Choosing a calculation strategy

When you enable an engagement type on a program, or a metric type on a distributor, Siren asks how to calculate the score for each trigger. That choice is the calculation strategy, and the question underneath it is simple: when a single trigger fires, who should it pay? The answer might be one person, the chain of people above them, or the chain of people below them. Those three answers map to the three strategies, and they are what the rest of this page works through.

Fixed answers the question with one recipient. It emits a single score for the collaborator who actually did the work, which is what most programs and most distributors want. It needs no collaborator group at all, and if one is bound it works regardless of how that group is structured. If you have never thought about spreading a score across more than one person, Fixed is the strategy to pick.

Once you want a trigger to pay more than one person, the direction of the payout decides the strategy, and both cascade options walk the triggering collaborator’s chain up to five layers. Upline Cascade walks toward the top of the chain, emitting a score for each collaborator above the trigger. Layer 1 is the direct upline, layer 2 is their upline, and so on. Reach for it when people above someone should earn from the work that person does, such as management overrides or team-lead bonuses.

Downline Cascade walks the other direction, emitting a score for each collaborator below the trigger so the people someone manages or leads each earn a share when the trigger fires. Layer 1 is the direct downline, layer 2 is their downline, and so on. It fits top-down rollups where a sale by a manager rewards the team underneath them, or team-goal incentives where a manager hitting a target pushes a bonus down to the reps they lead.

Both cascades require a linear chain or parent-child group bound to the program or distributor, and the picker hides them when the bound collaborator group is flat or unbound. See Why some calculation methods disappear for the reasoning, and read What is a cascade first if the idea of layers and walking a chain is new. The picker behaves the same way on the Programs Edit and Distributors Edit screens, so picking Upline on an engagement type and on a metric type means the same thing, walking upline from the trigger and crediting each layer. The two sides read the bound group and the per-layer args from different config locations: an engagement type reads from the program config, and a metric type reads from the distributor config. Each cascade page covers the exact config types in its Configuration section.

The picker is the only guard against an unworkable cascade, and it acts at edit time. If a cascade is saved and the bound group is later flattened, deleted, or unbound, the calc keeps the cascade strategy and pays out nothing rather than reverting to Fixed. See Cascade troubleshooting for the states that produce zero payouts after configuration and how to recover.