Question the Default: How a Simple Fix Can Save Millions in Ad Spend

Isin Acun
Mar 5
4 min read

An Interview with Dr. Uğurcan Dündar by Işın Acun, WU Vienna

When a company runs an online ad campaign on platforms like Google or Meta, it sets a daily budget, say $10,000. But that budget has to be spread across the day, and the platform needs a rule for how to spread it. Should it distribute the money evenly, hour by hour? Or follow the natural rhythm of web traffic, spending more when more potential customers are online? This decision, known as budget pacing, happens automatically in the background for millions of campaigns every day, governed by a default setting that most advertisers never question.

In Getting the Pace Right: Performance of Budget Allocation Heuristics in Online Advertising, Dr. Uğurcan Dündar (WU Vienna) and co-authors Nadia Abou Nabout (WU Vienna) and Bernd Skiera (Goethe University Frankfurt) show that the platform default, Even Pacing, which spreads the budget uniformly across hours, consistently generates lower profits.

An overlooked alternative, Waterlevel Pacing, follows natural traffic patterns—spending more during high-traffic hours when more users are online and less during off-peak periods. Waterlevel Pacing can outperform in profitability even sophisticated profit-maximizing models that optimize bids hour by hour when predictive measures contain errors, as they inevitably do.

Building on this, the authors develop a new heuristic, Adaptive Waterlevel Pacing (AWP), which weights each hour’s budget according to its expected conversion potential rather than traffic alone. AWP performs within a few percentage points of the maximum achievable profit model (a model with perfect knowledge of market conditions), while remaining simple enough to implement in practice.

“We found something counterintuitive: more sophisticated models only outperform simpler ones when their inputs are accurate. When forecasts are noisy, additional complexity often hurts more than it helps. It can amplify errors in prediction instead of correcting them."

— Uğurcan Dündar

The Workout Epiphany

What surprised Uğurcan was not that Even Pacing underperformed. His engineering intuition had already told him that defaults can be poor performers. The real surprise was how much small changes could outperform the industry standard. It was this realization that led to the paper’s most novel contribution.

The idea for Adaptive Waterlevel Pacing came during a period of obsessive focus. “My workouts were taking around two hours, because I was constantly thinking about pacing. In my dreams, commutes, showers, everywhere,” Uğurcan recalls. During one of those extended gym sessions, it clicked: what if they could embed the profit-maximizing logic into the simple, prediction-light allocation rule of Waterlevel Pacing? After some trial and error, AWP came together. The algorithm performs close to the theoretical optimum in a fraction of the computation time, making it practical even for advertisers without dedicated data science teams.

“The key issue isn't how sophisticated your model is. It's whether the added complexity is actually worth it given how well you can predict. And when that prediction ability is limited, more complexity can end up hurting more than helping."

— Uğurcan Dündar

Practical Advice for Managers

The message for practitioners is direct: question the default. Even Pacing is the default setting in major platforms like Google, Meta, and Microsoft because it is the simplest implementation rule: just divide the budget equally across hours, no forecasts needed. It is widely adopted because it is convenient, not because it is optimal. Uğurcan urges marketing managers to ask a simple Monday morning question: “How is my budget actually being paced?” If the answer is “the default,” that is a red flag.

“The temptation is always to reach for something more complex. But honestly, before you do that, it's worth just checking whether a simpler approach might already work just as well, or even better.”

— Uğurcan Dündar

For managers, the implication is clear: test alternatives, ask your agency to justify the pacing setting, and do not assume that more complex tools are automatically better. In this case, a simpler proportional rule, allocating budget in proportion to each hour’s expected value, outperformed full-scale optimization.

A Smooth Review and What Comes Next

The review process at IJRM was rapid and remarkably smooth. Uğurcan thanks the review team for being concise and constructive and gives credit to his co-authors for helping position the paper around a practical gap: no study had comprehensively compared the most common heuristics based on profitability. His advice for fellow PhD students: “Always keep your research question, contribution, and audience in mind when you are putting things together.”

With this job market paper now published, Uğurcan plans to stay in the digital advertising space, specifically questioning the effectiveness of other defaults (e.g. behavioral targeting). He is also pursuing a growing interest in field experiments, with two projects currently in progress.

Read the paper

Interested in exploring how budget pacing heuristics compare and what this means for your advertising strategy? Read the full paper here.

Want to cite the paper?

Dündar, U., Abou Nabout, N., & Skiera, B. (2025). Getting the pace right: Performance of budget allocation heuristics in online advertising. International Journal of Research in Marketing. https://doi.org/10.1016/j.ijresmar.2025.09.004

Meet Uğurcan

Assistant Professor, WU Vienna, Austria

If you weren’t an academic, what would you be?

“I would definitely be working in aviation, in a department where I can do optimization. Maybe in continuous process development, or scheduling crew assignments.”

What’s the best piece of advice you’ve received?

“My supervisor always told me that “Academia is not just a sprint; it is a whole marathon.” I still try to remind myself of this, almost every day. I really saw that if you try everything you can and be patient for long enough, most of the problems actually do resolve themselves.”

What’s the number one question you hope to answer during your career?

“Across everything I work on, the core question stays the same: what works, when does it work, and why? Ultimately, I want to help design decision-making systems that don’t just perform well in theory, but hold up under the uncertainty and messiness of real markets.”

This article was written by

Işın Acun

Ph.D. candidate at the WU, Vienna