The full scope

Amazon is one machine.
It runs at the speed of the gear nobody is watching.

This page is everything inside the partnership, in detail: the twelve gears I run, why each one exists, and why buying them separately quietly fails. If you are deciding whether one fee is worth it, read to the bottom before you decide.

Why together

You can hire someone to polish one gear. The machine does not care.

Stock feeds rank. Rank feeds traffic. The listing converts traffic. Conversion feeds rank back. Reviews compound conversion. Fees and pricing decide whether any of it makes money. And the P&L is the scoreboard that tells you the truth about all of it. Here is how the machine breaks when the gears are run separately. Every one of these is a failure I have lived, on my own brand or on brands I manage.

1. Inventory breaks PPC

Run out of stock and Amazon reacts in about 48 hours: your badge disappears, your ads get throttled, your rank starts sliding. When stock comes back, you keep paying, because the "bought in past month" counter is a rolling 30 day window. A week of stockout drags your social proof for a full month after it ends.

Lived it: during one stock-constrained week, with budgets and bids completely untouched, Amazon cut ad impressions and clicks nearly in half. The spend was available. Amazon had simply decided the listing deserved less traffic until the delivery promise was fixed.

2. PPC breaks inventory

The reverse is just as expensive. A SKU can sit at one sellable unit and still read "in stock," which means its campaigns keep charging while it cannot ship. Ads and inventory are one decision, not two departments with two dashboards.

Lived it: on a brand I took over, ads ran for two straight weeks on a SKU that could not ship, billing every day, before anyone connected the two systems.

3. PPC breaks against the listing

Ad clicks are bought traffic landing on a page. If the page does not convert, you are renting visitors you cannot keep. And keywords lie: terms describing what a product IS often convert at zero while terms describing what it DOES convert in double digits. You only find that when the person running ads is the same person rewriting the listing, because the fix is copy, not bids.

Lived it: on one account, every keyword describing what the product IS racked up clicks and produced zero orders. A keyword describing what it DOES for the customer became the best converter in the account. The fix was copy, not bids.

4. Fees rot silently underneath everything

Amazon bills by their record of your product, not by your product. Records go stale: you repack and nobody remeasures, the wrong item gets measured, or dimensional weight quietly takes over and you pay to ship air. No dashboard flags any of it.

Lived it: a multi-pack billed at more than double its actual weight because the box was mostly air. And a ten minute remeasure request on a product whose record was wrong got money refunded and cut the fee on every future order, permanently.

5. Reviews are two jobs, and most brands do zero of them

Job one is getting reviews to flow: automated, compliant, every order asked. Job two is reading them, yours and your competitors', as free market research. Customers write down exactly what to fix, what to build next, and when a product is aging out. Most brands let that stream pile up unread.

Lived it: a client's newest products sat at zero reviews for months, then got reviewed across the range within weeks of switching on compliant automation. The reviews now do selling the ads used to pay for.

6. Demand planning is what makes any of it durable

Every chain above ends in the same place: how much stock, ordered when, landing where. Forecasting off last month's sales fails on Amazon because ads, deals, seasonality and rank changes all move velocity. Demand planning is where marketing decisions and supply decisions meet in one plan, or you get chain 1 again.

The background: this is the discipline I ran for 16 years at PepsiCo, Unilever, P&G and Kenvue. It is the part of the machine agencies do not even offer.

7. The retail calendar multiplies everything, in both directions

Prime Day, Black Friday, Cyber Monday and your seasonal peaks are the biggest traffic days Amazon will ever hand you, and they are unforgiving of improvisation. Deal slots book weeks ahead. Event inventory is ordered months ahead, because stocking out mid-event means paying for the traffic surge and handing Amazon a stockout signal on the biggest data day of the year. Run the calendar well and one event can fund a quarter. Show up unplanned and it is the most expensive way to find every weak gear at once.

The scope, gear by gear

Twelve gears. One operator.

This is the actual work, listed the way I run it. Each gear says what breaks without it, because that is the honest way to evaluate scope.

GEAR 01

Inventory availability & inbound

  • Stock watch across available, reserved, inbound and FC transfers, not the summary number that lies
  • FBA shipments created, tracked and chased through check-in
  • Stockout risk flagged in days of cover, before it is visible on the surface
Without it — rank, badges and ad delivery throttle within 48 hours of a stockout, and recovery costs a month.
GEAR 02

Demand planning & forecasting

  • Velocity-based forecasts that account for ads, deals and seasonality
  • Reorder points and production timing tied to real lead times
  • One plan where marketing and supply decisions meet
Without it — stockouts on your winners, cash buried in your losers.
GEAR 03

Listing optimization

  • Titles, bullets, A+ content, image strategy, backend keywords
  • Variation families built correctly so reviews pool and pages consolidate
  • Copy written for the three questions Amazon's AI now reads your page to answer: what is it, who is it for, why this one
Without it — paid and organic traffic that arrives and does not convert.
GEAR 04

PPC management

  • Campaign structure, bids, budgets, keyword harvesting, negatives
  • Managed on TACoS (total sales), never just ACOS (ad sales), so organic growth counts
  • Spend synchronized with inventory so ads never run on stock that cannot ship
Without it — spend leaks into keywords that describe your product instead of keywords that sell it.
GEAR 05

