The Stash Edge

ReadingThe steal: don't build last-mile; rent it. Find a partner whose network already touches your buyers' zip codes, then feed them volume until the per-unit cost drops and your unit economics flip. CarParts didn't hire drivers — they added SKUs to an existing delivery backbone and watched the quarterly number climb. Run the same move this week: audit your current carriers' hit-rate in your top 10 metros, then bundle your repeat buyers into a small weekly batch and pitch a partner-rate test.

ReadingThe steal: identity-driven seasonal drops. Find a community with hard affiliation markers — school colors, team loyalty, mascot pride — and build your seasonal calendar around their moments, not retail holidays. BaubleBar didn't sell jewelry to college students; they sold school pride and timed the drops to championship season. Run this: pick a niche community with high identity-lock (alumni groups, regional sports clubs, sororities), design a 3-drop calendar around their signature moments (homecoming, rivalry week, championship), and pre-seed micro-creators from inside that community 4 weeks before drop one.

ReadingThe steal: don't choose one channel — stack them in the sequence that matches your unit economics. COS opened flagships where they had brand awareness, then used those stores to anchor e-commerce conversion in surrounding areas, then licensed partnerships to fill secondary metros without capital. Run this: if you're sitting at 50%+ DTC AOV margin, open one owned flagship in your highest-LTV city (not your biggest city — your highest-LTV city), run that store's traffic back to your email list, and license a distributor partner into two secondary metros simultaneously. Watch which channel unlocks first, then allocate next quarter's spend there.

ReadingThe steal: apply to LEAP if and only if you have documented repeat purchase data. Whole Foods isn't looking for the best story — they're looking for the lowest risk. Pull your repeat-customer cohort (orders two and three), calculate your repeat rate, and use that number in your LEAP pitch. Don't lead with the founder story or the product innovation. Lead with: 'Our first-purchase cohort has a 34% repeat rate,' then show the operational readiness (supply-chain health, packaging compliance, unit economics). That's the only thing Whole Foods cares about in the accelerator.

ReadingThe steal: audit your packaging against GS1 Digital Link standards now, not in 2027. If you're doing a reprint run in the next 12 months, add the QR code to the spec. The cost is negligible — 2-3% of packaging if you're already reprinting. Build the connected-packaging URL architecture (the destination page that loads when the code is scanned) in parallel. Don't wait for the deadline; move three months early so you have time to test the retail buyer's scanning workflow before compliance goes mandatory.

ReadingThe steal: audit your repeat buyers' friction points, not your point-system conversion. Run a 2-week friction audit: track every step from 'customer opens your app/site' to 'order ships.' Count the clicks, the form fields, the waiting. Then pick the one step that's causing the most dropoff between first and second purchase. Fix that one step. Don't add a loyalty program; remove a step. Most repeat-purchase programs fail because they add complexity when the real problem is friction on the first repeat attempt.

ReadingThe steal: if you have domain credibility in one category, audit the adjacent symptom categories your buyers mention. Midi's customers say 'my skin changes with my cycle' — that's the product line. Don't launch adjacent categories because the market is big; launch them because your existing buyers are already talking about the problem. Do a 30-minute customer interview blitz with your last 100 repeat buyers and ask: 'What else are you buying because of [your core product]?' Write down the patterns. Build one adjacent SKU around the most-mentioned problem. Pre-sell it to your email list before you manufacture. If 12%+ of your list buys at launch, build it. If not, pick another problem.