk about analytics and data-driven marketing today.

Maria-Lena Matysik: The pleasure is all mine.

What topic have you brought to the podcast today? And would you maybe also briefly place the topic in the context of your background—that is, how you came to work on it?

Peter Eigenschink: Yes, gladly. The topic I’ve brought today is Revenue Operations. I’m not entirely sure how well known it is in the marketing world. So, if I may, I’d just describe it very briefly.

Maria-Lena Matysik: Please do.

Peter Eigenschink: Basically, the topic comes from the B2B software world, especially from software-as-a-service companies, where it has now become very well established. In principle, Revenue Operations is an organizational system that is meant to help a company grow its revenue sustainably. This is especially helpful in the B2B context because there you have very long and very complex customer journeys. Depending on the industry, it can take months to years for a prospect to become a customer.

In smaller startups, everything related to this comes together with the founder. But as soon as these companies grow and want to scale further, this often splits into three functions: demand generation or marketing, sales, and post-sales—for example, Customer Success, where onboarding takes place, the customer relationship continues to be maintained, and topics such as customer retention and customer expansion are handled.

Of course, this allows the company to scale by setting up these processes and teams and deploying specialized tools and staff there. At the same time, however, it leads to these teams often working somewhat in isolation. They have their own metrics by which they’re measured: In marketing, that can be the leads generated; in sales, the conversions—that is, how many new customers are won in a quarter—and in Customer Success, it’s something like retention. But it can happen that these metrics are then often no longer fully aligned with the company’s revenue goals.

And that’s where Revenue Operations comes in as a function. Its goal is to build systems, processes, and technologies across teams or functions that ultimately bring the teams back into alignment with one another.

Usually, this is a very data-based function. The decisions made there are very data-based—or at least the goal is to make them in a very data-based way. If I know, for example, which customers often convert, I can feed that into my targeting and prospecting decisions. I can feed it into sales decisions, for example in territory design, lead routing, or lead scoring. That’s roughly the goal of this function.

And that’s also what interests me about it—to tie it back to my background as well. I come from IT and have worked a lot for IT companies. I was always surprised by how much sales there works on, let’s say, a trial-and-error basis. On the other hand, through my background in marketing analytics, I’ve seen that a lot is happening, especially in B2C. Work there is very data-based, and in some cases very modern models are used, if you think of targeting, promotion planning, or pricing, for example. In B2B, there’s relatively little of that, for various reasons, I’d say. That’s mainly because you have less data and the journeys are very complex.

I found that interesting, and that’s why this topic interested me so much: understanding how the data feeds in there, what the challenges are, and how close you really are to this ideal or target state of steering the company’s entire revenue engine in a data-based way.

Maria-Lena Matysik: So what you’ve seen is: In B2B sales or B2B marketing, there’s little data, the customer journey is complex, and the teams sit partly in silos. Let’s tackle exactly that problem—it’s particularly exciting.

Peter Eigenschink: Exactly.

Maria-Lena Matysik: Was there perhaps a trigger, a particular situation or circumstance, where you thought: Okay, this is exactly why this problem should finally be tackled?

Peter Eigenschink: I think what you often see and also hear is that these teams—if you think of marketing, sales, and Customer Success—often, if not working against each other, aren’t necessarily working with each other either. There’s a lot of friction at these boundaries. When, for example, a marketing qualified lead is handed over to sales, the sales team still often prefers to work its own leads rather than the ones coming from marketing. This is often accepted as a given, even though you could, of course, do something to make these leads better.

And it’s similar with the second handoff, between sales and Customer Success, where there’s also some friction. In sales, it’s primarily about closing the deal. Whether the customer leaves again a month later because it isn’t a fit after all, or because something doesn’t work the way it was originally thought in the sales conversation, isn’t that present or central in the sales process.

I found these observations very interesting. They, I’d almost say, puzzled me a little—because it would actually be possible to improve these things. And that’s exactly what interested me: why it’s so difficult.

Maria-Lena Matysik: Provocative question: Why do you think data, of all things, is the solution to these conflicts?

Peter Eigenschink: I don’t think data can solve this on its own. It is—and you see this again and again—of course a huge change management problem. But I do think data can help a great deal in getting a somewhat more objective view of the whole thing, one that leaves less room for debate—or that you can at least commit to across teams, as a shared approach or a shared data foundation. Often, it’s already an issue that metrics or definitions aren’t the same across the different teams.

So, to begin with, it’s very central to get a shared data foundation or a shared view of the situation without slipping into opinions, assigning blame, or whatever else. And there, I think, data can definitely provide a more objective view of the situation.

Maria-Lena Matysik: Absolutely plausible—I can follow you well. Then the next question, of course, would be: How do you solve the problem with data? From your perspective, what could contribute to resolving this conflict?

Peter Eigenschink: I think there are a lot of starting points, and these Revenue Operations teams are becoming more and more established in companies. As I hinted earlier, I tried to look at the topic through an academic lens. I had conversations with many people in this field—with experts who work in exactly these Revenue Operations positions or in one of the go-to-market functions, that is, marketing, sales, or Customer Success. The goal was to find out: Where do things stand there? What are the challenges? Where is data really used to improve decisions, and where is a lot still based on intuition? And where are possible starting points for improving the approach?

