E-commerce Founder // Built multiple 7-figure brands // Writing about backend systems that let you scale

Joined February 2024
68 Photos and videos
Pinned Tweet
Sold my ecom brand for 7-figures last year. Building the next one now. The last one I built in silence for 8 years. Huge mistake. Doing this one in public (here on X).
13
1
43
3,814
The best ad copy is already sitting in the words customers used when they were happy. A literal goldmine. Use Claude to mine it.
1
46
Your competitor 1 star reviews are the cheapest market research in the known universe.
39
I use Claude to pull ad angles out of my own reviews. I feed it the last few hundred customer reviews and ask for the recurring phrases people use to describe why they kept the product. Those become the hooks for the next round of creative.
2
1
86
I spent years learning what a buyer actually checks before they'll pay a real multiple for a brand. There are levels to everything.
1
3
160
In 15 days a parcel to Europe from China will be increased by €3,- per item. Means a potential increase of €9,- if you sell 3 items in once. Nobody is prepared for that. Your margins will be eaten. All EU dropshippers going to panic, believe me. We are testing out solutions right now and results are very positive
12
34
5,332
You paid for the mold. That doesn't mean you own it. Plenty of founders fund the tooling for a custom product and never get it in writing. The factory holds the mold, and the day you try to move suppliers, you find out the mold moves with them.
1
357
Pre-shipment inspection on every container. Supplier contracts that put defects on them. Quality controlled at the source instead of apologized for at the inbox. The small stuff.
1
125
An ad audience is rented. You stop paying, it vanishes. An email and SMS list is owned. It costs nothing to message and it doesn't disappear when the ad account does. Huge lesson.
3
145
The €70K supplier disaster started because of shipping. Under the terms we had, the choice was his and the cost of failure was mine. He picked the cheapest option. Now the terms are explicit. The exact line, who's liable at which point, written in.
1
153
A brand sells on how boring its paperwork looks under a microscope. Build the unglamorous parts early. They set the multiple.
72
The smart path to a real brand starts by selling someone else's product first. You test demand with a generic version off the shelf. The moment a product proves it sells, you stop reselling it and commission your own.
8
1
34
1,985
Most founders selling into the EU still think the new July 1 duty is a flat fee per parcel. It's not. It's €3 per item category, by tariff code. T-shirt socks in one parcel = €6. Add a third product type = €9. The more varied your mix, the more every shipment bleeds. We've been working on a fix with our logistics partners since May 👀
1
1
6
1,119
I gave Claude my SKU list and asked one question: which products are net negative after returns. Units sold, return rate, margin per unit, ranked by profit after refunds. Two products I'd been calling bestsellers came back red.
5
1
6
764
Retention is where the business is actually built. A customer who buys once costs full acquisition cost on that sale. A customer who buys twice halves it.
2
5
363
Before a supplier call I have a model brief me on the negotiation in their terms. I give it the order size, my target price, and the AQL level, and ask for the three points the factory will push back on and how a domestic Chinese buyer would counter each. Be prepared.
2
202
When I started building the brand to sell, most of the revenue ran through one Meta ad account. A buyer reads single-channel dependency as a countdown timer. One suspension and you're dead. By the exit, email and creators carried real volume on their own. The Meta account could have died on a Tuesday and the month would have survived.
1
1
4
344
Our creator channel did €10K a month last Q4. It's at €70K now, with over 1,000 creators on the brand. Nothing clever changed. I just wasn't feeding it. Month one of seeding looks like wasted money. Month twelve looks like a moat.
3
1
10
246
Every year factories in China close for two weeks for the new year. The disruption lasts six to eight. Workers don't all come back. Production backlogs stack. Lead times that were three weeks quietly become seven. I place the Q1 orders in December and write the timeline into the PO. The brands that get wrecked every February are the ones treating the most predictable shutdown of the year as a surprise.
2
205
Found a new channel, third day on a row hitting 10 ROAS. Can’t complain with the first test. This year it’s all about omni channel.
12
30
4,456
A dropshipper and I can run the exact same product from the exact same listing. Whoever has cheaper CPMs wins that week, and it resets every week. The moment my product has its own mold, its own pattern, a label that's legally mine, the comparison shopping stops. That's what it means to build a brand.
3
5
433