The Stash Edge

ReadingThe steal: audio in-store is free inventory you already own. By adding AI targeting and data matching, Dollar General converted a cost center (store music systems) into a margin-accretive ad network without new capital spend on hardware. A CPG brand can now run a test: seed product to 500 stores, buy audio placement in those same stores for 30 days, measure incremental velocity. The data tie means attribution is clean—no guessing whether the ad drove the buy.

ReadingThe steal: profitable DTC is not about going viral—it's about turning repeat customers into a retention flywheel. Reformation's filing proves that if you get repeat rate above 35%, you can fund growth without external capital. The play: measure repeat rate month-over-month, price at a margin that funds organic word-of-mouth (not paid ads), and reinvest retention spend into product. A small brand can run this: $50K cohort, track 30-day repeat, 60-day repeat, and 90-day repeat. If you hit 30% repeat by month two, you have the math to scale.

ReadingThe steal: don't seed all creators at once. Seed micro first (your network), prove engagement and repeat content for 3 months, then graduate to mid-tier. This creates a paper trail that retail buyers trust. When you walk into a buyer meeting, you have 12 weeks of content proof from micro, 6 weeks from mid-tier, and early signals from category ambassadors. You're not asking them to bet on virality; you're showing them a repeatable pattern. Run this: list 20 micro-creators in your niche, send 10 free units, ask them to post one organic video. If 8 post and average 5K views each, you have your proof point for mid-tier seeding.

ReadingThe steal: don't wait for your product to be relevant to culture—move to where culture is happening. The World Cup is free audience aggregation. NYX didn't ask women to come to them; they set up a booth where women already were. The mechanism is inventory-aware: each pop-up had finite sample stock, which created urgency. A small beauty brand can run this: identify the next 3 moments your audience will gather (concert series, sports event, festival), reserve a small booth ($2K–5K per location), send 500 units of your hero product, and staff it with a single person for 4 hours. If 20% of foot traffic samples and 3% of samples convert to repeat purchase, you've found your activation cost-per-repeat customer.

ReadingThe steal: the QR code is a redirect. Print once, update forever. A brand prints packaging with a QR code that links to a domain they control. When an ingredient changes or a promotion ends, they update the landing page—no new packaging order. Cost per update: zero. Cost per obsolete batch: eliminated. For a small brand doing limited runs, this is a margin play: instead of printing 5,000 units of packaging with dated info, print 5,000 units with a QR code and update the backend. If you revise a promotion or ingredient once per quarter, you save $800–2,000 per print run.

ReadingThe steal: streaming ads are more expensive than shelf media but allow for audience matching at scale. A retailer can now say: 'This person bought from category X in the last 30 days; show them a video ad for category Y on streaming.' The cost per impression is higher, but the conversion rate is higher too because the audience is pre-qualified. A small brand shouldn't buy this alone; instead, partner with a retailer's media network. Ask your retailer: 'Can you test our product in a streaming cohort matched to similar purchasers?' You pay for the media; the retailer does the audience match.

ReadingThe steal: if AI agents can buy media, they can optimize faster than humans. An agent could monitor foot traffic in real time and buy incremental DOOH inventory in zones where demand signals spike. For a brand, the watch is whether this pricing becomes dynamic—if you can buy DOOH by foot-traffic minute instead of by pre-bought calendar week, the math changes. Early signal: the platform was announced at Cannes Lions, which suggests it's in early adoption. Watch for the first CPG brand to report results.