Seller Sessions is the largest Amazon FBA and Private Label podcast for Advanced Amazon Sellers. It is the first of its kind, in terms of being raw, non nonsense and straight to the point. A lot of Amazon podcasts that came after has followed by example... Seller Sessions is published 4 times per week and often breaks new trends first in the industry.
Host Danny McMillan, is a world renowned public speaker and veteran Amazon Seller, Danny is also the co-founder of DATAbrill. DATAbrill manages Amazon PPC and advertising automation for 6, 7 & 8 figure Amazon brands.
Danny also works with Amazon in the UK to provide webinar content for their 3rd party sellers.
Each year he hosts Seller Sessions Live the annual conference for Amazon Sellers in the UK, bringing the worlds best speakers on the cutting edge of marketing on and off Amazon.
He is also the founder of SellerPoll, the official annual awards for Amazon Sellers and Brands.
Site
RSS
Apple
Data updated on 15/11/2026
Recent rankings
Latest chart positions across Apple Podcasts and Spotify rankings.
Shared links between episodes and podcasts
Links found in episode descriptions and other podcasts that share them.
700,000 Tokens to 6,000 with 1 File: Claude for Amazon Sellers
Friday, November 13, 2026 • Duration 34:57
Conversion Monthly is back with Danny McMillan, Dorian and Sim, and this month the theme is clear: for Amazon sellers using AI, infrastructure matters more than the latest model. The team look at what has really worked in their businesses since the last show, and what has burned time, tokens and money.
Sim explains how giving key staff the right tools doubled their off-Amazon marketplace sales, and how he gets the most out of a $20 assistant plan by having it run Codex for image generation. Dorian makes the case for building infrastructure and evaluations before solutions, and shows how he gives teams safe, limited access so they can build tools themselves.
Danny demonstrates a routing and index file that took one workspace from more than 700,000 tokens to under 6,000, and argues that the real value now lies in the systems that run your business, not in software that the next model can replace.
Key Topics
Off-Amazon growth - B&Q, Debenhams and other UK marketplaces, and what a short account suspension taught Sim about risk
Infrastructure before solutions - why bad input data, such as out of date landed costs, makes AI answers wrong
Building evals - how to check that the numbers your AI gives you are actually correct
Usage maxing - using a GrokBot assistant to spend a ChatGPT Codex allowance on image sets and TikTok slideshow ideas
Routing and index files - the Jake Van Clief folder system and how one file cuts token burn
The abstraction layer - why your business architecture is harder to replace than any software you launch
AI for teams - read-only data access, tracking outcomes, and cutting a thousand alerts down to three priorities
Multi-model setups - Claude desktop, separate cloud servers, open source orchestrators and Claude mods
Timestamps
00:02 - Welcome back: Danny, Dorian and Sim
00:19 - Sim: new UK marketplaces and doubling off-Amazon sales
02:10 - A short account suspension and spreading risk across seller accounts
04:23 - Dorian: infrastructure first, wrong profit numbers and building evals
06:42 - Simple wins: WhatsApp updates and why the model is not the point
07:32 - Sim's stack: Opus 5.5, Fable 5.2 and automated image generation
09:42 - Usage maxing with GrokBot, Codex and ElevenLabs voiceovers
12:09 - Token burn and the Jake Van Clief folder system
14:01 - Demo: from over 700,000 tokens to under 6,000 with a routing file
16:27 - The abstraction layer: why AI software goes out of date so fast
20:04 - Return on AI spend and what routing saves on each plan
21:58 - Dorian: input, instruction, output, and defining the workflow
26:28 - Why Sim uses ChatGPT for images alongside Claude
28:35 - Dorian: rolling out automation to larger teams
30:58 - Tracking results and focusing on three things, not a thousand
33:26 - Workflows compared: Claude desktop, parallel sessions and separate servers
36:41 - Open source orchestrators: pros, cons and setup
38:00 - Claude mods: a session status mod that shows when you are needed
40:17 - Markdown agents versus deterministic scripts
41:56 - Final words: API connections and the Solscope MCP
44:58 - Fable 5.2, a cloned Excel and whether software is solved
Key Takeaways
Start with infrastructure, not solutions - clean, current data and clear inputs decide whether AI output can be trusted.
Evaluate everything - check the input, check the output, and only automate once the results are consistently right.
One routing file can slash token use - point Claude to a routing and index file first so it reads only what the task needs.
Build the business, not disposable software - the architecture that runs your brand is the asset AI cannot copy.
Give teams controlled access - limited, read-only data lets your best people build their own tools safely.
Prioritise ruthlessly - automation should surface the three things that matter today, not a dashboard of a thousand.
Notable Quotes
"You give the key players in your business tools like the ones that are available now and great things happen." - Sim
"You don't start with building solutions, you start with building infrastructures." - Dorian
"It's not the models... It's basically everything ahead of it." - Dorian
"What it can't replace is your mechanism and your architecture that runs your business very efficiently." - Danny McMillan
Resources Mentioned
Claude - Opus 5.5, Fable 5.2, Claude desktop, Claude Code and Claude mods
ChatGPT Codex CLI - Sim's engine for main images and image stacks
GrokBot - personal assistant Sim uses to run Codex and manage usage
ElevenLabs v4 - voiceovers for TikTok slideshow content
Jake Van Clief folder system - routing and index files that cut token burn
n8n and Trigger.dev - automation platforms Dorian's teams build on
Hetzner - cloud servers Dorian uses to keep big projects separate
3T quote - open source orchestrator for multiple models, accounts and token tracking
Paperclip - open source tool for building teams of AI agents
Solscope - Amazon research tool with a strong MCP and API
Matthew Berman - YouTube creator who rebuilt Excel with Fable 5.2
Connect
Sim - LinkedIn, reach out for a Solscope discount code
Dorian - LinkedIn, drop him a message
Danny McMillan - danny@sellersessions.com
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan.
10 Hidden Costs of Automation for Amazon sellers
Thursday, October 15, 2026 • Duration 08:18
Danny McMillan reveals 10 hidden costs of automation for Amazon sellers: upkeep, stacked failures and lost judgment, plus a simple equation for true benefit.