To give a bit away up front: One of the most interesting takeaways I’d like to share is that, in my opinion, existing customer data is often used far too little. Of course, that’s also partly due to how companies grow. When a company is growing and doesn’t have many existing customers yet, you naturally don’t invest in post-sales and Customer Success, but rather in customer acquisition. That’s often where a lot of attention goes—including in terms of data infrastructure and analytics effort—while the number of existing customers keeps growing.

And that’s actually where incredibly rich information lies. These are the customers who already use my product and have already gone through the whole process. I have direct access to these customers and could ask them things directly. At the same time, I have my first-party information and can really observe how they use the product, whether they use it less or more, or whether they churn.

And I don’t have to hope that … Ultimately, that’s exactly what I more or less want to predict in acquisition: Is this someone who will use my product, who will also get value from it, and who will pay for it in the end—or not? With existing customers, I can observe exactly that: How do they use the product? Do they use it more at some point? Do they perhaps at some point use a second or third product of mine, or more features? Is the number of users of the product growing?

These are all extremely interesting signals that can ultimately also inform my prospecting and targeting activities and my sales activities. Who should I even be reaching out to? What does my ideal customer profile look like—based on how long customers really stay, that is, on a longer-term customer lifetime value, and on which customers really keep expanding?

That was one of my most interesting takeaways. It often gets a bit overlooked: Managing existing customers and retention management—that’s Customer Success’s job. But actually, existing customers should also be seen as a rich source of information. I think that’s one of the biggest levers there is—and it also helps close these feedback loops on a larger scale, ideally from the end of the funnel to the beginning.

Maria-Lena Matysik: Can you walk us through that a bit more? How exactly should I picture it? I have a database containing all my existing customers with their attributes—for example, as you said, which features they use, how long they’ve been customers, and so on. How exactly can I now connect this information with my prospecting team right at the beginning?

Peter Eigenschink: Maybe with an example from one of my conversations that I found particularly interesting—with [name to be verified], also a RevOps specialist. They started by analyzing their existing customers. They looked at exactly this usage data: Which features do the customers use? How often do they use the product? What Customer Success and support interactions are there? But also: How much do they pay each month, that is, how big is the account?

They in turn combined this with third-party signals. If, for example, a relevant contact changes—in their case, the company’s CFO—the risk that the customer will churn may be high. They drew on these various pieces of information to build a scoring model. It ranks existing customers by how complex the use case is, how large the account volume is, how likely a customer is to expand—that is, to use the product more, more broadly, or more deeply, adding more features or more users—and how high the churn risk is. Based on that, they then segmented their customers.

Maria-Lena Matysik: Mm-hmm.

Peter Eigenschink: They identified the customers who were really at risk of churning and scheduled quarterly business reviews with them in time. They identified those who were likely to expand their account and tried to support that.

At the same time, though, they also identified customers who actually aren’t a good fit for the company. I find that one of the most interesting aspects: that they figured out for themselves who the product might not suit so well—and actively helped those customers churn. Because they realized: These are actually customers we don’t necessarily want to keep, because the product doesn’t deliver what they hope to get from it.

Maria-Lena Matysik: Mm-hmm.

Peter Eigenschink: And they’re also not trying—at least not at the moment—to rebuild the product so that these customers stay at all costs. That’s also, I think, one of the most interesting aspects, because you can feed that back into customer acquisition. If I find out that there are customers who actually aren’t a good fit—maybe because they come from a certain industry, are a certain size, or have or haven’t seen a certain level of growth in recent years—then I can identify such customers early on as a poor fit.

At the same time, they also managed to identify the customers who benefit them over the longer term and have a high lifetime value. For example, they found that it works well for them to target smaller startups early on and then, so to speak, grow along with them. That was important information for the entire customer acquisition effort and for targeting in marketing.

Maria-Lena Matysik: And was it the case, for example, that an analysis of existing customers was done once a quarter and you saw: Okay, our ICP has changed—we based our prospecting on this ICP, but now we’ve learned that ideal customers actually look different? And then there was an ad hoc analysis and an adjustment of the ICP?

With the scoring model, I could imagine that it really happens more in real time: When some trigger fires, Customer Success should take action immediately. But listening to you, it seems harder to observe trends and then know: When do I feed these trends back up to the front of the funnel?

Peter Eigenschink: Yes, you’re absolutely right. There’s, of course, a certain inertia in this system.

Maria-Lena Matysik: It doesn’t have to be a bad thing.

Peter Eigenschink: No, it doesn’t have to be. I think that’s a general issue for these companies: Okay, I’m analyzing my customer base—but how do I reconcile that with my corporate strategy, which may also deviate from it somewhat? In this case, they certainly didn’t update the ICP in real time, but they did update it regularly, of course. And I don’t think Customer Success needs to act in real time in their case either. What’s more important is getting the priorities right—so that you don’t realize just two days before the renewal date that someone might churn, but ideally at the start of the quarter.

