How much should we pay inventors per disclosure, filing, and grant?

The three-stage pattern works because each payment rewards a different verified event. The acceptance award pays only when the review committee accepts a disclosure for filing, which makes quota-stuffing with thin submissions unprofitable. The filing award marks the real budget commitment. The grant award, typically the largest, lands when the patent issues, sometimes years later, and keeps the inventor connected to the outcome. Equal splits among named inventors keep the math clean and remove the incentive to trim contributor lists.

Two compliance notes belong in the design. Several countries, Germany, Japan, and China among them, have statutory inventor remuneration regimes that are legal obligations rather than perks; if you employ inventors abroad, route the program past employment and patent counsel. And awards are generally taxable compensation, so build the payout mechanics with payroll rather than as ad hoc gift cards.

Cash rewards the engineers who already recognize and document their inventions. It does little for the rest, and published commentary commonly suggests only about a third of engineers ever submit a disclosure. The missing two thirds are mostly stopped by recognition failure and writing burden, which money does not address. The fix for those is process: facilitated capture sessions where engineers talk and facilitators document, and a committee that decides quickly with reasons.

Per dollar spent, fast decisions and visible recognition outperform a bigger check. A filing announcement at the engineering all-hands, a plaque, and a decision in three weeks build more pipeline than doubling the grant award. Treat cash as the floor of the program, not the engine.

Watch for three patterns. Disclosure splitting, where one invention arrives as three submissions to collect three awards: acceptance-gated payment and an alert committee handle it. Name inflation, where managers or bystanders appear as inventors: beyond fairness, inventorship on a patent application is a legal determination with real consequences, so inventorship questions go to counsel, not to politeness. And cross-team resentment, where business units run different award levels for the same work: harmonize the schedule company-wide even if budgets differ.

Some companies add value-based awards later, sharing a slice of licensing revenue or commercial milestones. These can work but introduce valuation disputes; start with the simple milestone schedule and add complexity only when the basic program runs clean.