The $300 purchase order

I started MyAmici in 2005 because I thought labs were paying too much. Amazon was where you bought books. eBay was where everyone believed they were about to become a millionaire. I looked at both and thought a laboratory needs something like this.

I assumed the problem was price. What I found in the Understand phase changed the entire company, and it’s the single most useful thing I know about this category.

The shape of lab spend

Across the buying we see on our own platform, hundreds of laboratories, roughly 2,000 suppliers, around 20 million products, two patterns hold with remarkable consistency.

Now hold those two numbers together, because the implication isn’t obvious until you do.

Why negotiation does not work here

A scientist spends an hour getting quotations for a one-off item, compares them, and saves $10.

That hour cost the business somewhere between $60 and $150 in salary, before you count the experiment that didn’t advance while it was happening. The activity that looks like commercial diligence destroyed value.

This isn’t a failure of effort or of commercial instinct on the scientist’s part. It’s arithmetic. The economics of laboratory procurement are genuinely inverted compared with almost every other category, because the transaction values are small, the tail is enormous, and the people doing the buying are among the most expensive in the building.

There’s a well-documented general version of this. McKinsey’s work on tail spend finds that across industries, the long tail typically accounts for 80 to 90% of purchased items but only 10 to 20% of total spend, while consuming a hugely disproportionate share of procurement effort. Laboratories are that pattern in its most extreme form.

There’s also a nice piece of evidence on the friction this creates. A study across 25 universities by the Southern Universities Purchasing Consortium found that 24% of purchase orders were rejected for missing competing quotations, and that processing time was seven times higher where quotation rules couldn’t be satisfied. More than a fifth of respondents said they could source the item more cheaply themselves than the compliant process delivered. That’s a control designed to save money, costing money.

The two populations inside your spend

The mistake almost every lab makes is treating lab spend as one thing. It’s two, and they need opposite treatments.

The routine, repeat-bought core is high throughput, predictable and standardizable. It should be specified once, priced once, and then reordered without a human being involved in noticing, deciding or approving. Every manual touch on this population is waste, because you already know the answer.

The long tail is the opposite: unpredictable, individually low value, frequently urgent, and often genuinely specialist. You can’t standardize it and you can’t negotiate it item by item. What you can do is remove the friction, make it possible for the person who needs it to find it and order it in minutes at a price that was agreed before they went looking.

How the sectors differ

The balance between those two populations varies more than people expect.

In discovery research, the tail dominates, a very high proportion of what gets bought is unique to a project. But underneath it sits a routine core that’s usually being managed as though it were tail spend, which is where the easy gains are.

Contract research organizations have the opposite profile: high throughput of routine consumables, plus a need to source unfamiliar items quickly when a client project demands it. Two distinct requirements, and running them through one route serves neither well.

In GMP manufacturing, a defined portion of spend is fixed by validation or by customer specification and genuinely cannot be moved. That constraint is real. But it rarely covers everything, and the general consumables, PPE and laboratory supplies sitting outside it tend to receive very little attention precisely because all the governance energy goes to the regulated part.

Medical device businesses have two supply problems in one building: lab consumables behaving like any other lab, and device components requiring genuine strategic supplier management. The risk is applying the thinking from one to the other.

Aggregation

If individual negotiation doesn’t pay, the alternative is aggregation: agreeing prices once, across as much of your spend as possible, rather than transaction by transaction. That’s the logic of a group purchasing organization, and since we operate one I’m obviously not a neutral party.

What this means for how you design the process

The Reimagine question for lab procurement isn’t how do we make purchasing faster. It’s: given that most of what we buy we’ll never buy again, and that the average transaction is worth about as much as an hour of the buyer’s time, what should this process look like at all?

The answer we arrived at, and built a company around, was that the negotiation shouldn’t happen per transaction. It should have happened already, once, on behalf of everyone, so that by the time a scientist needs something, the commercial work is finished and all that’s left is finding it and ordering it.

You don’t have to arrive at our answer. But you do have to ask the question, because almost every conventional procurement playbook was written for a category with a small number of high-value, repeated transactions, and that’s precisely the opposite of what a laboratory buys.