The Synthesis blog

The math behind your brand, explained plainly.

How to give AI a version of your business it can actually reason about, and the operator math it needs to get right: landed cost, contribution margin, product structure, settlements and attribution. Written from what we see in real multi-channel brand data, with illustrative numbers in every example.

Start here · AI-ready data

What is AI-ready data? The data AI can actually understand

AI-ready data is business data an AI can reason about without you in the room. Every number has one written definition, a labeled source and period, and a link to the products, orders, customers and spend it belongs to. Facts only you know, like unit cost, are written down and approved. Answers are checked against the source of record before anyone acts on them.

Read the guide
AI-ready data

Why ChatGPT and Claude get your Amazon and Shopify numbers wrong

Paste an export into ChatGPT or Claude and the answer sounds right but is not. Ten ways AI misreads Amazon and Shopify data, and what fixes each one.

September 15, 2026 · 9 min read
AI-ready data

What is an AI harness? The part that makes AI answers about your business safe to act on

The model is the engine. The harness decides what it can reach, what it is told, what it may change and how its answers are checked. Five parts, for brands.

September 15, 2026 · 10 min read
Comparisons

Triple Whale and Polar alternatives for Amazon-heavy brands

Most Triple Whale and Polar alternative lists assume a Shopify store. Five questions to ask if Amazon is a big share of your sales, and six tools compared.

September 15, 2026 · 10 min read
Brand structure

SKU rationalization: how to decide which products to keep, fix or cut

SKU rationalization for consumer brands: judge product families first, rank variants by contribution margin per channel, count stock cost, then keep or cut.

September 15, 2026 · 10 min read
AI-ready data

Semantic layer vs ontology vs context layer: what a consumer brand actually needs

A semantic layer defines metrics, an ontology models the business, a context layer briefs the AI. One Amazon order shows what each adds and what to build first.

September 15, 2026 · 10 min read
Brand structure

Product hierarchy for consumer brands: from company down to the SKU

How to structure a consumer brand's catalog from company to brand, product family, variant and SKU, and why Amazon and Shopify listings do not do it for you.

September 15, 2026 · 11 min read
Unit economics

Is Amazon Subscribe & Save profitable? The break-even math for sellers

Subscribe & Save trades a discount on every order for more orders. Who funds each tier, the break-even math, and the traps that make it look better than it is.

September 15, 2026 · 10 min read
Measurement

Incrementality vs attribution vs marketing mix modeling: what each one tells you

Attribution hands out credit, incrementality tests measure extra sales, and a mix model reads spend against sales. What each answers, and where each misleads.

September 15, 2026 · 10 min read
Unit economics

How to calculate COGS and landed cost for Amazon and Shopify

The monthly COGS formula, landed cost per unit, what is not COGS (FBA and referral fees, ads), freight allocation and FIFO vs average, with a worked example.

September 15, 2026 · 10 min read
Unit economics

First-order profitability and CAC payback: how long until a new customer pays you back

First-order profit is a first order's contribution margin minus CAC. Payback is the days until a cohort earns CAC back. Formulas, a worked example, and Amazon.

September 15, 2026 · 10 min read
Measurement

Do Meta ads drive Amazon sales? How to measure it

Meta shows almost no purchases for ads that send shoppers to Amazon. Here are the three ways to measure what those ads really sell, and where each one breaks.

September 15, 2026 · 10 min read
Unit economics

Contribution margin for Amazon and Shopify: CM1, CM2 and CM3 by channel

How to calculate contribution margin for ecommerce when you sell on Amazon and Shopify: CM1, CM2, CM3, a per-unit waterfall, and break-even ROAS per channel.

September 15, 2026 · 11 min read
Unit economics

Break-even ROAS and MER for Amazon brands: the formulas, and where they mislead

Break-even ROAS is 1 divided by contribution margin before ads. How to set it per channel, why MER misleads when Meta sells on Amazon, and what to track.

September 15, 2026 · 10 min read
AI-ready data

Best MCP servers for Amazon and Shopify brands: three kinds, and a 20-minute test

Platform MCP servers, data-pipe connectors and a joined brand model each do a different job. What each is good for, where it stops, and a 20-minute test.

September 15, 2026 · 11 min read
Amazon

Amazon settlement report vs orders report: which revenue number is right?

Amazon's orders report shows what customers bought. The settlement report shows what Amazon paid you. Here is when to use each, and how to reconcile the two.

September 15, 2026 · 10 min read
Amazon

How to read Amazon's Search Query Performance report

Amazon's Search Query Performance report shows how often you were shown, clicked and bought on each search. How to read it, and the traps that make shares lie.

September 15, 2026 · 9 min read
Amazon

Amazon seller fees per unit in 2026: every fee, and what is left of a $24 sale

Every Amazon fee that can hit one unit in 2026, from referral and FBA to storage, placement and returns, with a worked $24 example of what is left per sale.

September 15, 2026 · 11 min read
Amazon

Amazon customer lifetime value: how to calculate LTV you can trust

How to calculate customer lifetime value on Amazon without the usual traps: cohorts, margin at fixed ages, buyer-data gaps, and a fair comparison with CAC.

September 15, 2026 · 10 min read
Amazon

Why your Amazon ad sales and ACoS keep changing after the week ends

Amazon keeps crediting ad sales to a week for days after it ends, so ACoS falls as the week ages. How much it moves by ad type, and how to compare weeks fairly.

September 15, 2026 · 9 min read