Automation looks like a pure time saver for Amazon sellers, but Danny McMillan explains why the true cost of automating your business is higher than most brand owners price in. In this solo Conversion Monthly show, he shows how automations that work fine alone can fail together once they are stacked on top of each other.
You'll hear ten blind spots, from failures that arrive all at once during a product launch or account suspension, to the problem of remembering how a workflow you built three months ago actually works. The core idea: automation does not remove the work, it shifts it from execution to maintenance, and most people price the build but never the upkeep.
Danny finishes with a rule of thumb for what to ask at 1, 5, 10, 20, 50 and 100 automations, a refinement loop, and a hidden cost equation you can use to check what each automation really gives back.
Key Topics
Stacked failures - why several automations breaking at once during a launch or suspension costs far more than each one saves
Execution versus maintenance - how automation moves the work rather than removing it
The 10 blind spots - from forgotten setups to lost judgment and unchecked results
The scale test - the question to ask yourself at each size of automation estate
The refinement loop - build, run, find the exception, add the rule, test, monitor, repeat
The hidden cost equation - time saved minus the cost of keeping the system alive
The 10 Blind Spots
1. Failures can arrive together - with 45 automations, ten falling over on launch day makes the time each one saves matter less
2. Before you fix it, you have to remember it - did you leave yourself clear notes on why and how you built it?
3. You become the bottleneck - if only you understand the automations, every repair and every team question comes back to you
4. One broken link can create several problems - a stock report that feeds a reorder reminder that feeds another task spreads one failure down the chain
The Scale Test Key Takeaways The Hidden Cost Equation Notable Quotes Resources Mentioned
Smart Problem Solving for Amazon Sellers: Fix 100 Issues at the Root
Wednesday, October 7, 2026 • Duration 08:41
Adam can't make it this week, so Danny takes the Broadmatch show solo with a method for anyone buried under a long list of problems. It started when a friend who uses Claude heavily could not see the wood for the trees: a hundred issues, each needing its own fix, and no way to review them all.
Perplexity gave the approach a name, collaborative upstream problem solving. You'll hear how it combines systems thinking, affinity mapping, root cause analysis and collaborative validation, and why the person holding the domain knowledge has to stay in the loop. In this real example, it turned roughly 10 days of work into 2 to 3 hours.
Danny walks through a five-step process, a Claude example about forgotten decisions and stale documents, and a one-paragraph summary you can drop into Claude to build your own skill.
Key Topics
Collaborative upstream problem solving - fixing the shared conditions behind many problems instead of each problem one by one
The four-part method - systems thinking, affinity mapping, root cause analysis and collaborative validation
Leverage points - the smallest set of changes that makes many problems disappear
Why domain knowledge still wins - the human in the loop checks that the right problems were solved
A Claude example - forgotten decisions, contradictory outputs and stale documentation as one shared context problem
Turning the method into a skill - reverse engineer the transcript in Claude
Timestamps
00:00 - Danny goes solo, Adam out this week
00:25 - Perplexity names the method: collaborative upstream problem solving
00:46 - The problem: a friend buried in Claude issues
01:36 - How do you turn 100 problems into a few?
01:58 - The combined approaches: systems thinking first
Key Takeaways The Five Steps Notable Quotes Resources Mentioned Connect
Running Claude and Codex From Your Phone For Amazon Sellers
Friday, October 2, 2026 • Duration 31:21
Danny McMillan and Ritu Java cover AI remote control, token-saving skills and burnout resets so you can step away from the desk and keep projects moving.
Ritu Java is back for the monthly Go With The Flow show, fresh from Accelerate and a lot of conversations about how fast AI adoption is moving. Her problem will sound familiar: chat is faster than she is, so she has become the bottleneck on her own projects. This episode is about AI on the move, running your work from your phone so you can walk, hike or cook without losing control.
You'll hear how Claude remote control and Codex remote work in practice, what the catches are, and where Gemini Spark fits in for anything that lives in the Google ecosystem. Danny shares a plan-reading tip that cuts cognitive load, then the pair dig into where agency token spend really goes, why deterministic Python and SQL beat AI judgment on cost, and how to keep a stack of 90 skills clean as models change.
They finish on the human side: Ritu's two past burnouts, Danny's reset habits, and why controlling your attention matters more than keeping up with every new release.
Key Topics
AI on the move - running Claude and Codex sessions from your phone while away from the desk
Claude remote control vs Codex remote - one-time setup, per-project setup, and the "laptop must be open" catch
Gemini Spark - an always-on assistant for Gmail, Calendar, Tasks and BigQuery
Plan-reading tip - ask for a natural language plan on top and the technical plan underneath
Where tokens really burn - judgment work, scraping, images, video, browser work and MCPs versus leaner CLI and scripts
Skill hygiene - using /insights and /doctor, merging duplicates and building a skill map
Burnout and resets - breaks, walks, meditation and protecting your attention
Timestamps
00:00 - Danny welcomes Ritu back for the monthly show
00:14 - Ritu on Accelerate and the AI adoption shift
01:01 - The foundation problem: chat is faster than the human
Danny McMillan and Shubhash unpack the Fable 5.1 vs Astra debate, the six-stage AI harness, and why your folder structure is the real cost lever.
Shubhash joins from his new home in Dubai to dig into the two biggest AI releases of the fortnight, Fable 5.1 and Astra, dropped within 72 hours of each other. Instead of relitigating which model wins, he makes the case that "which model is better" is the wrong question entirely, and spends the episode showing why.
You'll hear the real cost story behind cached tokens, why a shiny 3D render says nothing about which model is actually smarter at your business's real work, and Shubhash's six-stage "harness" framework for judging any AI tool. Danny closes with a teaser: an enterprise AI consultant's folder-as-operating-system approach that could cut token burn without touching the model at all.
