The Operating System We Built When We Could Not Afford a COO

In 2018, two years into running PerfumeM, I was the de facto COO of a business too small to actually have one. Inventory, fulfillment, customer service, supplier negotiations, returns processing, and finance all sat on my desk. The week after we lost $12,000 to a single botched international shipment was when I realized that scaling without an operations infrastructure was not a temporary phase. It was the business.

Nine years later, PerfumeM is bootstrapped, profitable since year two, and operates a 3,400-SKU catalog with a small team. What I built in the intervening years was less an operations function and more an operating system. A small set of rules and routines that compress the cognitive load of running the business so that operations does not become the bottleneck.

This is what I would tell another founder building an operating system for a business too small to have a dedicated COO.

### The Weekly Heartbeat Replaces Quarterly Planning

For the first three years I tried to run the business on quarterly cycles. Plan inventory quarterly, forecast revenue quarterly, review performance quarterly. The cycle was always wrong. Decisions could not wait three months and reviews three months out always lagged the reality of what was happening on the shop floor.

In 2020 I switched to a weekly heartbeat. Every Sunday evening, an hour-long routine I have not skipped in six years. Four things get reviewed: last week's sellthrough by SKU velocity tier, cash position and 13-week rolling forecast, inventory positions against the cover-window rules, and customer service queue health. That hour replaces what most operations teams use a full week of meetings to cover, because the data is the conversation. No interpretation, no consensus-building. Just the numbers and the next-week actions they imply.

The weekly heartbeat compounds. After 50 weeks of consistent review, you have a deep intuition for what abnormal looks like. Anomalies show up within a week instead of within a quarter. That gap is where most small operations problems become large operations problems.

### Decisions Get Routed By Reversibility, Not Importance

The conventional operations advice is to prioritize decisions by importance. High-impact decisions get more attention, low-impact decisions get delegated. The framework I actually use is reversibility. Reversible decisions get made fast and reviewed later. Irreversible decisions get the long deliberation.

A purchase order for $5,000 of a fast-moving SKU is high-impact but reversible. If demand drops, we have a markdown lever. That decision happens in under five minutes from the weekly velocity data. An eight-month supplier commitment for an exclusive distribution agreement is potentially lower-impact in pure dollars but irreversible because of the relationship dynamics. That decision gets three weeks of thinking and a conversation with a mentor.

This framework matters operationally because it changes what you put on calendars. Reversible decisions never get a meeting. Irreversible decisions always get a documented thinking session. The operating cost of small businesses is mostly the meetings about decisions that did not need to be discussed.

### Two-Bucket Cash And Tiered Inventory Both Follow the Same Logic

The cash discipline rule we run by is two-bucket: an operating account holding 60 days of fixed expenses (rent, payroll, software, debt service), and a growth account holding everything above that. The operating bucket is never touched for opportunities. The growth bucket pays for them.

The inventory rule follows the same shape. For our top 50 SKUs by weekly velocity we hold 8 weeks of cover. For SKUs 51 to 200, 4 weeks. Below 200, lean and accept occasional stockouts.

The pattern is the same in both cases. You define a fixed protected layer, you operate the variable layer aggressively, and you accept that the protected layer is the cost of being able to operate the variable layer without panic. This is also the entire risk management philosophy of the business in one shape. Most operations failures we see in peer businesses come from protecting the wrong layer or failing to define what the protected layer is.

### Customer Service Becomes Catalog Intelligence

For most ecommerce businesses, customer service is a cost center managed for response time. For us it became one of the highest-value operations inputs because we built a routine where every Friday I personally read the prior week's customer service threads. Not to answer them. The team handles that. To look for patterns.

That routine has produced more inventory decisions than any forecast model. A pattern of three customers in one week asking about a discontinued bottle becomes a sourcing project the following Monday. A pattern of returns concentrated in a specific SKU triggers a supplier conversation about batch variation. Customer service is where the future of the catalog reveals itself one ticket at a time.

The operational discipline is the Friday routine, not the customer service software. The software just records what people say. The discipline is reading it weekly with the lens of inventory and supplier decisions.

### What Bootstrapping Forces You To Build

The honest summary of what we built over nine years is that bootstrapping forced us to develop operating instincts most venture-backed competitors never need to. When a 12-month runway is your reality, you cannot afford a quarter of imprecision. The weekly heartbeat, the reversibility framework, the two-bucket cash discipline, the tiered inventory rules, the Friday customer service review. None of these came from a textbook. Each was learned the hard way and then institutionalized to prevent the lesson from being relearned.

Most small businesses I talk to think of operations as the function they will hire for once they get bigger. The lesson from running PerfumeM for nine years is that operations is not a function. It is the business. The earlier you treat it that way and start codifying your operating system, the better positioned you are when scale finally arrives. And the earlier you stop confusing planning meetings with actual operational improvement, the faster the business compounds.

If you are running a business too small to have a dedicated COO, the question is not how to delay the cost of operations. The question is what minimum viable operating system you can run in 90 minutes a week that quietly makes every other decision easier.