Exited my eCommerce brand | Co-Founder at Workspace6.io & MediaSpace | Helping B2B startups understand brand psychology to boost sales

I started a B2C brand with $188 at 17. Grew it to 8-figures in sales and sold for 7-figures.     Then I co-founded a 1,000 member private community for 7-9 figure operators. Combined member revenue is over $13bn.     Now I run Redux Payments, where we use AI to recover failed payments and fix involuntary churn for B2C subscription businesses.     This is the problem I’m obsessed with solving. I write about what I find, what works, and what doesn’t. I’m always down to nerd out if you want to jam!     ————————————————————————      Now building:    Redux Payments (reduxpayments.com) - Stop losing revenue to failed payments. We slash involuntary churn and grow ARR by 8%+ for subscription businesses by using AI to recover failed payments.————————————————————————      You can reach me via email: philip (at) reduxpayments.com

Philip Pages - workspace6

100+ Stripe audits and I missed this. 

Last week I spoke with a former J.P. Morgan payments consultant.

A B2C subscription company hired him after Capital One stopped approving all their transactions.

In Stripe, their chargeback rate was under 1%. Nothing looked wrong.

But something was very wrong...

He showed me a hidden number in Stripe that was 8x higher than the chargeback rate.

"This is what issuers are actually looking at."
__

He called it the claim rate.

Chargebacks + Early Fraud Warnings.

Here's what I didn't know:

When a customer tells their bank "I don't recognize this charge," the bank files an Early Fraud Warning with Visa.

No chargeback, no notification. The merchant sees nothing.

But Visa counts it. The issuer counts it. And they use it to approve future transactions.

This company had 5–7% of transactions flagged as EFWs. 

They had no clue.
__

Capital One saw the real number (8% including EFWs, not the 1% in Stripe) and shut off all approvals.

The merchant crossed a threshold they didn't know existed, based on a metric they never checked.
__

HOW TO CHECK YOURS

Go to Stripe Dashboard → Payments Analytics → Disputes.

EFWs show up there and in payment logs.

It's visible but buried deep enough that in 100+ audits, I never looked until last week.

If I missed this, most B2C subscription companies are too.

Go check yours.

Philip Pages

CEO @ Redux Payments | Stop losing money to credit card failures | Exited Founder

We're signing $50M/year companies. We're competing with teams 10x our size. And we're winning.   

Redux Payments has a core team of 9. Engineers, machine learning, go-to-market. 

Everyone owns something end-to-end. No one is riding coattails.

I'd rather have 9 people who are absolute killers than 90 who are just filling seats. 

Here's what I've learned building this way:

1/ Small teams move faster. 
There's nowhere to hide. Every person either contributes or the whole thing breaks. That pressure is a feature, not a bug.

2/ Ownership isn’t for everyone. 
Some people thrive when they’re accountable for the result. Others don’t. We've had to let people go who couldn't keep up with the pace. It sucks, but carrying dead weight on a team this small will sink you.

3/ Care shows up in the details 
Last week, a customer hit a bug. Our engineers shipped a fix the same day. That’s what happens when everyone actually cares about the customer. You don't get that at a 200-person company where the ticket sits in a queue for 3 sprints.

I’ve watched incumbents with massive teams get slow, bureaucratic, and complacent. In this new era, a small, hungry team that moves fast will run circles around them every time.

9 killers > 90 coasters

Last month, an iOS app doing $2.1M/MRR on Stripe came to us struggling with trials not converting. What I found was almost too obvious.

Trial starts looked great. Meta was performing. But nothing was converting to paid.

Their team was blaming the product, the onboarding, the pricing.

Then I pulled the card-level data.
__

Stripe doesn't surface this by default. You have to pull it from the API and I only know to look because I've audited over 100 Stripe accounts.

Credit cards? Normal failure rates.
Debit cards? Fine.
Prepaid cards? 95% failure rate.

Prepaid cards were 10% of their card mix but 25% of all payment failures. One in four failed payments was a card that was never going to convert.