Fee & packaging optimization

  • Quarterly fee audit: billed dimensions vs tape measure, size tiers, dimensional weight math
  • Remeasure requests filed when the record is wrong, refunds recovered
  • Packaging specs run through fee math before anything gets printed
Without it — a silent percentage of every order lost, forever, with no dashboard to flag it.
GEAR 06

Pricing architecture

  • Price points mapped against category fee cliffs (in grocery, two cents of price can nearly double Amazon's commission)
  • Multi-pack economics: the pick-and-pack fee paid once instead of three times
  • Subscribe & Save discount strategy on the products where repeat behavior earns it
Without it — a price that felt right, quietly handing Amazon your margin.
GEAR 07

Promotions & the retail calendar

  • Deals, coupons and S&S promotions run deliberately, not reactively
  • Prime Day and Q4 planned like a retailer plans a season: slots booked ahead, inventory built months out
  • Ad budgets staged before, during and after events to capture the cheap post-event rank
Without it — the year's biggest traffic days arrive unplanned, and you either miss them or stock out in the middle of them.
GEAR 08

Reviews automation

  • Compliant review generation on every order, new and mature SKUs alike
  • Review velocity tracked as a growth metric, not a vanity metric
Without it — new products stall at zero social proof while competitors compound.
GEAR 09

Review mining for product development

  • Monthly read of your reviews plus your competitors' reviews
  • Ranked fix-list: what customers actually complain about, in their words
  • Innovation signals and lifecycle flags: what to build next, what is aging out
Without it — product decisions made on opinion instead of what buyers wrote down for free.
GEAR 10

Catalog integrity & account health

  • Correct dimensions, attributes and variation structures on every record
  • Suppression fixes, policy compliance, account health watch
  • The unglamorous flat-file work that decides your fees and your findability
Without it — wrong fees, broken families, and listings that vanish mid-quarter.
GEAR 11

Search insights & experiments

  • Search Query Performance: impression share vs purchase share, where buyers are actually lost
  • The three-number funnel on every product: impressions to clicks to orders, each gap naming its own fix
  • Amazon A/B experiments on images, titles and A+ so changes teach something either way
Without it — you know sales moved, but never where buyers were lost or why.
GEAR 12

Reimbursements & recovery

  • Lost and damaged inventory tracked against Amazon's reimbursement policies
  • Refunded orders verified: unit back in stock or money back, never neither
  • Promised reimbursements confirmed as paid, because approved and paid are different events
Without it — Amazon's mistakes stay Amazon's gain.

And the gear that rules them all: the Amazon P&L and cash flow

Every gear above is only right or wrong relative to two numbers: your true contribution margin per unit after every Amazon fee, and your break-even ad spend. Get those wrong and everything downstream is guessing with confidence.

And a correct P&L still is not enough, because Amazon pays on a lag while ads bill on a threshold. A growing brand can be profitable on paper and out of cash in real life. I track both: the P&L that says whether the machine makes money, and the cash flow that says whether you can afford to keep it running.

Every recommendation I make arrives with its P&L consequence attached. That is the difference between an agency that reports ACOS and an operator who is accountable for money.

Accountability

How you see it working

Three artifacts, every engagement, from day one. This is the part most agencies will not sign up for.

A glide path

Before work starts, we agree on milestones and what to expect in each phase. Progress is measured against what was said, not against vibes. You always know where we are versus where we said we would be.

A locked weekly report

Every Monday, same sections, same format. Total sales and TACoS lead. Profit block, per-product scoreboard, inventory position, what changed, what happens next. Fixed structure means drift has nowhere to hide.

Scored calls

When I forecast something, it gets written down with a date, and later scored right or wrong, in front of you. You should know your operator's hit rate. Most agencies make sure you never do.

What good looks like

The shape of the first months

Every brand starts in a different place, so honest expectations beat promised multiples. This is the sequence the machine follows when it is run properly.

Weeks 1 to 4

Foundation and found money

Full operations audit. Fee record vs reality on every product, wrong dimensions remeasured, reimbursements recovered. Break-even math computed per SKU before a dollar of ad spend moves. Catalog defects fixed. Review automation switched on. The glide path locked together.

Months 2 to 3

Conversion and compounding

Listings rewritten against search-query data, spend moved from identity keywords to converting ones, pricing repositioned against fee cliffs. Conversion improves, Amazon widens distribution, organic share grows. Inventory plan running ahead of velocity instead of behind it.

Month 3 onward

Scale on purpose

The retail calendar worked deliberately: deals booked, event inventory built, post-event rank captured. Review mining feeding product decisions. The weekly report now reads like a compounding curve instead of a set of disconnected numbers, and every month closes with receipts against the glide path.

The honest math on the fee

$5,000 a month sounds like a lot until you price the alternative. A PPC-only agency runs one gear for $1,500 to $3,000 and leaves eleven unattended. The failures above do not announce themselves: a wrong size tier taxes every order, a stockout quietly resets months of rank, an unplanned Prime Day burns a quarter's budget in a week.

One prevented stockout on a hero SKU, one corrected fee record, one event run properly. Any single one of those routinely covers months of the fee. Running all twelve gears is how those saves stop being lucky catches and start being the system.

Anyone can make one gear spin.
The job is making the machine compound.

Stock that never breaks rank. Ads that never outrun the listing. Fees that never eat the margin quietly. Reviews that feed both sales and the next product. And a P&L that tells the truth about all of it, every week, in writing.

The audit is free, and it will show you which of your gears is slowing the machine, whether we work together or not.

Get your free operations audit