Laurence

Amazon sellers running ads typically optimize to ROAS, which ignores true unit economics like COGS, referral fees, FBA fees, and returns. Bids are often set uniformly across placements (top of search, rest of search, detail page) even though each placement has very different conversion rates and click values, which the page shows leads to overpaying and losing money per click in lower-converting placements. Sellers also must manually manage search-term harvesting, branded keyword defense, listing copy, backend SEO, and inventory reorder timing. Doing this manually or with simple ROAS-based tools means ad spend is miscalibrated to actual margin and customer value, and brand performance questions (like why CAC moved) require digging through reports instead of getting a direct answer.

Laurence is an autonomous ad platform for Amazon sellers that lets a seller pick an optimization goal, profit (true margin after COGS, fees and ad spend), CAC, customer lifetime value, or growth/share of voice, and then bids to that goal rather than to ROAS alone. It re-evaluates every keyword up to 24 times a day, reallocates budget to winning keywords, harvests search terms, and defends branded keywords from competitors. It prices bids per placement (top of search, rest of search, detail page) using placement-specific conversion rates rather than one flat bid, and factors in dayparting and incrementality (distinguishing branded searches that would convert anyway from incremental non-branded growth). It also surfaces listing optimization suggestions, backend SEO keyword refreshes, and inventory reorder points tied to sales velocity. A conversational interface ('Ask Laurence') answers plain-English questions about performance, logs every change made, and sends a weekly digest; it also connects via Slack, email, and an MCP-based agent integration.