Their Head of Growth told me: "We had no idea prepaid was even a category we could filter by"
__

A user sees the ad. Downloads the app. Signs up for a free trial with Apple Cash or a prepaid card with $2 on it. The $0 auth goes through. Trial count goes up.

7 days later, the real charge hits. There's no money on the card. It fails. It will keep failing.

No retry strategy in the world is going to put money on an empty prepaid card.
__

HOW TO CHECK YOURS

Pull your charges from the Stripe API or Sigma. 

Filter by card.funding = prepaid. Compare trial-to-paid conversion against credit and debit.

If prepaid is a growing % of your trial mix, you have a problem you're not measuring.

The fix isn't complicated. 

Detect prepaid cards at trial signup. Block them, require a micro-charge to validate funds, or route them into a separate cohort so your metrics stay clean.

You might find that 25% of your "growth" isn't real.

I personally tested Headspace’s payment recovery flow and, honestly, expected way more.

I signed up for a 7-day trial ($12.99) using a card with only $1 to see how they'd handle failed payment retries.

June 26th: First charge failed.

They sent a "your last payment didn't go through" email and began retries:

June 27th, 8:03am
June 28th, 8:19am
June 29th, 9:03am

After three attempts in three days, they cancelled the subscription.

PROBLEMS:

Retries all around the same timeframe tell me they're fixed "dumb" retries.

If these were spread out with seemingly randomized date/time patterns, that would tell me that they're using an intelligent model that's adapting the retries. 

But there’s no intelligent modeling here, no adaptation for paydays, month-end cycles, or specific bank error codes. 

Compare to Hulu. When their transaction fails, they retried me 13X over a 50-day window. 

That's 7 weeks of recovery opportunities Headspace abandons after 72 hours.

*Backend Recovery Logic Rating* = D 

Looking at their recovery emails, I really liked their copy: “Is this really goodbye?”, but their strategy behind it is flawed. 

They trigger an email with every single retry. 

Every email slightly increases chances of voluntary churn, so their goal should be the opposite: Recover as many transactions as possible behind the scenes so the customer never even knows there was a problem. 

*Recovery Email Strategy Rating* = C+ 

Overall, for their web based failed payment recovery I'd give Headspace a C. 

It’s wild to see a company of this scale leaving so much money on the table. It proves that even for the giants backend recovery is rarely on the radar. 

There is a massive amount of "found" revenue waiting for any brand willing to move past basic retry logic and actually upgrade their stack to solve it.

Head of Growth: "failed payments were $419K. Stripe's recovering 28%."

CFO: "WTF. that’s not what my numbers say. what does that 28% include?"

Head of Growth: "subscriptions that failed and got retried. it's right there in the recovery dashboard."

CFO: "does it include trial-to-paid failures?"

Head of Growth: "...what do you mean?"

CFO: "go to your Stripe revenue recovery page. hover over the info icon next to 'failed payments.'"

Head of Growth: "it says... 'excludes failures on the first payment after a trial.'"

CFO: "meaning every free trial that tried to convert to paid and failed? not in that number."

Head of Growth: "how many trials failed last month?"

CFO: "I pulled it. $550K. just from trials."

Head of Growth: "that's more than our entire subscription failure volume."

CFO: "and zero of it shows in your recovery dashboard. zero of it hits your 28% rate."

Head of Growth: "so the real number isn't $412K..."

CFO: "it's close to a million. and your recovery rate on trials? under 8%."
Head of Growth: "8%?? how is that possible?"

CFO: "prepaid cards. users with $2 in their wallet signing up for a free week. 10% of your card mix. 25% of all failures. 95% never convert."

Head of Growth: "and Stripe retries them the same way as everything else?"
CFO: "four attempts. same generic logic. gives up. no one notices because it's not even on the dashboard."

Head of Growth: "so our churn graph looks stable..."

CFO: "while the trial funnel bleeds out before we ever collect a dollar."

Head of Growth: "how long has this been happening?"

CFO: "$1.4 million in invisible trial failures. last 3 months alone."

Head of Growth: "this is priority one."

CFO: "now we're talking."