Key Topics
Fable 5.1 vs Astra - what a codebase-drift benchmark actually revealed, beyond the marketing screenshots
The cached-token price trap - why identical list prices don't mean identical bills
The six-stage harness - job, prompt, tools, model, failure handling, pass mark
Model-switching without the cost trap - avoiding preloaded tools when routing to third-party APIs
Claude Desktop as a single workspace - Danny's parallel-conversation, no-VS-Code setup
Folders as the AI operating system - a teased framework for scoping context with markdown files
Timestamps
[00:01] Shubhash joins from his new home - relocated to Dubai
[00:57] Agenda: Fable 5.1 and Astra, launched 72 hours apart
[01:24] Danny's pushback - "are they building anything?"
[01:50] Shubhash: "which model is better" is the wrong question
[02:06] Ellis's LinkedIn benchmark - Astra similar quality to Fable 5.1, lower cost
[03:13] Where Astra broke - drifting from existing codebase conventions on the hardest tasks
[04:29] Why comparison screenshots are confirmation bias - published by the model's own maker
[05:35] Shubhash's five-point checklist before adopting any new AI tool
[06:59] "A model is just an engine" - the car and harness analogy
Cutting AI Costs Without Cutting Output For Amazon Sellers
Wednesday, September 9, 2026 • Duration 29:59
Danny McMillan and Sim on model switching, OpenRouter cost hacks and whether AI spend for Amazon teams is actually worth it.
A quick solo show this week — Dorian's out sick, Matt's unavailable, so it's just Danny and Sim covering how to control AI spend without gutting output.
Sim walks through the real cost pressure of a 25-person team on Claude: five figures a year, uncapped Fable usage, and no way to see per-person burn. Danny counters with the setup he's built to fix exactly that — routing grunt work through OpenRouter and Kimi K3 without preloading tools, which took one job from £4-5 down to 35p. They land on a practical split: Fable 5 for planning, a cheaper model for the build, staged "cascade" plans to keep context windows under control. Sim also shares a genuinely wild same-day case study — a 100-video yoga app build for about £300 — and Danny pushes back on the whole framing: stop looking at AI as a cost line and start looking at what it's generating.
Key Topics Picking one AI provider and sticking with it - the cost of switching a trained team, and why Astra being better than Fable isn't reason enough to move Why teams actually burn tokens - heavy browser automation, parallel video work, and vibe-coders scaling far past their job description The OpenRouter cost hack - Kimi K3 without preloaded tools, cutting one job's cost by roughly 90% Cascade planning - staging big builds across multiple sessions to control context window burn Chinese model options - GLM 5.3 and LM Studio AI as free-to-cheap alternatives Grok as a low-friction accessory - personal automation for AI-hesitant team members ROI framing - weighing AI spend against what it actually produces, not just what it costs Timestamps [00:01] Quick solo show - Dorian ill, Matt unavailable, flight to Mallorca later [01:02] Wishing Dorian a speedy recovery, back next month [01:17] Kicking off: token costs, Astra's release [01:42] Why constant provider-switching burns team trust and training [02:10] History: Google/Gemini 18 months ago, Claude masterclass in December, full team switch [02:29] Sim's Opus 5 complaint - unclear responses, had to cross-check with Sonnet [02:58] Fable's cost problem - no way to cap runaway usage [03:28] Costs into five figures a year; splitting power users vs Chinese-model users [03:54] The NVIDIA DGX Spark purchase for local model runs [04:32] Danny's setup: Kimi in the terminal, Claude Desktop for daily work [04:59] The two real causes of heavy token burn: browser automation, parallel video editing [05:58] The "I've got ADHD" repo - controlling how Claude communicates [06:25] Claude Desktop's concise-response setting, plus widgets vs terminal mind-mapping [07:27] Auditing what each team member actually uses AI for before cutting costs [07:56] Why max-plan tool loading is free but third-party API tool loading isn't [08:26] "OpenClaude" - a terminal alias that loads Kimi without preloading tools [09:24] Sim's example: a warehouse dispatcher now vibe-coding time-tracking apps [09:54] The $20-to-Max "purgatory" middle ground [10:44] Danny's recommendation: Kimi K3 for the team, Claude Desktop kept for flexibility [11:12] Splitting by task: Fable 5 for planning, K3 for the build [11:40] Cascade plans - staging big jobs across sessions to control context [13:04] Real numbers: four parallel video edits, £4-5 down to 35p after removing tool preload [14:07] LM Studio AI and the free-credit-then-continue usage hack [14:43] GLM 5.3 - "Opus 5 good," and it speaks plain English [15:14] Grok/Grokbot - zero-friction agent automation from a phone [16:21] Building a review-flagging SOP from a LinkedIn post in one sitting [16:38] API scraper tip: Apify or Monid for cheap review scraping [19:00] Team breakdown: 5 on Max plans, ~18 on $20 plans [19:22] The frustration of no pooled token allocation across a team [20:12] The "Claude effect" - wait 21 days before reacting to a new release [21:08] Astra vs Fable, and why a full company switch is a bigger decision than it looks [22:04] Reframing: only 3-4 people actually need heavy-lifting models [23:37] Danny's pushback: what is the AI spend actually generating? [23:59] Case study: ~£300 to edit 100 yoga app videos with real footage and AI voiceover in a day [26:13] ROI framing: AI cost against revenue and headcount value [26:45] Team-wide Max plan cost: roughly £30k/year - "just a salary" [27:08] The local GPU reality check - a 90GB model locks the RAM, breaks other flows [28:03] RunPod and HyperStack as pay-by-hour alternatives to token metering [28:34] Wrap-up - Danny off to the airport Key Takeaways Constant model-switching costs more than it saves - retraining a team and rebuilding skills outweighs chasing the newest release; wait it out, then decide. Tool preloading is where the real API cost hides - stripping it out cut one job's cost by roughly 90% with no drop in output. Split the model by the task, not the team - planning on a stronger model, execution on a cheaper one, staged across sessions to control context burn. Audit usage before cutting spend - most token burn traces back to a handful of people doing genuinely heavy work, not the whole team. Cost only means something next to output - a £300 same-day build replacing two weeks of manual work reframes what "expensive" actually means. Notable Quotes "The last thing you want to be doing is getting people affiliated with the software, get them used to it, and then suddenly taking that away." - Sim "Once you took that out of the equation, I've got it down to about 35 pence for the same work." - Danny McMillan "Have you looked at what it's adding?" - Danny McMillan "It's not even remotely comparable." - Sim
Signal to Noise: AI Inbox Systems for Amazon Sellers
Friday, August 28, 2026 • Duration 31:36
Description: Danny McMillan and Ritu Java on AI-built email triage, pest control and client sentiment analysis for busy Amazon agencies and sellers.
