Most companies still can’t say how much money their data, automation and MarTech earn or save them. Frans Riemersma, founder of MartechTribe and co-author with Scott Brinker of the MarTech Landscape and the State of Martech report, argues that the hard part comes before any tool: deciding what you want it to do. In his second appearance on Data Driven Voices, Frans and host Emma Storbacka swap notes on use cases, scale, fast and slow data, and why AI leaves leaders with fewer places to hide.
In this blog, we summarize some of the key takeaways from the episode.
Buying technology is easy, deciding is hard
Frans started out as a marketer in the late 1990s and moved into coding, because features were what people listened to. His first shift toward value came some 13 years ago, on a global software implementation for a bank. The tool had 13 modules, and Frans recommended the bank buy five. It turned out the bank needed two, and two buttons in them made a difference of 1.2 million euros a year.
At the time he took it for a lucky story to tell at birthday parties. Years of building business cases with business school students taught him it was the rule. When you reverse engineer the value in a stack, you tend to find it in a small number of clear decisions about what the business wants to happen.
That’s also why a conversation about results feels uncomfortable. At Avaus, we started printing our client results on an office wall in 2012, and many of them were weak, even on million-euro implementations. Talking about results exposes past decisions, for the buyer and for the partner. Frans calls it courageous, and it’s where value engineering starts.
A lead scoring model is a capability
The companies that succeed define their use cases differently. In MarTech it’s easy to mistake a capability for a use case. A lead scoring model is a capability. Taking data into a lead scoring model and using it to decide which prospects your salespeople call next is a use case.
At Avaus, we break every use case into . The business result comes from the action, and the action is rarely technical. It’s operational work close to the customer and the sale, and it depends on leadership. Personalization follows the same logic. Its business case comes from how many personalized actions run, and across how many channels, markets, products and touchpoints.
Frans recognizes the pattern from his teaching. His students move from data to information to insight to actionable insight, and each step asks for a different mindset. He tells them to go to bed as a data officer and wake up as a new person, whose only job is to ask what the data says.
The split also works as a bridge between teams. When a use case reads as data, algo and action, IT, data and business stakeholders can each see their own part in it.
Outperformers do less, with more focus
Scale is the second thing successful companies get right, and Frans adds a twist to it. In his research he looks at the top 30 percent of companies by revenue per employee, because that ratio shows who performs best within an industry.
This way we can isolate outperformance and see what they do differently and systematically they do less with more focus. So they take away all the noise, all the stuff they don’t need.
Low performers try to be present at every touchpoint all day long, which Frans compares to stalking rather than marketing. Outperformers find the three to five things customers try to do with their products or services. Scaling the three to five journeys that drive half of your revenue is a far smaller task than scaling everything.
One retailer Frans describes bought three CDPs, each for a specific job, such as one e-commerce use case and identity resolution. Knowing what each tool was for gave them negotiating power, and they now run the stack in a way without calling it a strategy.
Focus still needs a plan for scale. In our experience, a single use case can involve around 30 people across teams, which is why takes a process and not only a backlog.
Fast data and slow data
Frans borrows a frame from Daniel Kahneman’s Thinking, Fast and Slow.
There is fast data and there’s slow data. And too often we treat data as data. It has all to be clean and complete. That’s the best way to paralyze your company and to burn your money and resources.
Slow data changes over three to five years. Customer lifetime value and market share are slow data, and so is a customer’s age. It calls for the deliberate thinking that leads to decisions, which is the thinking companies skip most often. Fast data lives in the moment, such as a visitor on your webshop right now, and it feeds dashboards, triggers and optimization.
The two work together. Slow data sets the guardrails, for example how large a discount a customer can receive, and a decision engine makes the fast choices inside them. Frans also questions the dashboard as the place where analysis ends. In his view, it’s where insight begins.
At Avaus, we draw a similar line between big brain and small brain analytics. Models such as CLV or lead scoring belong in your own data platform. Real-time choices, like which offer will convert a visitor, often sit better in the tool closest to the customer.
AI turns every insight into a decision
Thirty-eight minutes into the episode, Frans points out that neither of them has talked about AI yet. When the topic arrives, he ties it back to the discomfort of results. AI pushes people toward decisions. A prompt is a blinking cursor asking what you want, and for many people that is the hardest question of all.
Up until AI we were struggling to get insights. Now we have instant insights. So now is the question, of all the thousand insights, which ones matter? You have to make a decision again.
For years, management teams asked what they could do with AI. Now that the answer is almost anything, the question becomes what they should do. That is business strategy: where to compete, and where to place the bets.
Frans reads the hype cycle as a learning curve. At the peak, people externalize. They look for the best tool and move away from their own judgment. Outperformers come out the other side by internalizing, connecting data with experience. His attribution research with Scott Brinker found the same pattern. Outperformers used attribution models for the parts of a journey where the teams agreed the model matched reality, never for the whole journey.
Leadership, humility and the ability to respond
When AI makes insight cheap, leadership is what’s left. Frans sees humility in outperformers and separates it from weakness. Being humble means accepting that you can be wrong, and the leaders who claim to know the truth tend to be the insecure ones. He treats emotions as signals too, a dashboard most businesses choose not to read.
It’s like you’re driving in traffic and you’re ignoring all the signals and you have duct taped your dashboard. That is what we do in business.
He closes with a thought from his father: responsibility is the ability to respond. You can get it wrong, and after a few tries, asking why each time, you find the answer.
The marketing role is changing along the same lines. Frans cites recent research showing more Fortune 500 companies dropping the CMO title, while marketing moves into a wider growth or revenue role. In the Nordics we see the same shift toward . One commercial leader Emma met calls the data-driven sales engine she is building helvetinkone, Finnish for “the machine from hell”.
Key takeaways
- Define every use case down to the action. A model stays a capability until it changes what someone does
- Scale what matters. Three to five customer journeys can drive half of your revenue
- Separate fast data from slow data, and let slow data set the guardrails for fast decisions
- Outperformers, measured by revenue per employee, do less with more focus
- Treat AI as a decision problem. Instant insights make choosing harder
- Lead with humility. Admitting you can be wrong is what makes the results conversation possible
Common questions
What is the difference between a use case and a capability in MarTech?
A capability is something your organization can do, such as running a lead scoring model or personalizing content. A use case is a capability put to work on a business outcome, for example using lead scores to decide which prospects sales calls first. Avaus breaks every use case into data, algo and action, and the business result comes from the action.
What is the difference between fast data and slow data?
Slow data changes over years, such as customer lifetime value or market share, and supports deliberate strategic decisions. Fast data changes in the moment, such as a visitor’s behavior on a webshop, and supports real-time actions. Frans Riemersma of MartechTribe argues that slow data should set the guardrails that fast decisions run within.
How do outperforming companies use MarTech differently?
According to Frans Riemersma’s research on companies with the highest revenue per employee, outperformers do less with more focus. They scale the three to five customer journeys that drive most of their revenue, and they buy tools for specific, named use cases instead of trying to cover every touchpoint.
Why does AI make decision-making harder for leaders?
AI makes insights instant, so the bottleneck moves from finding insights to choosing which ones matter. The question shifts from what a company could do with AI to what it should do, which makes it a question of business strategy and leadership.
Inspiration for marketing, sales, and data professionals
Data Driven Voices is a podcast where Avaus together with industry experts, thought leaders, and partners discuss how to harness data, technology, and strategy to drive meaningful change and business results in primarily marketing and sales. The podcast shares actionable insights, success stories, and thought-provoking challenges to help professionals with new perspectives.
Related reading