Work from home is a scam and anyone who says they're more productive at home is lying. 

These are the 2 worst things 2 years of it cost me:

1) Mentally lost track of the competition. 

Seeing people show up day in and day out and grind toward a goal is a visual reminder of the effort it takes to actually build something real. 

If other people are doing it, you’ve gotta do even more. 
That pressure’s non-existent sitting alone at home. 


2) It killed serendipity. 

I never would’ve met my Redux Payments co-founder if we hadn’t been at the same Austin WeWork.

Years back, we were both working on opposite sides of the floor, doing our own thing. 
And every day like clockwork people would start heading home at 5pm.  
But he stuck around past 7, 8, 9:00, until the overhead motion detection lights went out and the cleaning crews started making their rounds. 
We didn't talk for weeks, maybe months. Until one day we started talking about AI or something. And just kept talking. 

We recognized the same thing in each other: the willingness to grind, work long hours, and get shit done.

That kind of thing doesn’t happen when you’re burning the midnight oil alone at home. 


It’s an anecdote of one, terrible I know. But I’ve never heard of a notable figure making it into the history books by doing anything in complete isolation. 

So yeah. I’m back at my old WeWork, working around people again. And a month in, it’s already a game changer.

Chuck Aikens

With over two decades of experience in content development and online marketing, I help brands harness the power of AI to create authentic, engaging content that resonates with their target audience and delivers on their brand promise.

At Tidal Wave Content, we've pioneered a unique method that seamlessly integrates AI throughout the content creation process. From initial insights and research to writing and optimization, our approach enhances human creativity and intelligence, resulting in impactful content at scale.

Our Content Manager Online platform enables marketing teams to produce high-quality, brand-aligned content efficiently.  This innovative tool, coupled with our strategic framework and proven process, empower brands to:

- Establish themselves as industry experts
- Optimize content for search visibility and social media engagement
- Create meaningful connections with their audience
- Drive brand awareness and business growth
- Increase traffic, generate more sales, and grow revenue.

I'm committed to equipping brand managers, marketing teams, and entrepreneurs with cutting-edge strategies, processes, and tools to accelerate their content marketing efforts. By responsibly leveraging AI, we create authentic conversations that truly impact our clients' businesses.

Let's connect if you want to explore how AI assistance and SEO best practices can revolutionize your content marketing strategy while maintaining your unique brand voice.

Founder

Tymoo is a digital marketing collective that empowers brand managers, marketing teams, and freelancers with the strategy, processes, and tools to accelerate growth.

🚨 If your AI agents don't have approval gates, you don't have automation - you have liability.

Since my last post about getting Open Claw running, I've gone from one AI assistant to a three-agent team handling real client work: strategy, content, follow-ups, the full operation. They named themselves The Crew. (Yes, they picked it. I just went with it.)

And the single most important thing I've learned so far? It wasn't about making them faster. 

It was figuring out where to slow them down.

After every client meeting, Morgan, my execution agent, extracts action items and proposes new work streams. My original design had her doing this autonomously. Faster. More impressive. Looks great in a demo.

But every work item is a financial commitment against limited resources. 

One overeager extraction, where the agent interpreted a casual conversation as a deliverable, could misallocate countless hours of capacity over the next few weeks.

So I added an approval gate. Charlie reviews Morgan's proposals before they ever reach me. Charlie reviews every commitment before it becomes real and flags me if something doesn't seem right.  This short delay and human-in-the-loop are negligible compared to the cost of getting it wrong.

Here's the pattern I keep running into:

1. AI amplifies output. 
2. Amplified output amplifies risk. 
3. Amplified risk demands governance.

I'm still figuring this out. But the more I build, the more I'm convinced that the teams who win with agents won't be the ones who remove the most guardrails; instead, it will be the ones who install the right ones to provide the needed governance.

I am documenting my journey as I go.  If you're building with agents and running into your own governance tradeoffs, I'd love to hear what you're learning. 

Drop a comment or DM me. These conversations are how we all get better at this.

