A founder showed me a spreadsheet once. Three suppliers across the top, fourteen rows down the side: unit price, MOQ, tooling cost, payment terms, lead time, certifications, photos of the factory gate. The sheet was sorted ascending by unit price. He wanted my opinion on column B, which was eleven cents cheaper than column C.
I asked him a different question. Of those three factories, which one would answer your WeChat message at 10pm on a Saturday in month fourteen, when your container is sitting at the port and the buyer in Rotterdam is threatening to cancel?
He did not have a row for that.
That is the pattern I keep seeing in distribution and export ventures, and it has cost more businesses their third year than any tariff schedule ever did. Founders treat supplier selection as a procurement exercise: collect three quotes, book a factory tour, negotiate four percent off, sign. But the counterparty you choose is not a line item in your cost structure. It is your cost structure. It sets your landed price, your lead time variance, your ability to introduce a second product, and therefore your pricing power with every customer downstream. You are not buying units. You are choosing the business you will be running in year three.
The procurement frame fails because the quote is not the price
Here is the mechanic that catches almost everyone on a first order from Asia.
The quote you receive is a sales instrument. It is engineered to win the order, and it assumes a set of conditions that will not survive contact with your actual business: a specific volume, a specific raw material lot, a specific exchange rate, a packaging spec nobody has stress-tested, and an order cadence you have not yet proven you can sustain. When any one of those shifts, and all of them shift, the price moves. Not because anybody lied to you, but because the number was never a commitment. It was an opening position.
I spent years building consumer hardware in China with the Jean-Michel Jarre venture, where one product eventually became eight. In that period I learned to read quotes as a forecast rather than a fact. The spread between the first quoted FOB price and the real, steady-state cost per unit after twelve months is routinely 8 to 15 percent, and it is almost never in your favour unless you built the relationship to defend it. Add the costs nobody quotes: engineering changes after the first production run, the rework on a batch that failed your own QC, the air freight you pay once because tooling was three weeks late, the sample rounds that multiply when the factory is guessing at what you meant.
The founder comparing eleven cents was optimising a number with an error bar of a dollar.
The second failure of the procurement frame is subtler. Three quotes gives you an illusion of a market. What it actually gives you is three counterparties who now know your volumes, your target price, and the fact that you are shopping. In a dense manufacturing cluster, the ones who matter talk to each other. Price discovery is useful. Price shopping, performed loudly, tells the cluster exactly how much leverage it has over you.
Do not chase the best factory. Pick the one where you matter
This is the rule I argue for, and it is the one founders resist hardest.
The best factory in your category, the one with the glossy line, the automotive clients, the ISO wall, the fourteen thousand square metres, will take your order. They will also schedule it behind the customer representing 22 percent of their annual revenue. When the component shortage hits, and it will, you learn what your position in the queue actually is. Not from a meeting. From a slip, then another slip, then a polite note about force majeure.
Run the arithmetic before you sign anything. Estimate the factory’s annual revenue, then estimate what your first full year of orders represents against it. If you are below roughly 2 percent, you are a rounding error, and you will be treated as one in every allocation decision for as long as you stay there. Somewhere between 5 and 15 percent you become a customer that matters: large enough that your delay becomes their problem, small enough that you are not their existential risk. Above 30 percent the dependency flips and becomes dangerous in the other direction, because now their survival is tied to you and they will say yes to things they cannot deliver.
That band is where pricing power lives. Not because you negotiated harder, but because of who takes your call when there is a decision to make about whose order runs first.
A slightly less impressive factory where you are in the top five accounts will beat a world-class factory where you are account number 180. Every time. The capability gap is usually closable with engineering attention. The priority gap is structural and no amount of goodwill closes it.
The three questions I ask before any first order
I do not ask about price in a first serious meeting. Price comes out of the relationship, not into it. I ask three things, and the quality of the answers tells me more than any audit report.
The first: who else do you make this for, and what happened the last time one of them changed their spec mid-run? I am not fishing for a client list. I am listening for whether they describe a process or describe a conflict. A factory that has been through engineering changes with a demanding customer will tell you about the change order procedure, the cost split, who ate the scrap. A factory that gets defensive has either never had a serious customer or never resolved the problem, and you will be the one discovering that.
The second: what is your current capacity utilisation, and where would my volume sit in the schedule? The number itself matters less than whether they answer at all. A partner who will tell you they are at 60 percent and hungry, or at 95 percent and would have to add a shift, is a partner thinking out loud with you. One who says “no problem, we have plenty of capacity” to every volume you name is telling you they will say yes to anything, which means their yes carries no information.
The third, and the one that has saved me the most money: what do you want from this relationship in three years? If the answer is more orders, you have a vendor. If the answer involves moving up the value chain, entering your market, developing their own engineering capability, co-owning a tooling investment, you have a potential partner with an interest in your survival. Those are different businesses. The second one will absorb a bad quarter with you. The first one will requote.
The first PO is a test, not a purchase
Treat your first order as a paid diagnostic, priced accordingly.
I size the first order to be commercially irrelevant and operationally revealing. Small enough that losing the whole batch is survivable, large enough that it runs on the real production line and not as a favour in the sample room. What I am buying is information: do they hit the date, do they flag a problem before it becomes my problem, does the documentation match the goods, does the packaging survive the real container and not a photo studio.
The date is the single most diagnostic variable. A factory that is four days late and told you eleven days in advance is a partner. A factory that is two days late and told you the morning of is a risk you have not priced yet. I have seen lead time variance of plus or minus 20 percent destroy a distribution business that had perfectly healthy gross margins, because the working capital cycle could not absorb it and the retail customers simply stopped forecasting them into the shelf plan.
Build the second order’s terms from what the first one taught you. Not from the quote.
Warm introductions beat platform sourcing, and it is not about trust
Platforms are fine for discovery and terrible for selection. The reason is not fraud. Most listed factories are real. The reason is that a platform relationship starts at zero social cost. If they disappoint you, nothing happens to them beyond a review nobody reads.
An introduction from someone who has placed real orders carries an obligation in both directions. The factory knows that failing you has a consequence in a network they care about. You know that abusing the relationship costs you the introducer. That shared exposure is the only enforcement mechanism that actually works at the scale most founders operate at, because your contract is not worth litigating over a container of goods and everybody at the table knows it.
Across twenty years in Hong Kong and the mainland, in banking, in hardware, in healthcare expansion, almost every durable supply relationship I have been part of started with somebody vouching. Spend your first ninety days building the introduction network rather than running an RFQ. It feels slower. It is the faster path.
What to do differently on Monday
Stop sorting the spreadsheet by unit price. Add two rows instead: our estimated share of their revenue, and who answers at 10pm. Size your first order as a test with a defined pass condition on date adherence and defect rate. Ask the three-year question in the first meeting and listen to what they want, not what they offer. And get introduced rather than discovered.
Then accept the consequence of the whole argument. If the factory sets your cost base, your lead times and your capacity to launch a second product, it is not a supplier you manage from a spreadsheet. It is the operating core of the business you are building, and it deserves the same attention you would give a co-founder.
I write from twenty years of building businesses between Europe and Asia. If your company is facing this, start a conversation.