Skip to main content

For business-to-business buyers, what is the best way to buy something? Looking at it from the other side, what is the best way to sell something to a business-to-business buyer?

Well, as an economist, I naturally defer to the answer, “it depends.” I need more information to answer.

First, what does it mean to say best?

Second, we need to know, best for whom?

The best outcome for B2B buyers means (as we say all the time around here) “buying the right solution, from the right supplier, at the right price.”  The right solution is the one that has the best problem-solution fit. The right supplier is the one who, at a minimum, is reliable, but is also some person of good character (keeping in mind that a company is an economic person). The right price is not necessarily the lowest one but is one that both sides perceive to be fair. The old rule is that if both sides are unhappy, then you’ve found the right equilibrium.

If we’re talking about the ideal paradigm, we can add two other dimensions. It should not take forever to purchase. It shouldn’t cost a ton of money to get from the point at which we decide that we have a problem to solve to the completion of the purchase of its solution. We need internal consensus when we purchase goods and services. Too often this is slow in coming. With the dawdling, we also see elevated transactions costs related to discovering suppliers and picking a winning vendor.

“Best for whom” can mean very different things based upon positioning. Buyers want value-for-money. Think about the commercials on television that show people paying wildly different prices for the same hotel room or the same flight. How maddening would it be to know that you overpaid? Or to learn later that there was a better solution you could have bought given the need instead of the thing you actually purchased.

The story is very different for sellers. For them, frequently winning means being the guy to get someone to pay twice what other people are paying for the same thing.

You can have a negotiation without an auction. In colloquial language, when people talk about negotiations, this is what they mean. It’s when a buyer discovers an individual supplier and engages in one-to-one discussions about price and terms.

Or, you can have an auction in which one side sets up a structured, simultaneous, competitive process with multiple players in parallel. Auctions are just a different type of bargaining arrangement. Most people don’t think of them as negotiations, but that’s what they are, in practice.

The journal Microeconomic Insights published a paper in October 2023 called “Auctions versus negotiations in the real world.” Economists love experiments. They compare the experience of two sets of landowners in Texas:

“At independence, millions of acres of unsettled Texan land were allocated to a fund. Landowners who purchased their land out of this fund before 1973 were given the right [to] undertake private lease negotiations for minerals beneath the further of their land on the states behalf – in exchange for a share of proceeds. Those that purchased land after 1973 did not get this right.”

The royalties from the minerals underground on this sold land still belonged to the state of Texas. Part of the purchase price when landowners bought the land before 1973 covered the additional right to participate in the economics of leasing access to these underground riches. For example, this being Texas, imagine an Exploration and Production company that would like to lease part of a ranch’s land. The E&P company contracts a lease for the minerals with the lease payments going to the state. For landowners who obtained their property before 1973, they negotiate the lease and receive a portion of the lease payments. For those who bought after 1973, they negotiate the lease and receive nothing.

Here’s what they found:

“In comparing leases signed between 2004 and 2016, we find that negotiated leases would have generated $185,000 more upfront revenue, on average, had they been awarded using an auction. This finding is intuitive; faced with the threat of more competition, it makes sense to expect that E&P companies would bid more in an auction than they offer in a negotiation.

“Our analysis also suggests, however, that auctions allocate assets more productively. Auctioned leases produce roughly 50% more output than negotiated leases. Combining discounted royalty payments and up-front bonuses, we estimate that auctions increase total payments to sellers by about $341,000 per lease.”

But here’s the best part. I’ve added the emphasis.

“Our data suggests it is the quality of a match between firms and leases, rather than simply the quality of the firm, that is the primary source of the productivity gains we measure. Auctions, it seems, bring in more competition than negotiations and ‘pick’ better winners.”


This example is for sellers looking to attract buyers in competition, but the principle is the same for B2B buyers. In the case of a buyer, auctions should be expected to do two things: lower the upfront cost and lower opportunity costs (by picking the solution that maximizes problem-solution fit so we are not living with a sub-optimal solution and the consequent inflated COGS or reduced revenue). Opportunity costs dwarf transactions costs and upfront pricing. Picking the right vendor is everything.

In the end, there is always a negotiation. The difference is with auctions buyers set themselves up for success at the bargaining table.

Now, you could say, EdgeworthBox, that’s all well and good, but it costs money to run auctions. It is expensive for suppliers to generate proposals. You’ve already told me that suppliers make an investment decision and elect often to not participate in auctions because they think they are rigged, having predetermined the outcome for what amounts to a pro forma consulting exercise.

For buyers to get the best outcome in an auction, they have to reduce the costs (for both sides) as much as possible and they have to make it a fair fight. That’s the promise of Generative AI in procurement. While the low-hanging fruit is some vaguely improved version of robotic process automation of invoices, the real value Generative AI will bring to procurement is the transition to truly functional auctions at almost any purchasing price point. If we can make it virtually costless for suppliers to respond, then there is no reason for them not to respond. If we can make it simple to evaluate proposals, then there is no reason not to stimulate competition.


Watch this space. At EdgeworthBox we’ve been collecting data to fine-tune large language models for precisely this purpose. This is all part of our broader effort to help B2B buyers

© 2022 Homework Fairy. | All Rights Reserved.