[Eridrus]

I am curious where you work; is it in finance?

I'm not sure I believe you when you say "there are no relevant milestones". From a business stand point it doesn't matter what technical milestones are hit, but almost every business in the world tries to determine if employees are doing as well as expected or better and there is often no clear link to the exact monetary value of any given dev work. There are obviously issues here, but I find it hard to believe any compensation structure exists that doesn't, if nothing else but because teamwork is hard to quantify.

If this project isn't expected to take more than a year to get off the ground, it doesn't sound significantly different from just having a very uncertain bonus.

In any case, if you have no preference in how they should be compensated, consider letting them decide how they would like to be compensated; in a bonus driven culture people may be quite happy to take a percentage they think is reasonable, or they may want to keep their existing structure.

If these developers are able to command such a salary elsewhere, I doubt you will have much luck getting them to take a salary structure they dislike simply because the stakes are so high.

[asbestoshft]

Yes, it is finance! Prop trading more specifically.

We could set up milestones such as (1) implement this exchange interface with this latency or (2) implement this trading algo with this latency but really none of that matters if you don't make money in the end. In fact you don't want people to focus too much on that. You don't really know what is important. Maybe latency really isn't very important. Maybe it is the most important factor.

I like the idea of letting them choose but my point is that they will choose to stay where they are unless you say that they can have their current compensation but then saying you get the same bonus you got last year regardless of how this new project works out defeats the incentive nature of having a bonus. Of course they would take that deal as they get the same amount as they would and there is some chance it will work out and they get more.

[Eridrus]

Some people are just completely risk intolerant, but if you can't convince anyone at your company to take a risk on something that could next them 10X their comp, then it seems more likely that they just don't think your idea will work.

One thing to consider is that you might not want the best of the best; their talents may be better utilised ekeing out small but highly leveraged gains on your cash cow. You may be better off with someone who isn't that great, but who can ask the more experienced devs for help.

Also consider that bonuses are not the only way to motivate people, most other industries manage to motivate people just fine without a huge sums.

Alternatively, if your goal is just to motivate people rather than conserve cash, let them get both; let them take the maximum of last year's bonus, or a cut of the new project; if they manage to make the new project work they do even better, if not the company eats the risk.