DTC Podcast – Bonus Episode: "How to Prove What’s Actually Working in Your Ad Spend — The Incrementality Mentality" with Stella's Brenden De La Rua

Release date: October 29, 2025
Host: Eric Dick (DTC Podcast/Newsletter)
Guest: Brenden De La Rua (Co-Founder, Stella; Media Buyer; SaaS Founder; Marketing Content Creator)

Episode Overview

In this bonus episode, host Eric Dick sits down with Brenden De La Rua of Stella to tackle one of the most persistent challenges in e-commerce marketing: understanding what’s truly incremental in advertising spend. The conversation covers the confusion around ROAS vs. ROI, making incrementality testing accessible, why traditional attribution models fall short, and how Stella aims to democratize advanced marketing measurement for direct-to-consumer (D2C) brands. Brenden also shares tactical insights on running incrementality studies, the evolving landscape of creative testing, and how he uses TikTok to educate and build Stella’s brand.

Key Discussion Points & Insights

The Problem with ROAS, ROI, and the "Black Box" of Incrementality

• ROAS often isn’t real ROI. Many brands mistakenly focus on platform-reported Return on Ad Spend (ROAS) without taking into account incrementality or saturation, leading to misinformed decisions.

  “So I always gravitated to it as a media buyer… performance is good, now I don't make a change, next week it's bad...what could it be? ...eventually the answer I stumbled upon was incrementality...Your ROAS does not mean ROI, right.” — Brenden (01:16)

• Saturation and diminishing returns: Brands can hit a "wall" where extra spend doesn't yield proportionate gains, but this is often invisible in traditional metrics.

  “Most of the time with channels … they might already be at a point of saturation, meaning if they spend more, they're going to start getting diminishing returns.” — Brenden (18:35)

• The complexity of incrementality is exploited by expensive vendors:

  “I think it's so confusing. It's over a lot of marketers' heads and that's why a lot of companies are charging so much for it because they're saying, ‘This will achieve more incremental revenue or profit for you.’ It's like a black box for how it works.” — Brenden (00:25, 24:56)

Stella: Making Incrementality Accessible

• Built out of necessity: After struggling with attribution confusion as a media buyer, Brenden co-founded Stella to provide D2C brands—especially in the mid-market—a more cost-effective, transparent, and easy-to-use tool for incrementality testing and media mix modeling.

  “Me and my co-founder Vinnie built Stella to kind of democratize this space and make incrementality testing much more accessible and easier to understand and much more affordable for D2C brands.” — Brenden (01:30, 00:42)

• Pricing philosophy: Stella intentionally undercuts legacy solutions, offering transparent and affordable pricing:

  • $2,000/month, cancel any time
  • $1,000/month (annual plan), unlimited tests
  • Seven-day free trial, no credit card required

  “Stella runs at $2,000 per month … an annual plan that's just $1,000 a month paid annually … unlimited uses of the incrementality testing tool.” — Brenden (17:00)
  “I want to have like the highest accuracy, but I also want it to be the most affordable tool by far, far. Like, I don’t want anyone to compete with us.” — Brenden (17:24)

What Is Incrementality, and How Should Brands Use It?

• Traditional attribution is misleading: Branded search and retargeting campaigns may show the highest platform ROAS, but are often the least incremental; they capture conversions that likely would have happened anyway.

  “Top of funnel campaigns are typically the most incremental, where the bottom of funnel campaigns are the least incremental ... branded search is just non-incremental. Even though it has the highest ROAS.” — Brenden (03:43, 04:06) “Retargeting is typically the highest ROAS, but usually the least incremental.” — Brenden (04:10)

• True incrementality is measured via geographic holdouts: Run regional experiments, measuring total revenue in "test" vs. "control" territories, not just the in-platform numbers.

  • Stella's approach: Helps users select the optimal test/control regions algorithmically, run statistically sound studies, and analyze true business impact using a brand’s source-of-truth platform (e.g., Shopify, WooCommerce).
  • Example: Only run CTV in specific states and compare total ecommerce revenue to the matching non-CTV regions.