But these are all living systems, of course. Basically, I’d say it’s never done. The whole point is to close these feedback loops so that information can keep flowing back. At the same time, as you say, it often makes no sense to feed it back in real time. You need certain cycles to be able to collect the information and then feed it back. That creates a certain inertia. What’s still important is being able to learn from the actual information.

Maria-Lena Matysik: Apart from the technical perspective—that is, how the data gets from one team to another: Did you also come away with learnings about how the people in the teams work together? For example, did the company introduce regular meetings explicitly intended for looking at this data together? Or were there processes for collaboration that worked better across teams?

Peter Eigenschink: I think what’s definitely essential is that this Revenue Operations role really exists—and that it also has a certain mandate, because it’s an important task. That also manages to align the teams to some degree. But in the end, as you say, it depends on how the people work together.

And the most important thing, really, is always—I’ve also experienced or picked this up in other conversations—that there’s direct value for the people who carry it out operationally. They need to get value from it, especially when things change. If you think of Customer Success in this example: You may have a lot of Customer Success managers who are already stressed. These teams often have a very tight budget and have to look after a great many accounts. Prioritization like this helps them directly, of course, when they really know: These are my most important accounts, these are the ones I need to take care of this month. Then they’re naturally also much more open to working with the Revenue Operations teams.

At the same time, of course, it also has to deliver this value directly to the other teams and functions. Ideally, it leads to everyone being able to work a bit more efficiently and having to do less by hand. After all, Revenue Operations often brings automation with it. That helps the teams overall to be under less pressure or to make better use of the resources they have. But at the end of the day, it also has to help in day-to-day operational work.

Maria-Lena Matysik: Absolutely, a very, very good point. In the end, it’s always a bit about “What’s in it for me?”—and then I’m in. You can think that’s good or bad, but if you know it, you can use it. Maybe one last question: If, for example, a RevOps manager or RevOps lead asked you: What would be the first step or an important point I could start with to think about my role in a more data-driven way?

Peter Eigenschink: I can really only repeat my point: I think the treasure lies buried in your existing customers.

A RevOps manager or Head of RevOps naturally has to achieve certain wins for the company as quickly as possible. These can lie in different places, and it always depends on all sorts of things. The biggest challenge is definitely change management, that is: How do I change processes? If you find out in which areas or at which interfaces there’s the greatest willingness to change something—or also a certain intrinsic motivation, perhaps because people can make their own working lives easier as a result—then of course that’s where you should start. The second issue is the data foundation, which is often poor or difficult.

Maria-Lena Matysik: Mm-hmm.

Peter Eigenschink: In sales in particular, there are some approaches that improve this. For example, analyses of call transcripts help populate CRM systems automatically—with very mixed results in terms of quality. But in principle, these are helpful approaches that save employees work and at the same time improve data quality.

In general, though, as I said, I’m a strong advocate of starting with existing customers. I think you gain a lot by observing your existing customers: by observing product usage and how customers interact with the Customer Success or support team—for which you can, of course, use transcript or text analyses just as well. And if you can then connect this information further through the funnel.

After all, the funnel is often very one-directional: Something goes in at the front and comes out at the end. Closing the entire loop would be the big goal, so to speak. But there are also parts you can close, often at the interfaces between teams, at certain handoffs. I think you can also achieve a lot of quick wins there by improving these handoffs and the processes around them. But you really do get the most direct and most reliable information from your existing customers.

Maria-Lena Matysik: The perfect closing words. Thank you so, so much for all the insights, your perspective on the topic, and, of course, for sharing.

Peter Eigenschink: Thank you very much for the conversation, Maria. It was an interesting conversation.

Maria-Lena Matysik: Thank you!

Outro

Maria-Lena Matysik: What I’m definitely taking away from this conversation: At several points, I had a little aha moment. And that’s exactly what I really love about interviews like this—that I sometimes hear something that isn’t really new, but that still clicks in a new way in that moment.

One takeaway for me was: Data can’t simply dissolve silos. Just because all the data sits in a data warehouse and is the famous single source of truth doesn’t mean the silos “in people’s heads” dissolve. But data can create a common foundation on which teams can talk to each other meaningfully in the first place. Especially when marketing, sales, and Customer Success look at different goals and metrics—that is, have a different focus—you often first need a shared view of reality. And that’s where data helps.

I also found the idea exciting that joint meetings or processes sometimes don’t help at all, and that what’s needed instead is an overarching role that brings exactly these perspectives together. In other words, someone who keeps the big picture in view and not just a single part of the funnel.

And probably my biggest takeaway from this episode: Existing customer data is often an underestimated lever. Anyone who understands which customers really fit, grow, or drop off can also align marketing and sales at the front of the funnel much more precisely. I found it extremely exciting to hear that from Peter’s B2B perspective.

And in the end, one thing is very clear to me once again: Data only works if it really helps people in their everyday work. If an analysis or a process doesn’t bring any real value and doesn’t solve a real problem, it won’t be used. Sometimes it’s that simple.

And with that, I’ve reached the end again 🙂 Follow me on LinkedIn and feel free to send me feedback on this episode—or your ideas for future episodes: topics or guests you’d like to hear on Analytics Sprechstunde.