Insights  /  Business

Your training budget fell and your hours rose

The 2026 ATD numbers look like a collapse in training investment. Read the second line and it is something more specific, and more useful to you.

By McKinley Malbrough III, J.D., MS-HRM. Published September 26, 2026.

What do the 2026 numbers actually say?

ATD's 2026 State of the Industry, released May 2026 and reporting 2025 data, puts direct learning expenditure at $846 per employee, down from $1,254 the year before. Over the same period formal learning hours used per employee rose from 13.7 to 16.7.

Spend down a third. Hours up roughly a fifth. That is not a retreat from training. That is a sharp drop in what employers paid per hour of it.

Why did the cost per hour fall so hard?

Three explanations sit on top of each other and all of them are probably true in part: cheap per-seat content libraries, AI-assisted production displacing custom builds, and budget pressure pushing buyers toward self-serve formats that keep the hour count up.

One caveat worth stating, because most write-ups skip it. ATD surveyed 340 organizations for the 2026 edition against 539 the year before, so some of that swing is sample composition rather than the market moving. Treat it as a strong signal, not a precise measurement.

What does that mean if I am buying training?

It means the thing your budget bought more of in 2025 was the format with the worst completion and the weakest transfer. Cheap hours are cheap because nobody is in the room.

It also means your per-employee benchmark moved, and if you are defending a number to finance using a figure from two years ago, they will find the newer one.

What does it mean if I am selling training?

Stop competing on the per-hour line, because you will lose to a content library every time and you should. Compete on the thing the library cannot do, which is a live room where people practice in front of someone who can correct them.

Small employers already know this. Per-learner spend at companies under 500 people runs materially higher than at large enterprises, because the fixed costs of content and platforms have nowhere to amortise. Those buyers are not shopping for cheap hours. They are shopping for the hours that work.

Is the answer just to charge less?

No. The answer is to be specific about what a live hour delivers that a recorded hour does not, and to price the outcome rather than the clock.

I wrote the mechanics of that in how to price a workshop. The short form: a named program with a defined behavior change is a different purchase than a timesheet, and it survives a budget cycle that a timesheet does not.


Keep going

Programs

Four live programs, capped at twelve people.

The playbook

My one page session plan, free.

Store

Programs, courses and merchandise.

Want the price list rather than the theory?

Every seat price, private session rate and reduced rate is published on one page, with no form in front of it.

See the pricing