June 2026 · 6 min read
What five years of fulfillment data teaches you
Single-order data is noisy. Five-year-aggregate data is informative. Six things we learned from the long log that are not obvious from any single quarter.
Single-order data is noisy. Quarterly aggregates are informative for capacity planning but still season-distorted. Five-year aggregate data is where the durable patterns show up. Six observations from our long log that are not obvious from any single quarter, and that have shaped the operational habits we now consider non-negotiable.
1. Reorder cohort matters more than first-order revenue. Most brand attention goes to converting first-time buyers. Our log shows that lifetime fulfillment value — total volume per brand over multi-year — correlates with the quality of the FIRST repeat purchase more than any other variable. The brand whose customer comes back the first time tends to keep coming back. The brand whose first repeat order has any friction (size confusion, decoration drift, delayed dispatch) loses that customer permanently.
2. Weekday dispatch patterns are remarkably stable. Order volume by weekday is one of the most stable signals in the data. Monday is consistently the heaviest dispatch day for catalogue brands; Thursday is consistently the lightest. This pattern holds across product categories, brand size, and customer demographics. Capacity planning that ignores the Monday spike will under-resource for fifty-two weeks every year.
3. Decoration method correlates with return rate, not customer satisfaction. Sublimation has the lowest return rate of the four core decoration methods. DTG has the highest. But customer-satisfaction scores (where measurable) do not correlate with return rate — they correlate with on-time dispatch. Customers will accept a DTG print that fades faster than sublimation if the order arrived when promised. They will return a flawless sublimation print if the order was a week late.
4. Geographic distance amplifies every other problem. An order dispatched to within 500 miles of our facility has a complaint rate of 0.4 percent. An order dispatched internationally to a non-US address has a complaint rate of 2.8 percent. The intermediate distances scale roughly linearly. International orders amplify every variable: carrier delay variance is higher, customs hold variance is higher, customer-service window mismatch is higher.
5. Brand stability outperforms brand growth in fulfillment efficiency. The brands we serve most efficiently are not the fastest-growing. They are the most stable. A brand with steady weekly order volume of 200 units consumes less of our operational attention than a brand whose volume fluctuates between 80 and 600 units week-to-week, even if the second brand averages higher.
6. Communication frequency predicts retention. Brands that exchange three or more substantive emails with us per month retain as partners at an 85 percent rate at year five. Brands that exchange under one substantive email per month retain at a 31 percent rate. The signal of communication frequency is not about emotional warmth; it is about operational alignment. Partners who are talking to us regularly catch small issues before they become churn-causing ones.
These observations have shaped how we onboard, how we structure capacity, how we cost orders, and how we measure our own performance. None of them are findings unique to our data — we suspect they show up in any sufficiently long fulfillment log. We are publishing them because the operational decisions they imply are easier to defend internally when there is a written reference for the pattern.