  “The most important part of a holdout study is selecting the right locations…Stella will tell you the best regions to compare to have a high correlation…” — Brenden (05:42)

• Role of surveys and MTA (Multi-Touch Attribution):

  • All three (MTA, post-purchase surveys, incrementality/causal analysis) fill different parts of the measurement "picture."
  • Rough rule-of-thumb: MTA = 40%, surveys = 20%, causal = 40% of the "truth".

  “I think MTA shows like 40% of the picture...Post-purchase surveys extend that to about 60%...Causal analysis like incrementality or media mix model comes in shows that additional 40%...” — Brenden (07:36)

Always-On Incrementality & Stella's Product Innovations

• Marketers want ongoing, easy-to-understand insights.
  • Snapshot vs. always-on: Traditional incrementality is a point-in-time read; Stella’s “Always-On Incrementality” attempts to deliver daily, automated, rolling causal analysis at the campaign level.
  • AI assistant: Stella’s AI interface provides recommendations, surfacing valid tests and advising on what actions to take. Inspired by marketers using ChatGPT for rough analysis, but made accurate and safe for big budget decisions.

  “Incrementality studies don't really work like [a daily dashboard]...They're like a snapshot in time...So we've built a tool called Always on Incrementality as like a response to those requests.” — Brenden (09:08) “Our idea was like, what if we made Stella to mirror like a GPT...Because ultimately we're building a data science tool for marketers, not for data scientists.” — Brenden (11:45)

Media Mix Modeling (MMM) Philosophy

• Blended open-source and proprietary approaches:
  Stella integrates Bayesian (preferred for flexibility with priors) and frequentist models, and is inspired by open-source initiatives (PyMC, Meridian/Google, Robin/Meta).

  "So we've leaned towards Bayesian models...We can put that IROAS as a what's called a Bayesian prior into our MMM to help calibrate...over time..." — Brenden (14:04) “...collaboration of building a model that is repeatable and accurate [but] easy to use…” — Brenden (15:03)

Practical Case Studies & Platform Findings

• Scaling with caution:

  • Run incremental scale tests (e.g., hold spend constant in some regions, increase by 250% or 500% in others) to map the marginal growth curve and determine spending limits before diminishing returns set in.
  • Example: Brands often learn they’re already at saturation on core channels.

  “What this test allows you to do is … start mapping out this marginal growth curve so you can see how far you can scale before you hit a point of diminishing [returns]…” — Brenden (18:34)

• Platform-by-platform incrementality ("thirstiness"):

  • Applovin: Incrementality almost always matches in-platform reporting (Iroas ≈ ROAS), likely due to click-based reporting.

    “With Applovin, we have seen that pretty much every single time we've run a study. ...Iroas in Stella is very close to the Iroas the applovin is reporting.” — Brenden (20:26)

  • Pinterest, Meta, Google: Generally positive incrementality, though it varies by campaign type and structure.
  • Two unnamed platforms are notably "thirsty" (claiming more credit than justified)—Brenden declines to name names.
  • CTV: Continues to be highly incremental across vendors.

  “Meta campaigns and their average incrementality factor is like 121%...Google is also pretty incremental...” — Brenden (23:00)

Common Fallacies & Misconceptions

• Incrementality is overcomplicated and oversold:
  • Smaller brands likely don't need advanced tools—diminishing returns analysis is most useful for $10M–$100M annual revenue D2C brands.
  • The industry often shrouds techniques in jargon unnecessarily.

  “I think incrementality has a place especially for brands that are spending enough. ... If you're below 10 million in annual revenue...you probably don't need an incrementality tool at all.” — Brenden (24:56)

TikTok Content & Community Building

• Started with silly, viral content, transitioned to marketing education.
  • Audience engagement has led to substantial business and real brand awareness for Stella.
  • Focuses on foundational, actionable advice for media buyers, especially those struggling with Facebook/Meta's changes (e.g., elevated CPMs, “Andromeda” algorithm update).

