Amtrak Paid for Hours With No Measure of What They Produced

← Back · ACT II · Chapter 7: Value Is Not Promised; It Is Proven

In February 2023, Amtrak awarded a technology services contract for supplemental staff on its eCommerce development projects. The maximum commitment was about $8.9 million. In August 2026 the company’s Office of Inspector General published an audit of how that contract was overseen, and most of the coverage I have seen since is about the paperwork.

The paperwork findings are serious. Of the 297 invoices the supplier submitted during the review period, “only 1 included all of the required supporting documentation.” The auditors questioned $7,753,279.24 as unsupported, which they describe as nearly all of the money the company spent on the contract. They note that the company did receive technology services; what was missing were the timesheets and reports that would show the invoices matched the work.

The finding I would want a buyer to read sits further in, and it is easy to skip.

What the Auditors Had to Work With

The audit set out to look at cost, schedule and performance. On two of those, the contract gave the auditors little to test against. The report says the contract “did not have explicit performance measures” and “did not have explicit schedule milestones,” and that the company “does not have a standardized process for documenting supplier performance.” So the auditors went back to the original solicitation to work out what had been expected at the start, and they interviewed company staff who knew the contract.

Here is how the report records the answer: “Company stakeholders told us they have not identified any issues with the quality or timeliness of the supplier’s overall performance.”

For a contract worth several million dollars, that sentence is most of what the report can say about performance. It tells you nobody raised a problem with the supplier overall. In a footnote, the auditors add that the company’s technical representative told them individual contractors who were not performing were replaced, and that they found no record of that process in the contract files. Whether the work was good is a separate question, and the file cannot answer it. The developers may have done excellent work, and nothing in the report suggests otherwise. But no one, the supplier included, can show it, because the contract set no measure of what good work would produce.

Hours Were What the Record Could Count

According to the report, the request for proposals first asked suppliers to price in “all costs of supervision, labor, materials, tools, supplies, insurance, bonds and any other items necessary,” the way a fixed-price deal is written. Later in the same solicitation, the company asked for bids on set staffing levels, billed on actual labor hours, which is how time and materials works. The contract kept the fixed-price terms anyway. The auditors do not fault paying by the hour as such. They say the services were better suited to that kind of contract, and they call the result a misalignment. Their summary states the consequence: “Contract terms may have limited the company’s ability to hold the supplier accountable.” A fixed-price contract, they add, would have needed “deliverables tied to specific milestones,” and this one had none.

They name the incentive as well. Paying on actual labor, the report says, “increases a supplier’s incentive to maximize the hours it bills instead of only the hours it needs to meet the contract’s deliverables.”

So the supplier was paid for hours, and the contract asked for weekly reports describing what each person worked on. That makes it a record of effort: time billed and tasks worked. Nothing the report describes in it pointed at whether the eCommerce projects got what they needed, so when the auditors arrived, what they could test across every invoice was whether the effort was documented. The file failed that test on 296 of 297.

The auditors are careful to say their results apply to this one contract and not to the company’s contracting in general, and I take them at their word. The reason to read the report anyway has little to do with one railroad. A buyer who signs for effort and leaves the result unmeasured tends to end up with the same kind of file: timesheets, complete or not, and nothing about what the time produced.

The Recommendations Follow the Same Record

The audit makes five recommendations. Four deal with invoices, the questioned costs and offshore labor. The fifth reads: “Assess whether to amend this contract to better hold the supplier accountable for its performance.” The auditors suggest either adding defined milestones or changing the terms to match the work. Management agreed to make that assessment, and then stated that it “does not believe that amending the current contract will add value.” It plans to keep monitoring the supplier through closeout and to document lessons learned for future sourcing.

There may be good reasons for that call. It still means the one recommendation touching performance was aimed at a contract with no explicit performance measures in it, was answered with a decision not to amend, and is headed for a lessons-learned document for whoever builds the next one.

What the Buyer Writes Before Award

Better invoice review would have caught the missing timesheets, and it still would not have told anyone whether the money did what it was spent to do.

Chapter 7 of The Value Standard is titled Value Is Not Promised; It Is Proven. This audit shows what proof needs first, which is a result written down before the work starts so there is something to prove against.

That part belongs to the buyer, who writes it before award and keeps it afterward. It covers what should be different when the work is finished, what evidence will show it, and who produces that evidence. For a staff augmentation contract it can be modest, like the releases the team was meant to ship or the backlog it was meant to clear. It needs to exist somewhere an auditor three years later can find it without rebuilding intent from the solicitation. An audit right will not supply it afterward, because an audit reaches the invoice and stops short of the result, which I wrote about in Your Contract’s Audit Right Can’t Tell You If the Purchase Worked.

A good supplier gains from this as much as the buyer does. On this contract, whatever the developers delivered is recorded as an absence of complaints. A supplier doing strong work has every reason to want the result written into the contract, because that is the only way the work shows up as more than hours billed.

There is a short clause for exactly this, the Verification Line, with a Word copy on the Guides page. It puts the evidence for the intended result, who produces it and by when into the contract before signature. It pairs with a document the buyer writes earlier, before suppliers bid, stating the result the purchase has to produce and how it will be measured. That document is a Request for Value.

Amtrak signed for hours, and three years later the hours are the only thing its record can be tested on.

A later post takes up what happens after the measure is set, when the people watching the contract change: Kansas Paid $2.8 Million and Plans to Return to the Old System.

References: Amtrak Office of Inspector General, Procurement: Company Can Significantly Improve Its Oversight of a Technology Services Contract, OIG-A-2026-009, August 28, 2026.

Keep reading: ACT II · Chapter 7: Value Is Not Promised; It Is Proven