The first thousand orders are forgiving; the habits you form while filling them are not. This playbook covers what to write down, measure and formalize between order one and order one thousand, so the systems the next ten thousand orders will demand already exist by the time you need them. It is written for founders and first operations hires working in the fragile window where everything still runs on memory.
Why the first thousand decide the next ten thousand
Every operation scales twice: once when volume grows, and once when the habits that handled the earlier volume finally break. The second scaling is the expensive one, because broken habits are invisible until volume exposes them. A spec that lives in a chat thread works until the supplier asks a question you answered six months ago; a cost model that lives in your head works until a tariff or a freight move moves three numbers at once; a delivery promise invented at order fifty still governs order eight hundred, where it is now wrong. The first thousand orders are the window in which writing these things down costs an afternoon instead of a crisis. Nothing below requires software, staff or capital — only the decision that the business will remember things outside your head.
Orders 1–100: write things down
The first phase has one job: create the records that make every later phase less expensive. Four documents earn their keep from the very first order. A one-page product specification — measurements, materials, tolerances, packaging contents — signed against a golden sample you keep and the factory keeps; without it, reorder two will not match order one, and you will argue about whose memory is correct. A landed cost baseline listing every line item that turns a factory price into a delivered unit, built with the method in our guide to calculating landed cost; this is the number every pricing decision will lean on. A delivery promise grounded in measured lane performance rather than optimism, because the promise you publish at order ten survives into order five hundred by inertia. And a supplier contact log recording who agreed what and when — names, dates, decisions — since supplier-side staff turnover is normal and your record is the only continuity that survives it.
Orders 100–500: measure things
With records in place, the middle phase turns them into measurements. Five numbers, tracked from this phase onward, tell you whether the operation works or merely runs. Defect and return rate by batch, so quality problems attach to production lots rather than to the world in general — the classification habits that make this useful are covered in our guide to quality control without touching the goods. Dispatch time measured against the promise, not remembered against it. Contribution per order — revenue minus product, freight, fees, fulfillment and refunds — because the profit conversation should happen on a number, not a feeling. "Where is my order" contacts as a share of tickets, which measures the tracking loop rather than the customers. And stockout days for the hero SKU, the quiet metric that costs more than all the others combined. None of these need dashboards yet; a spreadsheet updated weekly is the honest minimum, and the discipline of updating it is the system.
Orders 500–1000: formalize things
The last phase converts habits into agreements. The reorder that always happens gets terms: lead times, buffer depth, reorder triggers and capacity expectations written into the supplier agreement — the negotiation structures that get this done without damaging the relationship are in our guide to MOQ and terms negotiation. The hero SKU gets a qualified backup supplier, sampled against the same golden sample, because a single supply line is a failure scheduled rather than a risk debated. The returns policy gets dispositions — what gets restocked, what gets written off, who decides — before the first return wave forces improvised answers. And the weekly numbers meeting becomes real: same day, same six metrics, thirty minutes. These formalizations feel premature at five hundred orders and overdue at three thousand, which is exactly the gap this playbook exists to close.
The habit map by phase
| Phase | Habits to install | Habit to retire |
|---|---|---|
| Orders 1–100 | Written spec and golden sample; landed cost baseline; measured delivery promise; supplier contact log | "I'll remember what we agreed" |
| Orders 100–500 | Batch-level defect tracking; dispatch vs promise; contribution per order; WISMO share; stockout days | "It feels like margins are fine" |
| Orders 500–1000 | Supplier agreement with buffers and triggers; qualified backup supplier; returns dispositions; weekly numbers meeting | "We'll sort it when it breaks" |
The records worth keeping from day one
If you keep only one artifact per row, keep these. Together they form the memory a growing operation runs on, and every later system — inventory software, ERP, a partner's portal — is easier to adopt when they exist:
- Signed specification and golden sample reference for every active SKU.
- Landed cost model with every line item, dated, so changes are visible.
- Inspection and batch records linking quality complaints to production lots.
- Delivery performance log: promised versus actual, by lane, by month.
- Refund and defect log with reasons coded, not paraphrased.
- Supplier quote history, so future negotiations start from evidence.
- Packaging inventory and reorder points, because packaging stockouts stop dispatch as surely as product stockouts.
- The weekly numbers sheet itself — trends are the point, and trends only exist if the history does.
The pattern across all three phases is the same: convert something that currently lives in memory or in a chat thread into something written, dated and findable. Not because documentation is virtuous, but because memory does not scale, does not survive staff changes, and cannot be shown to a supplier, a partner or a new hire when the argument starts.
Frequently asked questions
When should I hire the first operations person?+
When the weekly numbers meeting and the written records exist, or you will be hiring someone to reconstruct them instead of running them. As a rough signal from the phases above: a first operations hire makes sense somewhere in the 500–1,000 order range for most small brands, once the formalizations make the role trainable rather than tribal knowledge.
Is a spreadsheet really enough, or do I need software?+
Through the first thousand orders, a disciplined spreadsheet usually beats software, because the discipline is the system and software purchased before the discipline exists tends to automate chaos. Move to a real inventory or operations system when the spreadsheet stops being updated weekly, when multiple people need concurrent access, or when channel count makes manual sync risky.
When is the right time to qualify a second supplier?+
Once a SKU is demonstrably load-bearing — the one whose stockout would hurt — which typically shows itself between orders five hundred and one thousand. Qualification means sampling against the same golden sample and passing the same verification as the primary; a name in a notebook is not a backup.
What is the most common waste in the first thousand orders?+
Relearning: renegotiating terms the supplier already offered, re-measuring costs that were measured before, re-deciding policies that were decided once and never written down. The testing habits that prevent wasted capital in this window are covered in our guide to testing products before scaling.
