Podcast Summary

Always Be Testing

Episode #119: Why 70% of Affiliate Programs Fail Before They Launch — and How AI Can Fix That
Host: Tye DeGrange
Guest: Kyle White (CEO & Founder, Refinery)
Date: March 10, 2026

Episode Overview

In this episode, Tye DeGrange sits down with Kyle White, a veteran affiliate and performance marketer, and currently CEO of Refinery—a tech solution using AI to help manage and optimize affiliate programs. Their conversation centers on why the majority of affiliate programs struggle or fail before launch, key lessons Kyle gleaned from high-profile roles (Overstock, Uber, Impact), how Product-Led Growth (PLG) reshapes affiliate strategy, and how AI can automate, refine, and scale successful partnerships in B2B SaaS.

Key Discussion Points & Insights

1. Origins & Inspiration Behind Refinery (01:37)

• Pain Point: Kyle observed, through consulting and direct management roles, that affiliate programs were slow to adopt new technology, causing inefficiency and wasted potential.
• “There has to be a better way… why isn’t this automated? So, let’s start building a solution…to unlock the power of AI, saving a lot of time and uncovering opportunities for affiliate managers hiding within their data.” (A, 02:10)

2. Product-Led Growth in Affiliate Marketing (03:49, 08:46)

• Uber Ambassadors Case Study:

  • Early influencer/ambassador affiliate program faced a “7% join rate,” but applying product and onboarding optimizations (brevity, reducing jargon, direct value) increased this to “like an 80% join rate.” (A, 04:26)
  • “Focusing on improving time to value—get a code, share it, get paid—was a game changer.” (A, 05:31)

• Defining PLG for Partners:

  • “PLG is all about retention via the product itself… using activation, engagement, resurrection—creating ongoing win-win loops where the partner sees value in promoting your brand and sticks around.” (A, 08:46)

• Actionable: Providing partners with useful, understandable metrics and insights (like LTV scoring) encourages ongoing engagement and loyalty.

3. The Science of Matchmaking & Relevancy (11:43, 12:25)

• Early Mistakes:

  • “At first we just followed the money—top performers from other programs. That did not go well.” (A, 12:44)

• Key Learning:

  • Weighted “relevancy” (vertical, audience fit, partner size) over payout: “Much better engagement and match rate.” (A, 13:01)
  • “Focus on contextual partners… know your audience, do your homework, use AI to reverse engineer searches and reach people where real synergy exists.” (A, 13:25)

4. Experimentation: Testing for Success (14:43)

• Most Underutilized, Powerful Experiment:
  • Tiered commission performance bonuses. “Give affiliates a clear target and a carrot—hit 5 sales, get a $10,000 bonus. It creates motivation and stretches them.” (A, 14:46)
• Proper Experimentation Tips:
  • Set historical baselines, watch for seasonality, isolate variables, compare apples-to-apples. “Be careful not to compare what isn’t apples-to-apples…it’s critical for meaningful tests.” (A, 16:58)

5. Biggest Lesson from 2025 (18:41)

• “Fail fast and iterate faster…do it, fail at it fast, but make sure you learn and grow from it quicker or you’ll just be stuck in the mud.” (A, 18:47)

6. AI’s Real Value for Affiliate Managers (Refinery’s Secret Sauce) (20:20)

• AI-Powered Partner Discovery:
  • “We don’t have large lists. We have a refined 5-10 people a week who will be your next top partners, found via AI using reverse-engineered search.” (A, 20:32)
• AI-Driven Insights:
  • “The wow moment…an insight that would take hours of spreadsheets is now instant. That’s $1-2K in savings plus more up-and-to-the-right performance.” (A, 21:10)

7. Industry Dynamics: The Threat & Promise of AI and Search (21:57, 25:16)

• Cautionary Take:

  • “If we [affiliate marketers] don’t get a seat at the table…AI search may bypass affiliate and either become the affiliate or cut us out. We need to keep pressing…advocate for content creators to get paid.” (A, 22:05 & 23:18)
  • Suggests potential for profit-sharing models between affiliates and AI platforms.

• AI Hot Take:

  • “Quantity of shiny objects will increase, but true quality will be much harder to find…we might be lost for a minute but if we keep alert, the end-of-bell-curve value will stick around.” (A, 25:16)

8. Partner Types: What’s Rising? (28:22)

• Excitement for Pay-Per-Performance TV & Audio:
  • “Full-funnel targeting—paper performance TV, audio hook…top-performing affiliates have some unique targeting, and these channels are big if your brand qualifies.” (A, 28:50)
  • Technology solutions for post-purchase targeting also highlighted.
  • “Glad to see the industry moving beyond last click attribution.” (A, 29:36)

9. Personal Insights & Fun Facts (30:09)

• Recent favorite purchase: Oofos slides (“feels like you’re walking on gushy clouds”).
• Hobbies/fun facts: Speaks Russian, enjoys film photography, Simpsons superfan, and was once featured in a live-action South Park commercial. (31:04)

Memorable Quotes

• On AI’s Real Impact:
  “This insight that it provided for me would have taken me multiple spreadsheets and hours… and now it’s instant. That’s the moment where you realize—wow, this thing is powerful.” (A, 21:10)

• On Product-Led Value:
  “If you create a nice virtuous loop, everyone wins. That’s when real success happens.” (A, 11:31)

• On Experimentation:
  “Fail fast and iterate faster… or you’ll be stuck in the mud.” (A, 18:47)

• On the Industry’s Future:
  “If we don’t get a seat at the table and work it correctly, AI search may bypass affiliate and… take a paycheck for itself.” (A, 22:05)

• Hot Take on Partner Types:
  “Full funnel targeting. TV, audio, tech partners—if your brand qualifies, these are the quality, top performers of tomorrow.” (A, 28:50)

Timestamps for Key Segments

• Introduction & Refinery’s Origin – 00:54–02:42
• Product-Led Growth & Uber Story – 03:49–05:43
• Minimizing Friction for New Partners – 05:43–07:03
• PLG Definitions for Affiliate – 08:46–10:54
• Data-Driven Matchmaking – 11:43–13:58
• Experimentation & Testing – 14:43–18:05
• 2025 Lessons & Learning – 18:36–19:26
• AI’s Real-World Use Cases – 20:20–21:38
• AI/Affiliate Industry Dynamics – 21:57–26:22
• Emerging Affiliate Types – 28:22–29:49
• Personal Q&A & Closing – 30:09–32:15

Tone & Style

Candid, collaborative, and rooted in field experience. The conversation flows as a mix of storytelling, hands-on tactics, and honest reflection. Both speakers balance optimism about technological advancement with measured caution about industry disruption, always centering on actionable value and relationship-building.

For anyone working in affiliate, partnerships, or SaaS growth, this episode offers both high-level strategy and granular, testable tactics—all shaped by the voices of practitioner-leaders.