Cut Amazon ACOS from 81% → 38% in 2024 while scaling daily spend 3×. Four channels live today.

The context

Nokkomo Mints is a DTC mints brand selling on Amazon and Shopify. When I joined as co-founder and CMO, the Amazon advertising operation was running at 81% ACOS: meaningfully negative unit economics. The account had a single Sponsored Products campaign attempting to do the work of a full structure: broad-match keywords with no negatives, no Sponsored Brands or Sponsored Display coverage, and creative that had not been refreshed in months.

The brand had reached the point where two options remained: continue spending against the current account structure, or rebuild from a clean baseline. We chose the latter.

What I built

A new Amazon Ads architecture from scratch.

• Campaign structure. Tight Sponsored Products with separate exact / phrase / broad campaigns per product. Sponsored Brands for branded defense + category headlines. Sponsored Display for retargeting cart-abandoners and competitor ASINs.
• Negative-keyword harvesting cadence. Weekly review of search-term reports. Bad search terms get added to a master negative list within 7 days, capped at a fixed waste threshold.
• Creative rotation system. Three image+headline variants per product running in parallel, swapped on a 14-day cadence based on CTR + CVR. Killed underperformers fast.
• Bid + budget logic. Day-parted bids reflecting actual conversion curves, not Amazon's default. Budget caps that flexed by day of week based on the prior 28-day attribution window.

A DTC funnel alongside Amazon so the brand wasn't single-channel-dependent:

• Shopify storefront with simplified PDP
• Klaviyo lifecycle: welcome, abandon-cart, post-purchase reorder
• Meta retargeting with Triple Whale attribution
• Cross-channel reporting so we could see actual CAC, not Amazon-only ROAS

Team

I built the intern program from scratch; it now runs 24 active interns across eight universities (UCLA, USC, Cal Poly SLO, SDSU, Michigan State, LMU, CSUN, and Illinois) across three workstreams (creative production, content operations, and Amazon catalog management), with rotating role assignments by quarter and weekly review cadences. The intern program also served as a structured pipeline for full-time hiring conversations and provided meaningful early-career experience to the team.

Outcome

• ACOS 81% → 38% in 90 days
• Daily spend scaled 3× while ACOS came down
• Brand became profitable on Amazon for the first time
• Built an intern program that started at eight and now runs 24 active across eight universities

Where it stands in 2026

The numbers above are that year's turnaround, not a running rate. Two years on, the brand sells across four channels (Shopify, Amazon FBA, Walmart, and TikTok Shop), carries two USPTO-registered trademarks, and leads with wholesale rather than paid acquisition. I report the 2024 figures because that was the turnaround I ran; the current picture sits beside them so the two are never confused.

Why this engagement matters

This was not an agency engagement. As co-founder I held equity in the outcome and full responsibility for the end-to-end marketing function: ad architecture, creative, Shopify storefront, and lifecycle email. The results reflect what is possible when an engineer rebuilds an advertising account from first principles, rather than operating within the defaults that a dashboard surfaces.