Fractional COO and CEO · consumer brands, $3M to $50M

Your numbers are trying to tell you something.

Ben Johnson reads a consumer business the way a good doctor reads a chart, tracing every symptom back to the thing causing it. He fixes that from the operating seat, then installs AI systems that keep it fixed after he leaves.

Check your symptoms

A read is a fixed-fee, three-week operations assessment. You keep the write-up either way.

Ex-Amazon operations · Registered Amazon SP-API developer · 137,799 rows reconciled, zero mismatches

Weekly scorecard · wk 40 sample data
Orders, last 7 days 4,212 Shopify Amazon + 3PL say 4,390. Which one is true?
Blended ROAS 4.1x platform-reported graded its own homework
Hero SKU, days of cover 9 PO 4471 late stockout before the PO lands
Open items from last review 7 owners: 0 nobody owns these

Symptom checker

What are you seeing?

Tick whatever sounds familiar. Most of these trace back to one usual cause, and the read card shows what Ben would check first.

Ben's read 1 of 6 ticked

“Shopify, Amazon, and the 3PL all report different numbers.”

Likely cause
Each system is right about its own slice, and nothing checks one against another.
Week one
Pull every source raw and check each one against a second, independent source.
The fix
Every channel lands in a warehouse the brand owns and shows up on one scorecard, built on definitions the whole team agreed to.
Keeps it fixed
A reconciler that runs 34 blocking checks before each refresh and halts when a number drifts.

This is the usual pattern. A read of your business takes three weeks.

Step one · the read

Three weeks inside your operation, then a written read.

The read is a fixed-fee operations assessment. Ben works inside the business for three weeks and hands you a written diagnosis you keep whether or not you hire him for anything else.

  1. 01

    Time study

    Where the team's week goes, in 30-minute increments, and which parts only the founder can do.

    finds the bottleneck, with a name on it

  2. 02

    Order and data map

    Every system an order touches, from cart to 3PL to ledger, and the exact place the numbers fork.

    finds where Shopify and Amazon stop agreeing

  3. 03

    Revenue roll-up by channel

    DTC, marketplace, wholesale, and subscription on one page, down to contribution margin.

    finds the channel that only looks profitable

  4. 04

    Supply-chain baseline

    Raw material to finished goods to cash, SKU by SKU, against the PO calendar.

    finds the stockout before it happens

The readout

  • The one constraint holding the business back
  • What every recommendation does to cash
  • An owner for each dependency

Step two · the fix

Then Ben fixes it from inside the business.

The read names what's wrong. Fixing it usually means resetting definitions, moving decisions to named owners, rewriting the buy plan, and running a weekly review that holds people to dates. For a few brands a year, Ben does that work himself from the fractional COO or CEO seat.

Plan variance and the watchlist

Every gap against plan gets an owner and a date, and stays on the list until it closes.

Supply chain and inventory

Buy plans, POs, and the 3PL, planned from what is selling and how fast.

The marketplace channel

Amazon economics read from the source, through the registered SP-API and Ads integrations.

Finance and cash

Cash on the same page as revenue, so a good month on paper is a good month in the bank.

The agency

Briefs, spend approvals, and an acquisition cost the agency and the CFO both sign.

The team

Decision rights moved off the founder and onto named owners.

Step three · the prescription

Every fix ships with a system that holds it in place.

A fix fades when it depends on someone remembering it. So each fix comes with a system Ben installs in a warehouse you own, with AI doing the repeat work and checks catching drift. The whole set is called Growth OS.

The reconciler

Pulls every channel into one warehouse you own and matches each source against an independent witness before anything reaches the scorecard.

Take: every refresh

The gatekeeper

Thirty-four blocking data-quality checks. When a number drifts from its source, the pipeline stops and tells you.

Take: before every publish

The stock watcher

Reads the 3PL, the ERP, and the PO calendar together and flags a late PO or a thin SKU while there is still time.

Take: daily

The signal cleaner

Server-side tracking on your own domain, so Meta and Google bid on purchases that happened.