AI Ops for Marketing · Founder, Tymoo · Where AI systems power real brand work

⁉️ Yesterday, I did an “AMA on AI tools” with the Forum for Naturals Summit at Natural Products Expo.  The conversation confirmed something I keep seeing across the industry.

The next phase of AI isn't about which tool to use. It's more about how we approach using AI.

Most people are in Phase 1 of AI: Prompt Engineering. Crafting individual instructions, iterating output, session by session. It works, but it has a hard ceiling. It doesn't scale, and it doesn't transfer.

The game has moved to Phase 2 (context engineering, where we feed AI info like our brand voice, personas, and product data, alongside your prompts) and Phase 3 (intent engineering, where we encode the specs and outcome you want into how agents and sub-agents do their work).

Those three phrases map to three AI identities:
-> Prompters use AI conversationally to draft, research, and brainstorm. They drive every interaction. This was Phase 1, and how we all started.
-> Builders connect tools, create reusable workflows, and build skill libraries. They're designing systems. This is Phase 2 and context changes the quality of outputs.
-> Directors orchestrate agents, write specs, and manage AI the way you'd manage a team. They define outcomes. This will be Phase 3 and requires software design-like thinking.

One phase isn't better than another. But you need to know which one you are today and which one you're growing toward.

What is your AI identity -- Prompter, Builder, or Director? That determines what you should pay attention to and where you should focus your learning.

Stop chasing tools. Match tools to jobs. The right tool in the wrong context wastes more time than no tool at all.

After your next really good AI session, where you feel proud of the deliverable, ask it to summarize what you did as reusable instructions. Save that file. Guess what? That's your first skill.

Your skill library may become the most important professional asset you own outside your professional network.

You don't need to master AI. You need to get intentional about how you use it.

One tool. One skill. One time a week. That's where you start.

#AI #NaturalProducts #CPG #Tymoo

Gave my 2nd talk at the Forum for Naturals Summit at Expo West.  This one was for the CMO Group.

After a little context, I asked everyone to pull up their brand's best-selling product page. 

Then we pretended like we were an AI shopping assistant looking for a question related to their product by a shopper.

For example: "What's a clean-label magnesium supplement for sleep that's under $30?"

Can you answer that question using only the text on the page? Not the images. Not the brand story. Just the structured text.

For most people in the room, the answer was no.

That's the gap. And it's not a technology problem. It's a product data issue.

AI shopping agents don't see your lifestyle photography. They read your product data. Your product data is your pitch to the machine.

I shared my 10-part framework (link to the deck in comments) with an emphasis on Hero Clarity, Key Facts Block, and FAQs.

Here's something you can do this week: work on your FAQs (question-answer pairs).

Layer 1: PDP FAQs. Five to seven product-specific questions on every product page. "What does this taste like?" "Can I take this while pregnant?" "How is this different from the capsule?" Each answer is 40–60 words, front-loaded with the direct answer. This is exactly what AI uses to recommend a product. 

Layer 2: Brand FAQ Page. One central page for shipping, returns, certifications, and sourcing. Keeps product pages focused. Gives AI a single authoritative source for brand-level answers.

Layer 3: Shopify Knowledge Base App. This is a Free app that feeds information directly to the AI shopper on ChatGPT. Not visible on your storefront. Just a backend feed for the LLMs. The key metric: unanswered questions are provided to you from OpenAI.

You don't need to rebuild your site. You need to organize what you already know.

AI agents are the new shelf space. Your product data is your pitch.  

Let's start by answering the questions your shopper and get them clicking on that "Add to Cart" button!

#ExpoWest2026 #ForumForNaturals #AISearch #NaturalProducts #CPG

taplio

The Battle

Philip Pages - workspace6

Chuck Aikens

Followers

Posts

Engagements

Likes

Performance ScorePhilip Pages - workspace6's LinkedIn Performance

Performance ScoreChuck Aikens's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Philip Pages - workspace6

Chuck Aikens

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Philip Pages - workspace6's LinkedIn Performance

Performance ScoreChuck Aikens's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScorePhilip Pages - workspace6's LinkedIn Performance