Don't tie bonuses to individual contributions. Individuals will work to meet their individual KPIs/metrics often to the detriment of the organisation.

If every employee gets a bonus based on the organisational performance then there is a greater incentive for team work and cooperation within the organisation.

I previously worked in a company with this policy. It meant the receptionist got the same bonus as the Principal Engineers. That policy might make some people unhappy if they think they deserve more compensation because they are more educated / smarter / experienced in a specialty. The thing is, the admin ans support staff worked hard to make the place run smoothly. The engineers were almost completely unburdened with flight, accommodation or training bookings. IT equipment was procured and commissioned in days, not weeks/months. If someone was roadblocked, the organisation worked to remove the roadblock.

It was the best company I have ever worked for. Sadly, the company was acquired and the cooperative culture died after a few years.

Other places I have worked are in contrast much more mercenary. Individual performance is measured and used to determine compensation and promotion. Individuals and internal teams actively compete with each other resulting in significant inefficiency. Internal politics dominate business making decisions.

You can still measure performance against KPIs but not tie it to individual compensation. Instead use the metrics to analyse where there might be opportunities for improving efficiency through restructuring / training.

Do you think that this successful bonus strategy would work if it was done as profit-sharing rather than bonus? I guess it is kind of the same thing, but there might be subtle differences. I don't know much about this area but I was always impressed by Balsmiq's profit-sharing ideas.

In some sense it is a profit sharing strategy. The owners will still have strategic spending goals for profits (growth, R&D, acquisition etc). Also, depending on how long the company has been around, the owners may also have to pay down debts taken out to start/grow the company.

That must be balanced against ensuring employees are adequately compensated vs the local market expectations.

Find a metric that is hard to game and aggregates the results you're interested in driving over a relevant period of time.

"Widgets that sold and had no returns logged" is better than "Widgets that passed QC" is better than "Widgets completed" is better than "Widgets started" for a widget assembly line worker.

Perhaps even better is "long-term profit and sustainability of Widgets, Inc" but that is harder to tie back to the individual widget maker. (A possible metric for the last is "arithmetic average of the share price of Widgets, Inc measured between 9 and 15 years from now", but that's hard to use to motivate someone to make more/better widgets this week.)

Use longer term metrics (grin), which means relying on multi-year performance rather than quarterly.