Ritu Java returns after recovering from a serious ankle injury, and dives straight into the theme of the episode: managing cognitive load as AI accelerates faster than we can adjust to it. You'll hear how Danny and Ritu are each solving the same problem from different angles — building systems that manage the noise so the signal gets through.
Ritu walks through two Claude-built systems running on her Google ecosystem: a "pest control" routine that reads her inbox twice daily and files unsolicited pitches before she ever sees them, and a client sentiment analysis system that reads years of email history to flag which client relationships are improving or deteriorating. Danny shares his own zero-inbox protocol — a labelled morning triage that clears by 8:33am UK time — and both agree the real skill now is boundaries, not more automation.
Key Topics
AI acceleration and cognitive load - why keeping pace with AI now outstrips our own working speed
Pest control inbox routine - a Claude + Gemini API system that filters unsolicited email using "left brain, right brain" logic
Reply radar - surfacing the emails that genuinely need a response, out of tens of thousands
Zero inbox protocol - Danny's labelled, time-boxed morning triage system
Client sentiment analysis - mining historical email threads to score relationship health
Cascading plans - reviewing AI-generated plans in stages instead of all at once, to protect decision quality
Timestamps
[00:00] Ritu returns after an ankle injury, catches up on the last few months
[01:34] The acceleration of acceleration - falling behind AI's own pace
[02:29] Danny's all-in-Claude mandate for his team
[05:36] Cognitive load and the need for "a skill to manage the other skills"
[06:39] Danny on verification systems, sign-off protocol and decision fatigue
[07:19] Cascading plans: reviewing complex work in stages, not all at once
[08:05] Why a shorter, higher-quality working day beats a fixed nine-to-five
RAG for Amazon Sellers and Claude Model Switching | Seller Sessions
Tuesday, August 18, 2026 • Duration 35:26
Week three of the all-things-Claude series brings Shubhash back after some time away. He walks through why sellers should build a RAG (Retrieval Augmented Generation) system: business knowledge trapped in a founder's head and a stack of spreadsheets creates a bottleneck, and RAG breaks it by answering questions from your own files with sources attached, no hallucination, no SQL required.
Danny follows with a practical rundown of the model-switching setup he's spent the last week building: a Claude Code Router (CCR) experiment that didn't work out, and the eventual fix via OpenClaude, OpenRouter and DeepSeek to save on token spend for grunt work without disturbing his main Claude Desktop workflow. He also covers running a second email address on a different domain via the Google Workspace CLI, and gives an honest verdict on OpenMontage for AI-assisted video editing: impressive, but still not a replacement for editing domain experience.
Key Topics
RAG systems for Amazon sellers - chunking, embedding, indexing, retrieval and answer, built on Supabase/Postgres with pgvector
Data drift and guardrails - how to stop a RAG system guessing, and how to version-control its answers over time
Model switching for cost control - why and how, via OpenClaude, OpenRouter and DeepSeek
Harnesses vs models - why a model performs differently outside its native environment
OpenMontage - an AI video editing repo, and why domain experience still can't be replaced
Timestamps
00:00 - Danny opens week three, hands over to Shubhash
Cognitive Overload: AI Maxing, Product Development and the Hidden Tax
Wednesday, August 12, 2026 • Duration 51:37
Danny McMillan returns after his longest break in almost ten years, with Seller Sessions approaching its tenth anniversary and roughly 1,300 episodes. This is the pilot of a new monthly roundtable with Sim and Matt (Dorian returns next month), moving away from the conversion show format towards raw conversation.
You'll hear how Sim's team runs product development end to end with AI: keyword-scored idea validation, brand director sign-off, Claude-generated product concepts rendered through Codex, and a launch pipeline already booked out to 2027. Matt shares how Productpinion prioritises features from customer feedback, and why prioritisation is the most undervalued skill in the AI era.
The back half tackles the big theme: cognitive load. Danny breaks down verification fatigue, context switching and AI maxing, and why the scarce resource is no longer time but attention and decision quality.
Key Topics
AI-driven product development - from keyword scoring to Claude SVG concepts and Codex-generated product renders
Team structure at scale - how ideas route through brand directors to sourcing across UK and Philippines teams
Hiring in the AI era - why refusing to use AI is now a dealbreaker, and why gutting teams for AI is commercial suicide
Free local AI tools - Fluid Voice (Whisper Flow alternative) and Meetily (Granola alternative)
Cognitive load and verification fatigue - the hidden tax of moving from doer to overseer
Timestamps
00:00 - Danny returns: ten years of Seller Sessions, new pilot format
01:50 - Sim's update: ditching ClickUp for a bespoke operating system
03:55 - Matt's update: closing the research loop in Productpinion, Florence CRO brain, MCP
05:39 - Sim's product pipeline: AI keyword scoring, brand director approval, deep research
07:07 - AI product development: Claude concepts, Codex renders, 3-in-1 product mashups
09:07 - Packaging designed for the main image, and how far you can push it
Building Repeatables in Claude: Skills, CLI vs MCP and Token Discipline | Go With The Flow
Saturday, May 2, 2026 • Duration 46:02
Building Repeatables in Claude: Skills, CLI vs MCP and Token Discipline | Go With The Flow
Claude Skills, CLI vs MCP and Token Discipline with Ritu Java | Seller Sessions
SEO Description
Ritu Java and Danny McMillan on building agentic skills, choosing CLI over MCP, plan mode discipline and the short window to ship before token costs reset.
