Laurence

Problem/Solution Report

The company frames today’s Amazon PPC as dominated by guesswork and misaligned incentives. Agencies tend to minimize time per account, rely on Amazon’s “suggested bids,” and use oversimplified formulas that treat keywords as binary good/bad rather than as curves where the profit‑maximizing bid varies with changes in conversion probability. Existing “AI‑powered” tools are criticized as mostly rules‑based systems—lower bids when ACOS is high, raise them when ACOS is low—without modeling uncertainty or thin‑data regimes. The result, according to Laurence, is that sellers overpay for clicks, miss profitable demand, and react too slowly to market movements.

Laurence’s solution is a full‑service Amazon Ads management model grounded in quantitative methods similar to those used in systematic trading. Practically, the platform models the full relationship between bid and conversion rate for each keyword; encodes a seller’s margin/ACOS goals as hard constraints; updates bids hourly using Amazon Marketing Stream; and “borrows” information across similar keywords to handle sparse data. The approach emphasizes operating under uncertainty—bidding cautiously when signals are weak and scaling aggressively when confidence is high—while maintaining interpretability (every decision is traceable with conversion estimates, confidence, and constraints).

For sellers, the stated value proposition is consistent improvement in efficiency and scale with transparency: “Others guess. We calculate.” The website and social profiles claim 30% ROAS improvement within ~90 days for sellers typically in the $1M–$10M+ revenue range, and YC notes the company was managing roughly $2.5M in live ad spend with a reported ~30% ROAS lift at the time of its posting. The differentiator is not just automation but the combination of engineering rigor, hour‑by‑hour recalibration, and profit‑margin discipline embedded directly into the optimization process.

Market and Competitors Report

Laurence operates within the Amazon retail media segment—specifically Sponsored Products/Brands/Display for third‑party (3P) sellers and brands on Amazon. Amazon’s “Advertising services” revenue was $56.2B in 2024, up from $46.9B in 2023, with quarterly disclosures through 2025 showing continued growth and year‑over‑year increases in the low‑20% range. Independent forecasts project Amazon’s retail media ad revenue could approach ~$79B by 2026, underscoring a rapidly expanding market.

The competitive landscape includes two major buckets: (1) software platforms for Amazon PPC automation/optimization and (2) agencies offering managed service. Prominent software options include Perpetua (Sellics/Perpetua), Teikametrics, Pacvue, Intentwise, Jungle Scout’s Downstream/Cobalt, Helium 10 Adtomic, Skai (Kenshoo), BidX, m19, AiHello, Seller Labs Pro, PPC Entourage, Sellozo, SellerApp, Trellis, and others. These tools span self‑serve to enterprise, often mixing rule‑based automations with some AI features, along with dashboards, analytics, day‑parting, keyword harvesting, and AMC integrations. Agencies (not exhaustively listed here) range from boutiques to large full‑service players—many layering these same toolsets with service and reporting.

Laurence’s stated competitive angle is depth in quantitative modeling under uncertainty (rather than rule‑based tuning), hour‑by‑hour bid recalibration via Amazon Marketing Stream, hard profit constraints, and decision traceability. Its messaging explicitly calls out the limitations of rules engines and the misalignment of agency incentives, positioning its approach as both more scientific and more aligned with seller profitability.

Founders and Background Report

Laurence is a New York–based startup founded in 2025 and currently a YC Winter 2026 company focused on full-service Amazon Ads management. Public profiles indicate a lean founding team (2–10 people) led by co‑founder Matthew Chen alongside co‑founder Leo Gierhake. The company emphasizes deep quantitative and engineering expertise applied to retail media advertising on Amazon.

Matthew Chen serves as Founder/CEO/CTO. His background spans Google and Meta, where he “built Data Science AI at Google” and worked on ranking models at Meta—systems where mistakes “can cost billions.” He also worked on data infrastructure for Amazon agencies, which informed his view that conventional agency/playbook approaches are fundamentally broken. Academically, materials indicate Math+CS and AI research at Cornell. He highlights applied modeling experience outside of work too, having trained his own models to rank in the top 1,000 out of 11 million entrants in fantasy soccer competitions. The company’s Team page further underscores his role spanning product/engineering leadership and hands‑on quantitative modeling.

Co‑founder Leo Gierhake brings a systematic trading background. YC lists him as a former researcher at Jump Trading where he built quantitative models in high‑velocity markets and notes additional signals of decision‑making under uncertainty (e.g., $250,000 in high‑stakes poker winnings). He previously founded a crypto investing platform at age 18 and studied Electrical Engineering at ETH Zurich—another indicator of a strong quantitative foundation applied to dynamic markets.

Laurence positions its founding DNA as “built by engineers, not amateurs,” contrasting the team’s experience in high‑stakes systems (big tech, quantitative trading, and large Amazon agencies) with the manual, rules‑based operations commonly used by agencies and software tools. The public LinkedIn profile corroborates that Laurence is headquartered in New York, founded in 2025, and operating with a small, technical team.