Take: every event

The Monday memo

An AI-drafted weekly business review written from the warehouse before the meeting. A person edits it and owns it.

Take: weekly

Drift alerts

Anomalies land in Slack before they land in the P&L.

Take: as needed

You own the warehouse and every record in it. If you and Ben part ways, it all keeps running.

See Growth OS →

The proof

Case files

Names are blacked out because most of this work runs under NDA. The numbers come straight from the systems. Revenue outcomes get added as each brand clears them for release.

File 01

Client Eight-figure supplement brand, mostly Amazon

137,799 rows

reconciled with zero mismatches

Symptom
Amazon reporting rebuilt by hand from exports every week.
Finding
Cut over to the live Amazon API and reconciled it against the exports it replaced.
File 02

Client Same brand

90,416 units

of hidden variance, about $375,692 at cost

Symptom
Inventory totals that looked fine.
Finding
The 3PL count and the ERP disagreed SKU by SKU while netting to under one percent overall. Surfaced 2026-08-03.
File 03

Client Same brand

45,000 units

in a reorder case with the math shown

Symptom
A late PO on a hero SKU.
Finding
The first exception the system flagged, turned into a demand-based reorder built on marketplace sell-through.
File 04

Client Beef protein brand: DTC, Amazon, KeHe, UNFI

~215,000

fake conversions removed from the ad signal every month

Symptom
Ad platforms reporting purchases that never happened.
Finding
Server-side tracking on the brand's own domain, delivered at event match quality 9.4 against a floor of 8.0.
Ben Johnson, founder of Zygo Consulting
Ben Johnson

Why root cause

Ben learned to look for the cause the hard way.

Ten years ago a gastroenterologist told Ben he'd be managing IBD for the rest of his life. He went looking for the cause instead, in food, sleep, movement, and stress. He got better, and along the way he got to know the founders building brands for people like him.

Amazon had already taught him the same habit for a business: when a number goes wrong, trace it back until you hit the thing that made it, then fix that. He ran operations there, built and ran a $70M direct-to-consumer portfolio, and held operating seats inside founder-led consumer brands including Ancestral Supplements, Paul Saladino MD's media company, and Dr. Gabrielle Lyon's practice.

Zygo is that habit, installed. Ben does the read and builds the systems himself, then hands them to your team, which is why they keep running after he leaves.

the brands that help people deserve operations this good

  • Ran operations at Amazon
  • Built and ran a $70M DTC portfolio
  • Registered Amazon SP-API developer and Ads Partner
  • TikTok Shop app developer
Read the full story →

How to start

Read first. Everything else is optional.

  1. 01

    The read

    Operations assessment · fixed fee · three weeks

    The time study, the order and data map, the revenue roll-up, and the supply-chain baseline, ending in a written readout you keep.

  2. 02

    The prescription

    Growth OS install · fixed fee

    The warehouse, the checks, the scorecard, the weekly review, and the alerts. Manual exports light up the dashboard in week one while API access comes through.

  3. 03

    The seat

    Fractional COO or CEO · limited

    For a few brands a year, offered after an install. Ben runs the weekly review himself.

Field notes

Notes from inside the numbers.

Short writing for operators who need the reporting to hold up before they make the next call.

AI Operations

AI Agents in Ecommerce Operations: What's Real in 2026

AI agents reliably run support triage, review synthesis, creative variants, and catalog work in DTC. Where they hold up, where they break, and the human gates.

Aug 24, 2026 → Leadership

Building the Ecommerce Team That Owns Outcomes

The hiring sequence for a DTC brand from $1M to $10M+, which metric each role owns, what to keep outsourced, and how to tell an outcome owner from a task owner.

Aug 24, 2026 → Scaling

Ecommerce KPIs by Stage: What to Track at $1M vs $10M

A $1M brand needs four ecommerce KPIs. A $10M brand needs about a dozen with named owners. Here's the stage-fit list, the review cadence, and what to cut.

Aug 24, 2026 →

Book a read

Find out what your numbers are hiding.

The read takes three weeks at a fixed fee, and the written diagnosis is yours whether or not you go further with Ben.

See Growth OS