Episode Summary
Week 4 of the month, Go With The Flow, and Ritu Java is back from her travels. The world has shipped fast since the last episode: Codex 5.5, Claude 4.7, an Amazon Ads MCP and a fresh round of panic over the rumoured removal of Claude Code from the $20 plan (it was a 2% AB test, not a rollout). Ritu and Danny use the noise to make a sharper point: this is the moment to stop chasing models and start building repeatable systems on the platform you have already chosen.
Ritu walks through the three eras of PPC Ninja's automation stack. Apps Script bulk file generators three years ago, Netlify hosted UI apps last year, and now agentic skills that her team chats with in plain English to produce upload ready Amazon bulk files. The same shift applies to data: BigQuery accessed through the Google Cloud CLI rather than through MCP, because CLI is leaner on tokens and works better when the job is heavy on data rather than tool surface. Danny mirrors the move with his event-ops CLI for WordPress, WooCommerce, Stripe and FooEvents reconciliation, and his four tier ExtractFlow cascade (HTTP, headless, stealth, agentic) that bypasses the limits of any single browser tool.
The second half is a discipline talk. Plan mode every time. Push back on the first plan because Claude over engineers by default. 30% of your time on workflow scaffolding so the other 70% can be real building. The 21 day Claude rule: when a shiny new tool fires the dopamine, wait 21 days before refactoring around it. Left brain tasks (counting, SQL, deterministic logic) belong in scripts. Right brain tasks (judgment, creativity, hypotheses) belong in the model. Mix them inside a single skill. Skills are micro pieces of your workflow, not magic, and Claude can write them for you from an existing SOP.
Key Topics
The three eras of PPC Ninja automation: Apps Script, Netlify UI apps, agentic skills
CLI vs MCP: when to choose each and why CLI is more token efficient for data heavy work
Timestamps Key Takeaways Notable Quotes Resources Mentioned Hosts What's Next About Seller Sessions
Related Shows Based on Content Similarities
Discover shows related to Seller Sessions Amazon FBA and Private Label, based on actual content similarities. Explore podcasts with similar topics, themes, and formats, backed by real data.
5. The setup ages while you run the brand - a new supplier, marketplace or stock process can leave an automation running on assumptions that are no longer true
6. Every workaround adds something to remember - small fixes for a missing SKU or a special stock figure pile up across a dozen automations
7. You can automate the task and lose the judgment - a person questions an odd order, a workflow passes it straight through unless a check was built in
8. You still have to check the result - a workflow can run successfully and still produce a wrong report that drives a stock or spending decision
9. The repair is only part of the work - after a fix you may need to clear duplicate tasks, missed updates and incorrect records
10. The interruption can cost more than the task saved - the real cost includes the important work you could not do while maintaining the system
1 automation - can I keep this working?
5 automations - can I remember what each one does?
10 automations - can I see the failures and identify the owners?
20 automations - can I change one without breaking another?
50 automations - can I manage the testing, the dependencies and the recovery?
100 automations - am I running my business, or maintaining my automation estate?
Price the upkeep, not just the build - automation moves your work from execution to maintenance, it does not remove it.
Stacked failures hurt most when you are busiest - a launch or account suspension is exactly when several automations may need you at once.
Leave breadcrumbs - clear notes and maps of how automations connect make a repair possible three months later.
Keep human judgment in the loop - a workflow that runs is not the same as a workflow that is right, so build in checks and test the result.
Watch where your time goes - at scale, you can end up maintaining workflows instead of doing the work only a brand owner can do, like the product pipeline.
Time saved minus monitoring, fixes, testing, recovery and ongoing refinement equals the actual benefit.
"Automation does not remove the work. Effectively, it shifts some of the work from execution to maintenance." - Danny McMillan
"The time each one saves matters less when all need attention at once." - Danny McMillan
"The first version automates the process you thought you had. Maintenance adapts it once you know what you actually have." - Danny McMillan
"The trap is counting the hours each automation saves whilst ignoring the hours it takes to keep them alive." - Danny McMillan
Claude, or your favourite AI - paste this episode's transcript in and prompt against it to find the blind spots in your own setup
Maps of how your automations connect - a way to see how they integrate and overlap before you change one
Breadcrumb notes - short records of why each workflow was built the way it was
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan.
```
03:57 - Why domain knowledge matters: avoiding the wrong problems
05:03 - Clusters and markdown handover: 2 to 3 hours versus 10 days
05:18 - The five-step practical approach
06:22 - Claude example: forgotten decisions and stale documents
07:06 - Avoiding fixes that break other outputs
07:25 - A hypothesis, not a diagnosis
07:44 - Turning this into a Claude skill
07:55 - The method in one paragraph
08:23 - Wrap-up
Solve the conditions, not the problems - look for the shared causes generating a long list of issues.
Cluster before you fix - affinity mapping turns 100 items into a handful of meaningful groups.
Treat drivers as hypotheses - validate them with the person who holds the domain knowledge before acting.
Test one intervention first - make a small change and check whether several related problems improve.
AI executes, you think - the human in the loop and domain expertise decide whether the fixes are right.
1. Capture examples - the problems, plus the outcome the person wants
2. Cluster the symptoms - group problems by meaningful relationships using affinity mapping
3. Propose underlying drivers - recurring conditions that might explain each cluster, held as hypotheses
4. Validate together - check the groupings, exceptions, constraints and what is fixed
5. Test one intervention - a small change, then check whether related problems improve
"I'm not trying to solve a hundred problems faster. I'm trying to find the smallest set of changes that could make many of these problems disappear." - Danny McMillan
"AI isn't just your assistant. You're driving the process yourself. It's you who's doing the thinking." - Danny McMillan
"I solve the conditions creating the problems, not the problems themselves." - Danny McMillan
Perplexity - used to find a concise name for the thinking approach
Affinity mapping - the established technique for grouping many issues into meaningful clusters
Claude - used to organise the material and build the markdown handover files
Cron job checks - a scheduled way to keep documentation current and replace stale information
Questions on the method? Email Danny McMillan at danny@sellersessions.com.
