Try a Driver What-If

An Allocation Split step shares an amount between members in proportion to a driver, such as headcount. The What if window shows how that split would change if some driver values were different, or if the step split by another column, from its driver table or from another table in the project. Each scenario is calculated to the cent from what the step allocated on its last run.

Open the Window

You can open it for an Allocation Split step that has an output table, if you can read steps and tables in the project:

On the workflow canvas, right-click the step and choose What if….
In a workflow’s step list, click the flask icon in the step’s What if… column.

What the Window Shows

From top to bottom:

A summary of every scenario.
Notes about the step, and about anything that limits the figures. The sentences below them refer back to these notes.
The scenarios. Click a scenario’s name to make it the active one. Under them, Saved scenarios: lists the scenarios saved for this step.
The driver list: each driver line, with its driver value now, its share now, and what it was allocated on the last run.
The active scenario’s lines, then its Member list, showing how much each member’s amount would change, and its breakdown if you choose one.
Later steps, when Weave can follow a step that reads this step’s output, this step doesn’t round what it writes to whole numbers, and the active scenario has its figures.
Comparing the scenarios, when more than one can be calculated.
Other drivers, after you click Try other drivers.

Add a Scenario

Click Add a scenario. The new scenario is named Scenario 1, Scenario 2 and so on; use Rename… to give it your own name and Remove… to delete it. You can have up to 10 scenarios.

A scenario either changes driver values or splits by another column.

Change Driver Values

Type a new value in the driver list’s New column (for example, New headcount). The figures are recalculated a moment after you stop typing. To go back to a line’s current value, delete its new value. A scenario can have up to 20 new values.

Split by Another Column

Choose a column under Split by another column:. The list offers the driver table’s other numeric columns. After you click Try other drivers, it also offers the columns Weave found in other tables, tried or not, each shown as the column name followed by from and its table’s name. While a column is chosen, the scenario’s new values are set aside, and they come back if you choose None again. To try both, add another scenario.

If a column can’t be used, the scenario’s lines say why.

Save a Scenario

Click Save…, beside Rename… and Remove…, and give the active scenario a name. Everyone who can use What if on the step can see saved scenarios. If a scenario with that name is already saved for the step, you’re asked before it is replaced. Save… is there only if you can change the project.

A saved scenario keeps only its inputs — its new values, or the column it splits by and that column’s table — and is calculated again each time it is opened, from the step’s latest run. Changes you haven’t saved are lost when you close the window.

To use a saved scenario, choose it under Saved scenarios: and click Open as a new tab. Changes you make to the tab don’t affect the saved scenario until you save it again. Rename saved… and Delete saved… change the saved list itself, and like Save… they’re there only if you can change the project. Deleting can’t be undone.

A scenario saved by an AI assistant can hold changes the driver list can’t show, such as one change to every period of a member, or a change by a percentage. When you open it, the window says how many, includes them in the figures, and keeps them when you save.

Try Other Drivers

Click Try other drivers to see what would happen if the step split by each of the driver table’s other numeric columns. The Other drivers section lists each column it tried, the one that would move the most first, and says why any column couldn’t be used. Click Add as a scenario on a column to try it as its own scenario.

Weave can’t tell a code or an ID stored as a number from a measure, so check that a column is a measure before relying on it.

Drivers from Other Tables

Under From other tables, the section also tries numeric columns from other tables in the project. A table is tried only when it has every key column of the driver table, by the same name, each holding the same kind of value. Weave leaves out any table built from an allocation step’s output, since a driver made of allocated figures would feed an allocation back into the split, and the section names the tables with matching key columns that it left out. Weave tries up to 8 columns from other tables, taking the tables in alphabetical order, and names any it didn’t try. You can still choose those under Split by another column:.

Having the same key columns doesn’t make a column a sensible driver, and moving more doesn’t make it a better one. Check what a column measures, and that it’s counted at the same level of detail as the current driver, before relying on it.

If a step in the project reads or writes a table Weave can’t match to one in the project, Weave can’t tell which tables hold allocated figures, so it doesn’t try other tables at all, and the section says so.

Find a Row

The driver list shows its first 1,000 lines, but every scenario’s figures include all of them. Type in Find a row to list the lines whose key values contain what you typed, ignoring case. The count under the box says how many lines match. Click Clear search to go back to the full list.

Break Down Each Member’s Change

Once the active scenario has a Member list, choose a column under Break down by:, just below it, to see how much of each member’s change falls on each value of that column. For example, choose an account column to see which accounts a cost center’s gain comes from. The Breakdown by table shows this for the active scenario. Your choice stays when you switch scenarios. Choose No breakdown to hide the table.

The list offers the columns of the step’s output table that come from the rows being split. It leaves out the member and group columns, the driver, the figures the step writes (such as allocated amounts and shares) and system columns. If a breakdown can’t be calculated, a sentence under Break down by: says why, and the member list still shows each member’s total.

The table shows at most 1,000 rows, starting with the members whose change is largest. When there are more, the scenario’s lines say so, and the last member shown may be missing some of its rows.

Carry a Scenario into Later Steps

When Weave can follow at least one later step that reads this step’s output, this step doesn’t round what it writes to whole numbers, and the active scenario has its figures, the Later steps section offers Carry “Scenario 1” into later steps, named for that scenario. Click it to see what that scenario would do to the tables those steps write, and to the tables after them, as far as Weave can follow them. Nothing is re-run and no table is written.

The section then shows what the scenario would change in the last tables Weave reached, and in any table that feeds one Weave couldn’t calculate, then one line for each later table: the step that writes it, and how much its values would rise and fall, or why Weave didn’t calculate it. Under the line for each last table, a list shows that table’s largest changes, one row per line of the table, with how much the changed column would rise or fall.

If the scenario’s figures leave out any group, if the scenario changes too many lines to carry, or if Weave can’t read what it needs, the section says so instead.

Weave follows Table Extract, Table Union All, Allocation Split (which the window calls Split (No Target Filters)) and Allocation By Assignment Dimension (which the window calls Split using Dimension) steps, and only where their saved settings let it calculate the change exactly. A later Allocation Split step is followed only where the shares it would apply now are still the ones it applied on its last run. A later Allocation By Assignment Dimension step reuses the shares from its last run, but only where its rules, as saved now, would split every changed line the way they did then, and its driver table hasn’t been updated since that run. Weave doesn’t check changes to the hierarchies those rules take their targets from. For a step it can’t follow, the section says why, and the tables after that step aren’t calculated. Weave also has a time limit; a table it didn’t reach in time is named.

Carrying reads every later table again, so Weave does it only when you ask, and for one scenario at a time. Your next change clears it; click the button again to carry the scenario as it is now.

Read the Results

Zeroed in this scenario marks each driver line that the active scenario would allocate nothing to.
The Member list adds each member’s groups together, so a gain in one group and a loss in another offset each other. The scenario’s lines say when that makes the member figures add up to less than the total moved.
Later steps are calculated only when you ask. The notes say how many later steps read this step’s output, and how many of them Weave can follow from their settings. See Carry a Scenario into Later Steps.
The figures cover every period. There is no period choice in the window yet.