Elsa Ögren – Snap Inc.

Marketing has never had more data, tools and ways to measure effectiveness. Yet for Elsa Ögren, Head of Marketing Science Northern Europe at Snap Inc., that does not necessarily mean marketers are making better decisions. After starting her career on the media agency side and spending seven years at Meta before joining Snap, Elsa has seen measurement evolve from traditional brand metrics, through the rise of digital attribution, to today's more fragmented landscape. Her view is that sustainable brand growth requires something more fundamental: aligning what the business wants to achieve with what campaigns optimise towards and, crucially, what marketers choose to measure.

In this conversation, she explains why no single methodology provides the full answer, why experimentation is essential for innovation, and why the future of Marketing Science may be less about building dashboards and more about helping organisations make better decisions.

Aligning Business Goals, Optimisation and Measurement

For Elsa, one of the biggest opportunities for improving marketing effectiveness sounds deceptively simple: make sure the business objective, campaign optimisation and measurement are all working towards the same outcome “Almost every time I start working with a new client, I ask quite quickly: what is the goal? What do you want to achieve?” she explains. That first part is often relatively easy to answer, whether the objective is to drive sales, traffic or another business outcome.

The next step is ensuring campaigns actually optimise towards that objective. Technology has made this increasingly straightforward. The problem often appears at the third stage: measurement. “Business goal, optimisation and measurement should all be connected and aligned,” Elsa says. “And that third part, measurement, is where we often see it fall apart.”

When measurement is heavily driven by attribution, teams can end up optimising towards whatever the measurement system rewards. “What happens when teams then come in and optimise is that they will optimise towards what your measurement system is rewarding,” she explains. The consequence is that media which does not generate many direct clicks can be undervalued, including brand-building activity. At the same time, more credit may go to media associated with purchases that might have happened anyway, while factors such as offline sales, seasonality and promotions can be overlooked.

“Business goals, optimisation and measurement need to be aligned. And that third part, measurement, is where we often see it fall apart.”

From Measuring Everything to Measuring what Matters

Digitalisation fundamentally changed marketers' expectations of measurement. As more consumer behaviour became trackable, performance marketing and attribution gained prominence, allowing marketers to connect activity to lower-funnel outcomes much more easily. But Elsa believes that accessibility also created an unintended consequence: the industry became accustomed to the idea that everything should be measurable.

That assumption is becoming increasingly difficult to maintain. Privacy-related changes and the loss of signals have made attribution less effective and accurate, while at the same time encouraging renewed interest in approaches based on aggregated data, including Marketing Mix Modelling. But Elsa sees a broader lesson in this evolution.

“What happened with digitalisation, when we became used to being able to measure everything, was that we also became a little obsessed with being able to measure everything.”

In reality, she argues, measuring everything is neither possible nor necessarily desirable. Instead, measurement should begin with prioritisation. “We need to have a plan where we prioritise: what are the client's most important questions?” Elsa says. Those questions can then be assessed against both their potential impact and the effort required to answer them.

Testing many times comes with both an opportunity cost, such as holding out part of the audience from receiving media, and a resource cost in terms of time and effort, which makes the decision that follows particularly important. “I always want to make sure that the value of testing is greater than the investment,” she explains. “And the value is very closely connected to whether you act on the results you get.” That means being clear beforehand about what decision the research will inform. “If we just test and don't do anything with the results, then we should skip it.”

It also means accepting that not every decision needs the same level of evidence. Tactical choices can sometimes rely on established best practices or previous learning, while larger strategic decisions may justify robust experimentation. And after years of trying to quantify everything, Elsa believes marketers may occasionally need to become comfortable again with something less measurable: “To some extent, we may have to rely a little more on gut feeling and intuition.”

“If we test something and then don’t do anything with the result, we might as well not test it.”

Why Marketing Measurement needs Triangulation

If measuring everything is unrealistic, relying on one measurement method is equally problematic. Elsa describes the alternative as triangulation: combining different approaches while understanding what each can and cannot tell you.

MMM, for example, can provide a valuable cross-channel view of marketing effectiveness and help capture effects that short-term attribution can overlook. But because these models are largely based on historical data, relying on them alone creates another risk. “If we only rely on MMM and don't work with other measurement methods, we'll never learn anything new,” Elsa explains. “We'll just copy and paste what we've done before.”

That is where experimentation becomes important. MMM can identify patterns and generate hypotheses, while experiments such as      lift studies can test whether changing an approach actually creates incremental impact. “You can use MMM to come up with hypotheses or see trends but then testing them is crucial for innovation.”

The point of triangulation is therefore not to accumulate as many metrics as possible. It is to use methods for what they are good at and allow them to complement each other. “It is about understanding what advantages and purposes certain methods have, and what gaps they have that can then be filled by others.”

“If we only rely on MMM and don't work with other measurement methods, we'll never learn anything new. We'll just copy and paste what we've done before.”

 

Testing as an Engine for Brand Growth

This philosophy makes test-and-learn more than a measurement exercise. For Elsa, it is also an important ingredient in innovation. One of the foundations she believes organisations need is “a test-and-learn mindset embedded in the culture.” While that may sound straightforward, getting an entire organisation to embrace continuous learning is considerably harder in practice.

Even when a business is performing strongly, Elsa argues that marketers should continue challenging what they do. Organisations risk continuing to repeat yesterday's successes while the market around them changes. “You have to continuously work on that part, which becomes a way of innovating and changing what you're doing today,” she says. “Especially in an environment where there is so much change all the time.”

This is particularly relevant in a rapidly evolving media landscape. At Snap, Elsa sees new advertising formats and measurement solutions emerging regularly, creating opportunities that cannot always be evaluated using historical evidence. Best practices can provide a starting point for lower-risk decisions, but experimentation becomes more valuable when marketers are considering larger strategic changes. Growth, in that sense, comes not from constantly changing direction, but from continuously creating new evidence about what could work better.

“Even if things are going really well, I think it's important to keep asking: how can we become even better?”

The Changing Role of Marketing Science

Looking ahead, Elsa expects AI to make analysis faster and increasingly accessible. But as technical tasks become easier to automate, she believes the human side of Marketing Science will become more important. The value will increasingly lie in asking the right questions, understanding the context and quality of the data, and translating findings into insights and recommendations that people can actually use.

In that sense, the role is shifting from being primarily a technical one towards becoming more of a “conductor”: bringing together different sources of evidence and turning them into direction. AI can accelerate the analysis, but marketers still need to decide what is worth measuring, what the evidence really means and, ultimately, what to do with it. That reinforces Elsa’s broader point: better measurement is not about producing more data, but about using it to make better decisions.