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan.
03:15 - Claude remote control: how it works
04:43 - The terminal quirk for activating remote control
06:23 - Answering Claude from the car (and why not to)
06:43 - Danny spots the new remote control toggle in Claude Desktop
07:56 - The catches: laptop must be open, set up per project
08:22 - Codex remote: desktop and mobile mirror each other
09:21 - Sharing one GitHub between Claude and Codex
10:02 - The protocol: plan at the desk, follow up on the move
11:05 - Danny's plan tip: reject early plans, ask for a plain-English top half
12:40 - Gemini Spark for the Google ecosystem
13:58 - How steep is the Claude to Codex learning curve?
15:06 - Sonnet 5.5 efficiency versus Opus 5.5 cost
15:49 - Running 86 to 90 skills and watching token usage
16:41 - Left brain versus right brain: where tokens skyrocket
18:15 - MCP versus CLI token cost
19:10 - Running /insights and /doctor over your skills
21:38 - Maintenance: old scaffolding can slow newer models
22:23 - What Ritu actually does on the road
23:18 - Notepads and breadcrumbs: Apps Script on Gmail
24:14 - Client sentiment analysis from email
25:10 - Staying fresh and Ritu's two burnouts
26:52 - Danny's resets: walks, meditation and breaks
28:31 - Wrap-up: the luxury of experimentation
30:15 - Danny: control your time and attention
31:12 - Where to find Ritu
Plan at the desk, steer on the move - do the heavy work in one sitting, then answer questions and corrections from your phone.
Know the catches before you rely on remote control - the laptop has to stay open, and Claude needs setup project by project while Codex is enable-once.
Ask for a plain-English plan - a natural language top half with the technical plan underneath makes plan review far easier on your brain.
Push repeatable work into scripts - Python and SQL cost nothing once built, while judgment, scraping, images, video and browser work eat tokens.
Maintain your skills - merge duplicates, add a skill map, and retire guard rails that newer models no longer need.
Protect your attention - pick a handful of things to follow, set boundaries and take real breaks before burnout does it for you.
"Chat is faster than me, and I'm seeing that I am becoming the bottleneck." - Ritu Java
"We thought AI could make us freer than before. No. I feel like a slave. But it's fun." - Ritu Java
"Scaffolding or guard rails you put in eight months ago can actually slow the model down." - Danny McMillan
"Control your time and attention, because AI is going to drop something new every single day." - Danny McMillan
Claude remote control - lets you answer a running Claude Code session from the Claude app on your phone
Codex remote (ChatGPT Codex) - mirrors desktop and mobile, enabled once for every project, Mac only
Gemini Spark - always-on assistant in enterprise Google Workspace for Gmail, Calendar, Tasks and BigQuery
/insights and /doctor - built-in commands for reviewing and tidying your skills
Google Apps Script - used to mine email for notes, ideas and client sentiment
Notion and Freeform - Danny's breadcrumb tools outside Claude
Ritu Java - find her on LinkedIn under her full name
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan.
[08:28] The real price difference - cached token reads, 25c vs $1 per million
[09:35] Season ticket vs pay-on-the-gate pricing
[09:54] How few sellers actually examine their AI bill
[11:41] Avoiding the third-party API cost trap - tool preloading
[12:20] Real example - a $4.50 job cut to 35p once preloading was stripped out
[12:40] Where Astra genuinely wins - 3D renders, exploded views, web design
[13:46] Splitting by job type - Astra on design, Claude on knowledge work and recovery
[14:30] The six-stage harness framework introduced
[17:12] "The model is the sex and sizzle" - Danny's framing
[18:21] Shubhash's VS Code setup - Claude and Codex extensions handing off work
[19:52] Danny's Claude Desktop-only workflow, no VS Code
[20:22] Finder and Spotlight optimisation, Kimi K3 kept to the terminal only
[23:39] Widgets over walls of text - "like a kid's colouring book"
[24:24] Natural-language plan on top, technical plan underneath, on request
[24:45] Freeform and stylus for complex problem-solving
[25:50] Raycast as a free Spotlight replacement
[27:08] Why "you don't need instruction files anymore" ignores who's paying for the tokens
[27:56] Teaser - an enterprise AI consultant's file-structure system
[29:15] The folder-as-AI-operating-system concept
[31:27] Three-layer breakdown - the map, context rules, work tools
"Which model is better" is the wrong question - the harness around the model decides the outcome far more than the model itself.
List prices hide the real cost - cached-token rates can be 4x apart even when headline pricing looks identical.
Split tasks by strength, not loyalty - Astra ahead on visual and web design work, Claude ahead on knowledge work, long documents and error recovery.
Tool preloading is the hidden API cost - stripping it out cut one job from $4.50 to 35p.
Your folder structure is your harness - scoping Claude to only the context a task needs cuts token waste before any model swap is needed.
"A model is just an engine. If you haven't built the car, the chassis, the tyres, everything else around it, a faster engine isn't getting you anywhere." - Shubhash "You don't pay your goalkeeper up front just because he's the best athlete in the squad." - Shubhash "The model is the sex and sizzle, not the bit underneath it." - Danny McMillan "Before you chase the next model, write down only what you know... get that right and we are good." - Shubhash
Fable 5.1 - the latest release under discussion, benchmarked against Astra on real codebase tasks
Astra (GPT-6) - OpenAI's release, strong on visual/web design work, weaker on holding existing coding conventions
Raycast - free Spotlight replacement Shubhash uses for local search
Claude Desktop - Danny's single-workspace setup: parallel conversations, in-app browser and file viewer
Freeform (iPad + stylus) - Danny's tool for mapping out complex problems before handing them to Claude
Shubhash - Not a Square, now based in Dubai (relocated from London)
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan. Ritu returns next week for Go With The Flow; Shubhash is back next month.
