Forecast columns and grid layout

What it does

The Layout step defines every column in the forecast grid — what each one aggregates, how it is calculated, and whether a user can type over it. It is where most of the actual design work in a forecast build happens.

Key facts

Four column types: Rollup (aggregates the rollup entity), Calculated (a formula over other columns), Simple (manually uploaded via Excel, which is how Quota works) and Hierarchy related (an attribute pulled from the hierarchy entity, always single line of text).
The normal starting move is Add from option set > Opportunity > Forecast category > Auto-configure, which builds a column per option set value using Microsoft's defaults. Auto-configure only offers itself when the rollup entity is Opportunity or Opportunity Product.
The out-of-the-box forecast category values are Pipeline, Best Case, Committed, Omitted, Won and Lost. The default grid shows Quota alongside them.
Rollup column settings: Selector (which option set value), Amount field, Date entity, Date field, Data type, Allow adjustments, