Always Be Testing

Michelle Eatherton, Chief Business Officer at Liminal, shares how she convinced her co-founder to build the company partner-led from the very first call, before there was even a product. She and Tye dig into why mid-market enterprises are the underserved sweet spot for AI governance tools, and what it really takes to earn a hyperscaler co-sell relationship.

Jordan Bentley, founder of Audio Hook, explains how he built a performance-based podcast advertising model where brands only pay for real outcomes, not impressions. He and Tye talk about why AI made ad production easier (not the strategy), and why geo-based incrementality testing beats promo codes for proving what's actually working.

Michael Gants, founder and CEO of Encore, breaks down how apps can turn unused "streak" and celebration screens into a new revenue stream, without a single ad or paywall. He shares how brands like Disney, Uber, and Apple are paying to reach users in their happiest moments, and why founders should be more afraid of a broken CAC/LTV ratio than of running a monetization experiment.

Brook Schaaf, co-founder of FMTC and author of the new book Affiliate Hypotheses, joins Tye to make the case that affiliate marketing might be the internet's best — and sometimes only — way to monetize the open web. They dig into why walled gardens like Google and Meta get outsized credit for conversions, and why AI search is putting the open web's survival at risk.

Vikram Chalana, CEO and founder of Pictory, joins the show to unpack how his AI video platform went from 50 customers to 5,000 in just 60 days — and why the breakthrough came from changing the market, not the product. He digs into Pictory's State of Video report (AI adoption has no Silicon Valley bias — Florida and Indiana are leading users), the design principles that keep Pictory out of the "timeline" trap of tools like Premiere, and what's next: code-generated, interactive video that could reinvent how content gets made and consumed.

Remy's Danielle Sakher explains how her team balances speed with observability and security as AI coding tools move from novelty to enterprise infrastructure. She shares what it's really like selling an AI builder to security-conscious industries like financial services, plus a wild stat: alpha users have already spent $300K in AI compute.

Darryl Yearwood, who leads partnerships at Apollo, makes the case that a smaller, higher-quality partner roster beats chasing volume every time. He and Tye get into what "not plug-and-play" really means in affiliate marketing and how to build partnerships that actually move revenue.

Avishag Daniely built One Layer's entire marketing and partnerships function from scratch — coming from a background in IDF cyber intelligence. She shares how her military analytical training shaped her approach to B2B GTM, what it takes to build pipeline when no one's heard of your company yet, and why she treats every marketing decision like an intelligence problem.

Affiliate fraud investigator Ben Edelman joins Tye for an emergency episode breaking down the "Phia fiasco" that's rattling the affiliate marketing industry. Edelman, who's spent over 20 years catching affiliate fraud and has collaborated with Tye on forensic cases throughout his career, walks through what makes this case different from the usual stand-down violations that have hit shopping plugins like Honey and Microsoft Edge in the past.

Logan Dunn, Head of E-commerce at Wyze, ran 100 ads at once, cut traffic, and watched conversions go up. In this episode, he breaks down the counterintuitive paid media playbook behind Wyze's D2C growth — including why narrow targeting beats broad reach, how in-house creative unlocks faster iteration, and what most brands get wrong about scaling paid spend.

Nancy Harnett (Instantly) and Guy Yalif (Webflow) break down exactly how SaaS companies can grow their partner programs from single-digit contributions to 25% of pipeline and revenue. They cover partner ICP alignment, building a best-in-class partner experience, and why most programs stall before they ever scale.

Yoray Halevy joins Tye DeGrange to dig into why brands can't rely on Vogue-style coverage to build credibility — and what they should be doing instead. Yoray shares how his company creates authentic content that speaks to specific audiences, why last-click attribution misses the full picture, and how a paid media distribution mindset unlocks the true value of content. They also explore how AI is changing the content-to-distribution relationship, and why experimentation is the key to knowing what actually works.

Tom Rathbone joins Tye DeGrange to break down the real challenges of measuring CTV — from probabilistic Nielsen-era methods to where the industry stands today. Tom shares how TVScientific approaches measurement, why holdout testing is still underutilized by most brands, and how to build a test media plan that actually generates signal. Plus: why single-touch attribution fails in a multi-channel world and what it takes to move from vanity metrics to a long-term growth strategy.

For Mother's Day, Tye DeGrange hands the mic to his mom — Ginger DeGrange. Rodeo queen. Horse trainer. Summer camp founder. 36-year instructor who's taught 10,000+ students the art of horsemanship at Santa Rosa Junior College. This one's full of great stories: the OJ Simpson trail ride, a student who went on to dine with the Queen of England, competing at the Grand National, and the old reinsman who taught Ginger that quiet confidence beats loud energy every time. Plus real lessons on building confidence, earning trust, and leading with feel — on and off the horse.

Adam Dewer leads partnerships at Soci.ai — an AI-powered platform helping 800+ multi-location brands manage their local presence across social, listings, and reviews. In this episode, Tye and Adam get into the "zero platform" future of local marketing, where AI agents replace dashboards entirely — handling updates, review responses, and social posts through a simple SMS conversation with the business owner. They also dig into Adam's approach to building and measuring partnerships (he tracks just two numbers: net new revenue and expansion revenue) and why he'd take 10 incredible partners over 100 mediocre ones any day. What you'll learn: • Why "automagically" is the most dangerous word in AI marketing • How to identify your AI strength areas before trying to scale • Why agency partners are Adam's #1 channel play right now • The early signals that tell you a partner is going to scale • How to tee up a partner for success without making it transactional • What the "zero platform" future looks like for a local business owner Connect with Adam: linkedin.com/in/adamdewer | soci.ai

The first episode of SaaS Class tackles one of the most pressing questions in B2B marketing: how do you win when your buyers are finding answers on ChatGPT and Perplexity instead of Google? Tye sits down with Nancy Harnett (Head of Affiliate & Offsite AEO at HubSpot) and Guy Yalif (Webflow, formerly CEO of Intellimize) to break down what's actually working in the race to become the most visible brand in AI search. Nancy reveals how HubSpot generated 200K+ AI citations in 6 months and became the #1 most visible CRM across AI search. Guy shares how Webflow appears in 67% of the AI answers they care about — more than any other CMS including WordPress. What you'll learn: Why AEO is a natural evolution of great SEO, not a replacement The 4 pillars of AEO: Content, Technical, Authority & Measurement Why affiliate marketing is now the "new backlinks" — driving top and bottom-funnel AI impact How HubSpot used Reddit to go from ~100 to 146K AI citations in 7 months Content depth and localization: what actually moves the needle How to build a closed-loop measurement system to iterate fast

In this episode of Always Be Testing , host Tye DeGrange sits down with Lacie Thompson, an experienced agency leader, affiliate and partner marketing expert, and former operator at brands like Expedia and Blue Nile. After successfully scaling and exiting her agency, Lacie is now building Media Pack—an AI-powered platform designed to modernize and streamline how brands buy media. The conversation explores Lacie’s journey across brand-side, agency, and startup environments, highlighting how real-world frustrations led her to build solutions rooted in efficiency and transparency. From the broken workflows of flat-fee media buying to the lack of standardization across the industry, she unpacks the core problems that still slow teams down—and how AI is now making it possible to rethink the entire system. Beyond product and tech, the episode dives into the mindset behind building and scaling—from making decisions beyond data, to taking risks before feeling “ready,” to focusing on what truly creates differentiation. It’s a grounded and insightful look at what it takes to build in today’s AI-driven landscape while staying anchored in real customer needs.

Kyle White is the CEO & Founder of Refinery , an AI-powered affiliate technology solution. He brings a rare blend of experience across affiliate management (Overstock, Uber) , product development , and data-driven experimentation . He previously built tools at Impact, led global affiliate/ambassador programs at Uber, and now focuses on making affiliate programs smarter, more automated, and more profitable using AI. In this conversation, Kyle breaks down how AI is transforming affiliate marketing, how brands can adopt product‑led growth principles, and why relevancy—not payouts—is the strongest signal in partner matchmaking. He shares lessons from scaling Uber’s Ambassador Program from a 7% to an 80% activation rate, what most brands get wrong about affiliate recruitment, and how to design tests that actually produce meaningful insights. The episode also covers trends shaping the future: the threat of AI search bypassing the affiliate channel, the rise of pay‑per‑performance TV & audio partners, and why the next era will require filtering through more noise to find true value . Kyle also shares personal stories—building a company with a newborn, being featured in South Park, speaking Russian from missionary service in the Baltics, and his love of film photography.

In this episode, I sit down with Doug Bell — Fractional CMO, former leader at Automation Anywhere, LeanData, and Searchmetrics, top‑40 Substack writer, and co-host of Cannonball GTM — to unpack why modern go‑to‑market playbooks are breaking… and what the next generation of GTM looks like in an AI‑driven world. We go deep into the new patterns shaping high‑growth companies — and the uncomfortable truths most teams don’t want to face.

This episode explores the future of search and digital discoverability with an experienced SEO executive and growth advisor who has worked with high-growth marketplaces and SaaS companies. The conversation unpacks how the mechanics of online discovery are being reshaped by AI, shifting user behavior, and the declining reliability of traditional traffic channels. Together, we examine why visibility is no longer just a marketing function but a product and experience challenge — where trust, intent, and usefulness determine whether users discover and choose a product. The discussion also covers how teams should rethink SEO from a tactical ranking exercise into a strategic growth capability, what durable acquisition looks like today, and how organizations can adapt their playbooks to stay relevant as search continues to evolve.