Episode 15 - Your Product Is Probably Ready Before You Think
Published Jun 23, 2026 · 30 min
Listen to this episode
Summary
What happens when a digital agency starts building and launching its own software products?
In Episode 15 of Not Brothers, Mark and Ryan talk about the lessons they’re learning from bringing Oodle-built products like Herald and Nebula to market. After years of helping clients promote finished products, they’re now dealing with the messier front end: product, price, placement, positioning, launch decisions, feature scope, and the temptation to keep polishing forever.
They dig into why products are often ready before the builders think they are, how “done is better than perfect” applies to SaaS launches, and why overbuilding can make a product harder to sell instead of easier.
They also talk about the product design trap of adding too many permissions, settings, toggles, and safeguards for problems that do not exist yet — plus why simple user experiences are much harder to create as products become more capable.
Later, the conversation turns to AI: how hallucinated research can quietly poison go-to-market work, why teams need better verification habits, and how skills, agents, orchestration, and daily briefs can create real productivity gains when used with discipline.
Chapters: 00:00 — No topic, so let’s talk about what we’re building 00:22 — Launching Herald as Oodle’s first product test bed 01:25 — Moving from promotion into product, price, and placement 03:11 — Nebula, feature creep, and products being ready before you think 05:44 — Shipping early enough to get real user feedback 07:05 — Done is better than perfect 08:35 — Day one starts when real users touch the thing 09:48 — Finding and removing features nobody actually uses 12:21 — Don’t add complication until it is necessary 13:03 — Permissions, theoretical problems, and soft safeguards 16:59 — Settings, toggles, and exposing complexity in the right place 18:24 — Why simple UX keeps getting harder 21:24 — Eating our own dog food while building products 22:18 — Using AI for market research without accepting bad data 24:10 — AI is not a Google search result 26:34 — Skills, repeatable workflows, and progressive disclosure 29:46 — Running multiple AI sessions without losing the plot 31:45 — Orchestrators, review agents, and long-running AI work 34:10 — Applying agent workflows beyond development 35:32 — Daily briefs, AI loops, and reclaiming focus
Chapters
- 00:00 — No topic, so let’s talk about what we’re building
- 00:22 — Launching Herald as Oodle’s first product test bed
- 01:25 — Moving from promotion into product, price, and placement
- 03:11 — Nebula, feature creep, and products being ready before you think
- 05:44 — Shipping early enough to get real user feedback
- 07:05 — Done is better than perfect
- 08:35 — Day one starts when real users touch the thing
- 09:48 — Finding and removing features nobody actually uses
- 12:21 — Don’t add complication until it is necessary
- 13:03 — Permissions, theoretical problems, and soft safeguards
- 16:59 — Settings, toggles, and exposing complexity in the right place
- 18:24 — Why simple UX keeps getting harder
- 21:24 — Eating our own dog food while building products
- 22:18 — Using AI for market research without accepting bad data
- 24:10 — AI is not a Google search result
- 26:34 — Skills, repeatable workflows, and progressive disclosure
- 29:46 — Running multiple AI sessions without losing the plot
- 31:45 — Orchestrators, review agents, and long-running AI work
- 34:10 — Applying agent workflows beyond development
- 35:32 — Daily briefs, AI loops, and reclaiming focus