Skip to main content
Formula issues usually become visible as wrong numbers, blanks where a value should exist, or logic that changes unexpectedly across time or dimensions. When that happens, start with the formula itself before you assume the problem is in a board or report.

​
Start at the source variable

The fastest debugging path is usually:
  1. open the source variable
  2. review the exact formula text
  3. confirm the references still point to the intended variables
  4. check time and dimension context
  5. review the result at both detail level and summary level
This keeps the investigation focused on the place where the logic actually lives.

​
Wrong number of arguments

Some errors come from calling a function with too many or too few inputs. Check:
  • whether the function expects one value or several
  • whether you accidentally left out a closing value in a conditional branch
  • whether the formula became malformed during editing
If the formula is difficult to parse, simplify it before you debug it further.

​
Bad or stale references

References can break when variables are renamed, moved, or replaced. Review:
  • the exact variable names in the formula
  • whether linked or referenced variables still exist
  • whether the formula now points to a broader or narrower variable than you intended
When references are unclear, use Variable references to reset the formula to the smallest reliable building block.

​
Aggregation mismatches

Sometimes the base formula is correct, but the summary meaning is wrong. Common signs include:
  • a total that looks too high because the metric is being summed
  • a quarter or year view that does not match the intended business meaning
  • a formula that behaves correctly by month but not in a higher-level review
When this happens, review the formula together with Time aggregation.

​
Dimension and context issues

Formulas can behave differently when a variable is dimensioned. Check whether:
  • the formula expects a single dimension value or an aggregated result
  • the result is being reviewed on the correct row or item
  • the formula should live on the aggregate row instead of being copied into detail rows
If the model structure changed recently, review Using dimensions with variables before rewriting the formula.

​
Blanks and fallback behavior

Blank handling is easy to miss. Review:
  • whether the formula can return blank
  • whether downstream math expects a number instead
  • whether a date or driver input is still empty
Keep the fallback rule explicit. Do not assume a blank value will always behave the way you want at every step in the model.

​
Cutoff mistakes around the Last actual date

Many formula issues appear right at the boundary between actuals and plan. Check:
  • whether the actual formula and planned formula are both defined correctly
  • whether the formula is using lastActualDate intentionally
  • whether the first planned period behaves differently than expected after close
If the problem appeared after a close update, review the Last actual date before changing the formula.

​
When to use broader troubleshooting

If the formula itself looks correct, the issue may be elsewhere. Move to Formula issues when you need to investigate:
  • source data problems
  • scenario context outside the variable
  • downstream board or view behavior
  • model-wide issues that are not caused by the formula text itself

​
Prevent repeat issues

Most recurring formula errors come from hidden complexity. Shorter formulas, stable naming, explicit cutoff logic, and helper variables reduce the chance that the same issue returns later.