Resources Mentioned OpenRouter - model-switching gateway, used here to route to Kimi K3 without preloaded tools Kimi K3 - the cost-efficient model for day-to-day build work GLM 5.3 - Chinese model tried via LM Studio AI, described as "Opus 5 good" LM Studio AI - harness for running Chinese models, including a free-credit usage workaround Grok / Grokbot - low-friction personal AI agent with built-in scheduling Apify / Monid - API scrapers used for cheap-at-scale review data pulls NVIDIA DGX Spark - local hardware for running in-house model flows RunPod / HyperStack - pay-by-hour GPU rental as an alternative to token-based pricing Connect Sim - Amazon seller and co-host
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan.
[10:03] Ritu's two access routes into Google: MCP connector vs CLI
[10:56] Claude routines explained - the "cron job" for your inbox
[11:55] The pest control system: left brain keyword matching, right brain Gemini analysis
[16:04] Reply radar - catching the emails that were missed
[18:07] Danny's zero inbox protocol: labels, triage timing, draft handling
[20:39] The "sixth follow-up email" rant - boundaries with cold outreach
[27:17] Danny on spotting relationship drift before the client raises it
[30:12] Wrap-up: signal to noise as the theme of the episode
AI is accelerating faster than we can adapt - the gap between AI's pace and our own decision-making speed is the new operating reality.
Fuzzy logic beats static rules - Ritu's pest control system combines keyword detection with AI judgement, catching what a plain Gmail filter would miss.
A time-boxed triage beats an always-on inbox - Danny's zero inbox protocol proves structure, not more tools, is what protects focus.
Sentiment analysis surfaces relationship drift early - mining historical email tone gives an objective signal alongside human account management.
Cascading plans protect decision quality - reviewing AI output in stages, not all at once, avoids the fatigue that degrades judgement.
"We're slower than the AI now in terms of how fast things are moving." - Ritu Java "AI is not an enabler for us to think on our behalf. It's our job to verify it." - Danny McMillan "In order for us to be efficient with our productivity, we need something to manage us." - Ritu Java "Without perspective and perception, empathy can't live." - Danny McMillan
Claude routines - scheduled, cron-style Claude tasks used to run the pest control and reply radar systems
Gemini API - powers the "right brain" analysis layer in Ritu's pest control system
Google Workspace MCP connector - built-in Claude access to Gmail and Drive
CLASP - CLI tool for Google Apps Script development
BigQuery CLI - command-line access to agency ad, organic and rank-tracking data
Ritu Java - AI for E-commerce newsletter; speaking at the Amazon booth this month
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan. Adam Heist returns next week for the Broadmatch Show.
06:54 - Danny on data drift: the chicken-and-egg problem of trusting Claude to catch its own drift
08:18 - Core of it simplified: chunk, guardrails, plain-English queries via Slack or Claude
09:15 - Building on the same Supabase project from earlier sessions, switching on pgvector
10:12 - Guardrail rule: don't guess, cite sources, flag when there's no answer
10:41 - Version control: feeding user feedback back into the system
11:27 - Danny's segment: model switching, why and how
13:22 - The CCR experiment: a gateway/pass-through that didn't preserve context or dependencies
15:14 - Why OpenClaude was the better route: terminal-based, dependencies not preloaded
16:44 - DeepSeek vs Fable 5: close, not equal, but roughly 90% cheaper for grunt work
17:40 - Harnesses matter as much as the model itself
20:38 - Budget-capping your OpenRouter API key (and the horror stories of not doing so)
22:52 - Why Danny moved off Claude in Chrome for email rebuilds - context window visibility problems
23:37 - Solving it with dual Google Workspace access via CLI (Seller Sessions + DataBrill)
25:17 - OpenMontage: plug in the repo, but expect to still do heavy lifting
27:36 - The aggregator analogy: why "done for you" video/design without domain experience is a scam
30:46 - Round-up: three steps to start a RAG system yourself
34:56 - Where to reach Shubhash
RAG breaks the founder bottleneck - your spreadsheets and documents become queryable with sources attached, cutting out the "ask the founder" loop.
RAG is simpler than it sounds - five steps (chunk, embed, index, retrieve, answer), no model training, no GPUs, built on Postgres/Supabase you likely already have.
Guardrails beat guessing - if there's no answer in the data, say so rather than fabricate one.
Model switching only pays off with the right setup - a terminal-based OpenClaude + OpenRouter + DeepSeek combo saved roughly 90% on grunt-work token spend.
AI video editing still needs domain experience - OpenMontage is a genuine leap forward, but a done-for-you button-press is the same mistake the Amazon aggregators made.
"The business knowledge lives in the founder's head and a stack of spreadsheets that only they can navigate." - Shubhash
"It's easier or better to say, I don't have enough data to make a conclusion to this, than to guess." - Shubhash
"Harnesses are just as important, to a point, as the model itself." - Danny McMillan
"If you said to a video editor, I'll just press a button and it makes me a video - they'd say that sounds like a scam to me." - Danny McMillan
Supabase / Postgres + pgvector - the database layer for storing and searching RAG "cards" by meaning and keyword
OpenRouter - model-switching gateway used to route to DeepSeek; supports API key budget caps
DeepSeek (v4 Pro) - the model Danny settled on for cost-saving grunt work, roughly 90% cheaper than Fable 5
OpenClaude - terminal-based alternative to a full CCR gateway, dependencies loaded on demand
Google Workspace CLI - used to run two email accounts through Claude without a browser context-window problem
OpenMontage - GitHub repo for AI-assisted video editing, works alongside a Claude video/Remotion flow
Shubhash - Not a Square Email: shubhash [at] notasquare.io
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan. Week four returns with Ritu for Go With The Flow.
10:46 - Team workflow: brand directors owning P&L, sourcing handoffs
14:21 - Danny on gutting teams for AI: who maintains the machines?
15:56 - The hiring line: refuse to use AI, you haven't got a job
17:48 - Claude across every department: projects, Claude Code vs Cowork
20:10 - Free local tools: Fluid Voice for dictation, Meetily for meeting notes
21:57 - Marketplace arbitrage: moving proven products between Amazon marketplaces
23:21 - Matt on signal to noise: wasting tokens instead of wasting time
24:57 - Time blocking and prioritising features by customer impact
31:46 - Asking the right question: the Claude Science deep-dive example
36:41 - AI maxing, context switching and high-stakes decision quality
42:07 - Claude telling you to go to sleep
45:16 - Danny's framework: reject the first plan, decision sprints, deliberate decompression
50:47 - Where to reach Sim and Matt
AI has made product development fun again - unique product concepts generated with Claude and Codex, feeding a pipeline mapped to 2027.
