Skip to content
PlaidCloud

Recursive and Reciprocal Allocations

Some cost models need the target of one allocation to become a source of the next — for example shared services that consume each other (IT serves HR, but HR also serves IT). PlaidCloud models this by chaining allocation steps into explicit, staged passes, not through a single automatic solver.

When You Need Multiple Passes

Section titled “When You Need Multiple Passes”
  • Reciprocal services — two cost pools that consume each other.
  • Layered spreads — divisional costs cascade through a hierarchy, and the lower levels must absorb the upper levels before re-allocating.
  • Iterative balance — the model approaches a stable result after several passes (pool A allocates some to B, B allocates some back, and so on).

How to Model It

Section titled “How to Model It”

PlaidCloud has no built-in “recursion mode” and no automatic convergence solver. You build the recursion yourself by adding a separate allocation split step for each pass and wiring them in sequence:

  1. Run the first pass as a normal allocation split.
  2. For the next pass, build a source that includes the previous pass’s output — typically by unioning or joining the prior result back into the pool — and allocate again.
  3. Repeat for as many passes as your model needs.

Because each pass is an explicit step, you decide how many passes run. The engine does not solve the reciprocal relationship as a single simultaneous equation, and there is no tolerance setting that stops the loop automatically. Add passes until the residual still to be re-allocated is small enough for your purposes.

Tips

Section titled “Tips”
  • Start with a single-pass allocation and confirm the simple model behaves as expected before adding passes.
  • After each pass, check the remaining unallocated residual. If it isn’t shrinking pass over pass, revisit the driver relationships — a genuine circular dependency won’t settle no matter how many passes you add.
  • Every pass reprocesses its source, so recursion is expensive on large datasets. Test on a slice before running across the full source.
Section titled “Related”