Entrepreneur-Familyman-Adventurer. I believe in Trust, Transparency and Customer Success. I believe in Force existing in every process for the good to happen.
Win-back loop. 50–70% of buyers lapse, so the system watches repeat windows, ships the right prompt, and routes lapsed cohorts back before the LTV math goes flat. Skip cohort timing and you just mail everyone.
6% referral fee. 25-35% all-in. TikTok Shop looks cheap until fulfillment, affiliate cuts, and returns hit GMV. That stack is the price of transaction flow, not customer ownership.
15% referral + 4.99 FBA + 2.80 returns on a $39.99 ASIN. That receipt is why settlement statements lie and P&L wins only after CM math. Claude Code reads it, not a junior buyer.
Klaviyo sold a dashboard humans used to run email. An agent can write, segment, and trigger the same flow without the workbench. The interface was the constraint, not the data.
5-12% of lapsed customers come back from triggered win-back. Same offer, different timing: 3 emails, escalating only after the soft reset. It replaced the annual 'we miss you' blast.
Bid tweaks is what an agent now does. Margin floor, creative judgment, and channel mix is what the operator does instead. Set break-even ACOS per SKU this week, then let Claude Code obey it.
$50 CAC, $33 contribution, 28% repeat rate. That’s why most Shopify brands lose money on the first order and wait on a second. Win-back is the biggest salvage pool; Shopify is where the customer relationship stays.
6% referral fee. 30% all-in. TikTok Shop looks cheap until FBT, affiliates, and returns hit the P&L; GMV can rise while margin vanishes. It pays transactions, not customer ownership.
15% referral. 4.99 FBA. 7% returns. On Amazon, profit lives after the fee stack; settlement deposits lie, P&Ls lag, and break-even ACoS is just contribution margin in a spreadsheet.
Klaviyo sold email automation. An agent can write, segment, and send the same campaign without the dashboard, report, or Friday approval. The cadence was the constraint, not the value.
$18k in lapsed-customer revenue. Triggered 30-60-90 email plus SMS beat the annual 'we miss you' blast; one path recovered 9%, the broadcast 2%. Same list, different pipeline.
Bidding by gut is what an agent now does. Defining the margin floor, creative rule, and channel mix is what the operator does instead. This week: set break-even ACOS per SKU, then let Claude Code follow it.
$50 CAC. $80 first order. $33 gross profit. That looks fine until shipping, processing, and fulfillment turn a 50% Shopify margin into ~30% net. The second order is where the model stops lying.
6% fee on TikTok Shop, but 25-35% all-in once FBT, affiliates, and payments hit. GMV can look great while margin quietly bleeds; the channel is a transaction, not a customer ledger.
15% referral + FBA + storage + returns on Amazon. That's before ads. Break-even ACoS is just contribution margin %, so the receipt, not GMV, tells you if the SKU earns its keep.
Klaviyo sold a human-operated inbox and report stack. An agent can run segmentation, send logic, and win-back timing without the dashboard. The interface was the constraint, not the data.
Weekly MBRs are what an agent now does. Cohort reads, re-lapse checks, and win-back sequencing are what the operator does instead. This week, wire lapsed segments into Klaviyo before the next Friday review.
Managing bids is what an agent now does. Setting the margin floor and channel mix is what the operator does instead. This week, wire CM% into Amazon, Meta, and TikTok before spend moves.
70% lapsed base. On Shopify, win-back is the biggest pool you can still monetize. Most brands keep buying first orders because the customer graph lives in a dashboard, not in the model.
6% headline fee. 30% all-in on TikTok Shop. Freight, affiliate rev-share, and returns eat the rest. That replaced the Amazon pricing habit of treating referral fee as the whole bill.