  “I just started talking about marketing because that's all I do...I get the comment I get the most is like, I feel like we should be paying you for this, like, content...” — Brenden (27:06)

The Post-Andromeda Meta (Facebook) Approach

• Ad account structure:

  • Use a simplified two-campaign strategy:
    • Testing (broad, heavy exclusions, ~10% budget)
    • Scaling (broad, fewer exclusions, 10x+ budget)
  • Pipeline: Identify winning ads in Testing, transfer them (by Post ID) into Scaling.

  “...two campaign approach, a testing campaign and a scaling campaign. And that's it...If ads are winners against a new audience, I copy the post ID and move it to the scaled campaign with a 10x budget.” — Brenden (30:20)

• Creative diversity boosts incrementality even in broad-audience campaigns. Use varied creative to access new sub-audiences and increase reach.

  “Creative diversity 100% can increase the incremental reach of your campaigns, even if it's a broad audience, you're targeting everyone in America.” — Brenden (32:54)

• How to analyze ad creative: Profile new ads by “Neanderthal review” (describe the first three words that come to mind) to ensure diversity, and adopt metrics like view-through rate and social shares, not just direct CPA/ROAS.

  “Reviewing your ads like a Neanderthal, like, saying the first three words that come to your head to understand if the ad is diverse enough.” — Brenden (33:20)

AI, The Future of Media Buying, and Human Judgment

• AI won’t fully replace human decision-making soon, but tools like Sora, UGC video, and Stella’s AI-powered assistant will make marketers more effective and creative.

  “I'm very hopeful...I think AI is going to play a big role. Obviously I don't think you can like get away from it right now, but we'll see how long we have until it just takes over completely...” — Brenden (36:40)

• Still a need for human context: Media buyers must interpret and apply insights, especially as each ad platform tries to take as much conversion credit as possible.

Notable Quotes & Moments

• On ROAS vs. ROI:

  “Your ROAS does not mean ROI, right.” — Brenden (01:18)

• On Incrementality Pricing Irony:

  “The minute you sign for those platforms, you're immediately less profitable. Like, and by a lot.” — Brenden (15:58)

• On overcomplicating incrementality:

  “I think people are overcomplicating it…They’re like, ‘I need to be in the loop because people are talking about it, but I don't understand it or how to use it.’” — Brenden (25:00)

• On TikTok as a growth engine:

  “I'm speaking to tons of people that are finding me from TikTok…I'm getting back into the world of just like very basic…media buying tips…I'm like helping people.” — Brenden (27:06)

Timestamps for Key Segments

• 00:00–01:40: ROAS ≠ ROI and Stella’s mission
• 03:29: Why media buyers miss incrementality and common mistakes
• 05:25: How to set up an incrementality study in Stella
• 07:32: The role of multi-touch attribution and surveys
• 09:08: Always-on incrementality & daily dashboards
• 11:31: The AI/ChatGPT-like interface of Stella
• 13:21: Stella’s approach to media mix modeling (MMM)
• 15:44: Pricing transparency and philosophy
• 18:34: Case studies — how clients pivoted spend and avoided saturation
• 20:26: Platform-by-platform findings (Applovin, Meta, Pinterest, Google)
• 24:56: Misconceptions and overcomplication in the incrementality industry
• 27:06: Using TikTok to educate and attract customers
• 29:49: Meta "Andromeda" algorithm and ad account structure
• 32:54: Creative diversity and measuring incremental impact
• 36:40: The future of AI in media buying

Conclusion

This episode serves as a comprehensive masterclass in modern marketing attribution, incrementality, and how to avoid the pitfalls of over-investing in what platforms say is working. Brenden’s perspective—equal parts technical and practical—offers genuine tactical takeaways for any D2C brand operator, media buyer, or marketing strategist aiming to squeeze more real business value from their ad spend.

To try Stella, visit heystella.com and follow Brenden on TikTok at @brendenbuilds.