Kovalchick Analytics

About

I am a physical therapist who sells on Amazon and got tired of guessing.

Kovalchick Analytics is one person. That is a limitation and it is also the reason the work is any good.

Mathew Kovalchick, PT, DPT

I am a Doctor of Physical Therapy. I work per diem at UCLA Health in Santa Monica and I live in Venice, California. A few years ago I started a small Amazon brand selling exercise and recovery products, because I kept writing the same home program on paper for patients and wanted something better to hand them.

Running that brand meant running its advertising. I tried the tools. Some of them made things worse, and the ones that did not still could not answer the only question I cared about, which was whether a given campaign was making me money after cost of goods, Amazon fees, fulfilment and returns. They all reported ACoS. None of them knew what my products cost me.

So I built the thing that does. It pulls my account every day, keeps a full local mirror, computes a breakeven return on ad spend for each product from its real unit economics, and makes bid, budget and negative keyword changes against that number rather than an account average. Every change it makes lands in a ledger. It has been running on my own money since April 2026.

Then a couple of other sellers asked whether I could point it at their accounts. That is what this is.

Why "Analytics" and not "Agency"

I am not selling you consulting hours or a strategy deck. I am selling a system that reads your account and acts on numbers, operated and watched by me. The name is meant to be literal.

What I bring that a PPC generalist does not

Two things, and they are related. The first is that I sell physical product on this platform with my own money, so every decision the system makes has been tested on an account where being wrong costs me directly. There is nothing theoretical about the profit floor. I added it because I needed it.

The second is clinical habit. In physical therapy you do not get to claim an intervention worked because the patient feels better. You measure, you keep a baseline, you name your outcome before you start, and you are honest when the data does not support the thing you hoped. That is the same discipline that makes me measure contribution profit per product per week rather than quoting you an ACoS, and that makes me say plainly that one account is not a study.

Being one person, honestly

The upside is that the person reading your account is the person who wrote the system, and you are never handed off. The most common complaint in public reviews of the bigger tools and agencies is account manager turnover, and it is structurally impossible here.

The downside is capacity. I keep the client list small because past a certain number I would be triaging instead of thinking, and you would feel it. If I am full I will say so rather than take the money. If you need somebody on call at 2am across four time zones, I am the wrong choice and I will point you somewhere better.

What the record shows

Every figure here comes out of my own database and I will walk you through the derivation if you want it.

On my own brand

Measured weekly, per product

  • Weekly contribution profit up 5.2x between 19 July and 30 August 2026, from $83 to $432
  • Total advertising cost of sales of 24.4% across that period
  • Profit and loss per product per week, including cost of goods, Amazon fees, fulfilment and refunds
  • 4,181 successful account changes since April 2026, each with a before value, an after value and a reason
On a client account

What per-SKU margin found

  • Three products scored against their own breakeven returns of 1.95x, 2.33x and 3.17x, computed from their real cost of goods
  • An account level view showed 6 profitable campaigns and 6 losing
  • Scored properly it was 6 profitable and 13 losing, at a net $540 a month
  • The account was being steered to one 3.00x target that fitted none of the three products

The offer is simple

Let me audit your account for free. If the audit is any good you will know whether to keep talking, and if it is not you have lost twenty minutes.