This Year We Turned Down More Orders Than We Took

On the evening of October 9th, around 8:30, I sat in the office going through this year's ledger. Orders taken: 147. Orders turned down: 163.

Xiao Wang from sales scratched his head at those numbers and said, "Boss, are we being too picky?" I said not picky enough — in the rejected column, four orders slipped through that we later lost money on.

This piece is about our three rules for rejecting orders, and how those rules came to be.

We don't take orders with murky credit backgrounds

In May, a buyer claiming to be from Nigeria wanted twelve excavators at a total price of USD 960,000. His email was urgent — shipment had to go out by the end of the month, deposit arriving next week.

I asked for company registration documents and past bills of lading. The registration papers he sent had a crooked seal, and the registration number didn't exist in Nigeria's corporate registry. Then we checked the email domain — registered in March that year, a free mailbox with a shell wrapped around it.

I had Xiao Wang reply asking for 30% prepayment plus an irrevocable letter of credit. We never heard back. We've had nine orders like this this year and rejected every one. A USD 960,000 order makes everyone's eyes light up, but an order where the money never lands is a bigger pit the larger it gets.

Our process now is to check three things for every new customer: registration info, past bills of lading, and payment method. If any one doesn't add up, the order gets parked. These checks cost no money — just two days of time — and two days filters out most of the landmines.

There's a more hidden variant: a real company, a real buyer, but the paying account sits in a third country and doesn't match the contract entity. In March we ran into one — a Tanzanian company wanting payment routed through a personal account in Dubai. I refused. Get tangled in money laundering once and you never climb out.

We don't take orders pushed below the cost line

In July, a Southeast Asian customer wanted twenty loaders. He came in with a competitor's quote, USD 1,100 below our cost line per unit.

Xiao Wang wanted to take it — big volume, thin margins, fast turnover. I laid out the math for him: losing 1,100 per unit means losing 22,000 on twenty units, and that's before counting three months of tied-up capital. That's not thin margin; that's working for someone else for free.

We offered an alternative: ten units, price unchanged, spare-parts kits thrown in. He thought it was too few and walked. In October he came back — the cheaper supplier's machines had arrived, three had hydraulic pump problems, and nobody was answering for it. This time he signed for eight units at our price.

A customer who squeezes on price isn't a bad customer; he just wants to save money. But I need him to understand: on a low-priced order, the seller always claws it back somewhere else — machine condition or after-sales. Once you lay that out plainly, half the customers who walk away come back.

We don't take orders that ask us to alter data

This one is a red line. In September, a customer asked us to change an excavator's hour reading from 9,800 to 4,200 — said it would sell better over there. I said no can do. The contract, the report, and the nameplate have to match.

He raised the offer — an extra USD 1,500 per machine. Five machines, 7,500 in total. I still didn't take it. Xiao Wang asked me why; the money was real.

I told him about a stupid thing we did back in 2019. A road roller had its hours rolled back by the previous owner and we didn't catch it; the customer found out after delivery. We took the machine back and covered the freight — 58,000 yuan all told. The lesson sits right there: fake the data once, and you pay for it with three years of reputation.

Since then we've added one more step when buying machines: the hour reading has to line up with the wear on the engine and hydraulic pump. If it doesn't, we push the price down or refuse the machine outright. This year we screened out nineteen machines with data problems at the buying stage — nineteen landmines defused before they left the yard.

The cost of saying no

Saying no has a price. Of those 163 rejected orders, roughly thirty were clean money — money we could have earned. The profit we gave up adds up to several hundred thousand yuan a year.

But the other page of the ledger reads: four after-sales disputes this year, all tails left over from orders of the previous two years. New orders this year: zero disputes.

Xiao Wang's commission isn't tied to rejections — rejected orders still count toward his workload. I set that rule myself. Otherwise the rules don't hold, and sales will find ways to slip bad orders in.

Better to do less than to do bad orders. That's not a principle; it's a habit bought with losses.