What You Don't Bid, You Renegotiate at Invoice

The sleeve you had to buy a million at a time

A refrigerated-food brand ran a competitive RFQ across its packaging, awarded the business, and started buying. Months later its finance team escalated. The invoices did not match the award.

The clearest case was an egg-oval sleeve. It is not a high runner for them; it moves in modest releases, and the artwork changes often enough that a large standing quantity is a write-off waiting to happen. The awarded supplier was invoicing against a 1,000,000-unit minimum. The category lead on the buy side had an email trail from the month before production start confirming minimums in the 200,000 to 300,000 range. Their answer when the million-unit number came up was flat: "We would never sign up for that. That is why we ran the RFQ."

Two other things had happened in the same period. Several awarded items were being purchased from a second supplier rather than the one that won them. And the person who had owned the supplier relationship had left the company mid-dispute, taking the only informal version of the understanding with him. His last recorded position was that the conversation had gotten difficult and the items might need to move somewhere else.

Nobody was arguing about the awarded unit price. They were arguing about the quantity you had to commit to in order to get it, which is the same argument with a different name.

An award is only as strong as its fields

A competitive event produces a price. It does not produce a contract unless the terms that determine what you actually pay are fields on the sheet: unit price, the quantity band that price holds at, minimum order and minimum release, the index and baseline the price moves against, who eats artwork changes and obsolete stock, and which items and plants the award covers.

Anything that is not a field lives in a note, an email, or a person's memory. Notes get read by the person who wrote them. In a bid-design session I sat in recently, a twenty-minute debate about whether to have suppliers confirm their own price baseline ended on one sentence: the suppliers do not read the notes. So we publish the baseline and tell them to bid off it.

The failure mode is asymmetric, and that is what makes it expensive. The buyer treats the award as a control document. The supplier treats it as a price list. The invoice arbitrates, and the invoice sides with whoever holds the term in writing. Add normal turnover and the informal side of the deal evaporates entirely, because the only copy of it walked out with a laptop.

There is a second-order effect worth naming. When some awarded items get bought outside the award, the volume the supplier priced against stops being real, which hands that supplier a legitimate reason to reprice everything else. Leakage does not stay local. It reopens the award.

And there is a trap specific to sourcing events. In this case the million-unit minimum did not enter as a negotiated term at all. It entered as benchmark data, historical numbers supplied to describe the current state. Data you hand a supplier to explain where you are becomes that supplier's assumption about where you are going, unless a field on the sheet says otherwise.

Design the bid field before you write the RFP

The work happens before the package goes out. A recent bid design for a molded-plastics category, roughly $4M of incumbent spend across two suppliers, ran through it in one sitting.

1. Publish one baseline. The buyer names the resin level and the date every supplier prices against, and movement off it runs through an escalator and de-escalator agreed in advance. Let each supplier pick its own baseline and you never get apples to apples, no matter how carefully you normalize afterward.
2. Give alternates their own lane. Current price for the current spec in one field, alternate price and alternate spec in two more, with room for the supplier to explain the change. An idea that arrives in a field is comparable. The same idea in a cover email is a distraction that arrives after you have already scored the round.
3. Decide what does not belong in round one. That team deliberately held a specialty label application out of the first round rather than let four suppliers each price it a different way and make the round uncomparable. Rebates and incentives got pushed to best and final, so nobody could buy the award with a promise.
4. Terms and conditions travel with the sheet, and additions close before it ships. On another engagement the working rule was explicit: anything you want in the purchase-order terms goes in now, because the terms are being incorporated into the RFP sheets this week.
5. Bid the quantity structure, not just the price. Minimum order, minimum release, price-break bands, and the artwork-change and obsolescence treatment, per item. The refrigerated-food case is the whole argument for this line item.
6. Agree the reconciliation mechanism while you still have leverage. An outdoor-furniture maker mid-event took a roughly $50 per ton corrugated increase across about $700K of spend. That conversation is short when the baseline and escalator are already on the sheet and long when they are not.

My rule after enough of these: if it can move the invoice, it is a field, not a note.

What a well-run award looks like

Every term that can change what you pay appears as a field on the bid sheet, and the note section carries instructions only, never commitments. One baseline is published by the buyer, and price movement runs through an agreed escalator. Minimum order and minimum release sit on the award sheet per item, and they match the release pattern in the last twelve months of actual purchase history rather than the benchmark file. Invoice-to-award variance is a monthly count, off-price lines and items bought outside the award, and every exception has a named owner and a close date. No single person is the only holder of a term; if they resign on Friday, the award is still enforceable Monday.

The event was supposed to end this

The RFQ existed to escape exactly this situation. It produced a better price and left the terms where they had always been, in somebody's inbox, until the person who owned that inbox left. Price is what you bid. Cost is what you did not.