Augment, don't replace - if someone was worth hiring, AI should multiply their output, not justify cutting them.
Prioritisation is the undervalued AI skill - just because you can do everything doesn't mean you should.
Verification fatigue is real - build in decision sprints and deliberate decompression, and reject Claude's first plan on sight.
The scarce resource is attention, not time - your night schedule and recovery feed the next day's output.
"AI is enabling the boring to get released and the fun stuff to happen." - Sim
"Instead of people wasting time, now they're just wasting tokens." - Matt
"Prompts don't matter, but asking the right question unlocks everything." - Danny McMillan
"AI doesn't just speed you up. It puts you on permanent oversight duty, and the cost of that duty is your attention and your judgment, not time." - Danny McMillan
Fluid Voice - free, open source local dictation with on-device models; a Whisper Flow alternative
Meetily - free, open source meeting summariser that runs privately on your machine; a Granola alternative
Claude / Claude Code - the AI platform used across both Danny's and Sim's teams
Codex - used alongside Claude to generate product concept images
Productpinion - Matt's shopper testing platform, now with MCP support and draft polls
Sim - on LinkedIn (genuine reach-outs answered) Matt - on LinkedIn or via productpinion.com
Seller Sessions is the leading podcast for advanced Amazon sellers, hosted by Danny McMillan. Dorian returns next month.
Token economics, the rumoured $20 plan change and why it was a 2% AB test
The short window before subsidised tokens get repriced
Plan mode discipline and the "push back on plan one" rule
Danny's 30 / 70 framework: workflow scaffolding vs building
The 21 day Claude rule for resisting tool churn
Left brain vs right brain task design inside a single skill
The PPC Ninja "5 Whys" skill: deterministic SQL plus non deterministic hypotheses
[40:53] Beginners: let Claude write the skill from your SOP, use the interview pattern
[42:39] Skill cascading: URL to research to article to LinkedIn post to tweets to slides
[44:42] Mixing deterministic and non deterministic inside a single skill
[45:39] Wrap up, signal to noise, who is it for
Pick a platform and stop chasing models. A new model ships every week. Time spent benchmarking is time not building. Double down on Claude (or whichever you chose), use the 21 day rule, and let the ecosystem catch up to the shiny thing in your feed.
CLI for heavy work, MCP for soft touch. MCP loads tools and skills into context and burns tokens. CLI uses programs already on your machine. For data heavy jobs (BigQuery, AWS, WordPress at scale), CLI wins. For light cross app workflows, MCP is fine.
Build repeatables, not one offs. Subsidised tokens will not last. The 100planreportedlycostsAnthropic100planreportedlycostsAnthropic5,000 to serve. Spend the window building scaffolding that compounds, not 14 hour vibe coding runs.
Plan mode every time, then push back. Claude over engineers by default. Generate the plan, then say "you have over engineered this, although I want it elegant, go back and review." Plan two is the one you start from.
30% on workflow, 70% on building. Each new dependency, MCP, skill or repo you add to your workflow compounds across every future project. Stop building only the apps. Build the things that build the apps.
Left brain in scripts, right brain in the model. Counting, SQL, deterministic logic belongs in Python the moment you can offload it. Save the model for hypotheses, judgment and creativity. The PPC Ninja 5 Whys skill mixes both inside one flow.
Skills are micro pieces, not magic. Take an SOP, ask Claude to interview you with decision panels, and let it write the skill. Then cascade skills together: URL to research to long form article to LinkedIn post to tweets to slide deck.
"Instead of doing one offs, it is time to build repeatables. The more people can learn that skill now, the better it will be, because a year from now you may not have access to the same tokens." Ritu Java
"If you see something and it looks sexy and it has sex and sizzle and your dopamine is screaming to go after it, wait 21 days. Either Claude will have it, or someone will have a repo, and you can combine it." Danny McMillan
"Always use plan mode. Never accept plan number one. Tell Claude: you have over engineered this, although I want it elegant, go back and review. Then start from plan two." Danny McMillan
"I build the things that build things. I build the scaffolding the team needs so they can build on top of it." Danny McMillan
"Spend 30% of your time on your workflow and 70% building. The 30% compounds across every project." Danny McMillan
"If we just hand six months of ad, organic, ranking and SQP data to Claude with no structure, it is going to mess up. It will give you an 80/20 you are not satisfied with, because it is not equipped to handle that volume without scaffolding." Ritu Java
"WordPress is horrible to work with through MCP. It falls over all the time. CLI can be amazing for certain things." Danny McMillan
PPC Ninja : Ritu's Amazon PPC software and agency, base for the BigQuery + CLI stack discussed
Claude Code : Anthropic's CLI for Claude, the primary surface used in the episode
Ritu Java : CEO and co founder of PPC Ninja, Amazon PPC software and agency. Specialises in automation, BigQuery pipelines and agentic workflow design.
Next week: Ritu and Danny pick up routines and the new Claude scheduler.
In 8 days: Seller Sessions Live 2026 in London on 9 May. Last week to lock in any final discounts.
Seller Sessions is the leading podcast for serious Amazon sellers, hosted by Danny McMillan since 2017. Go With The Flow is the weekly automation strand where Danny and Ritu Java work through agentic flows, MCPs, CLIs and skills, in real time, on the same stack their teams ship every week.
Episode published: 1 May 2026 Series: Go With The Flow (Week 4 of the month) Keywords: claude skills, claude code, cli vs mcp, mcp model context protocol, claude 4.7, codex 5.5, amazon ppc automation, bigquery cli, agentic workflows, plan mode, token optimisation, claude.md, agents.md, ppc ninja, ritu java, seller sessions podcast, go with the flow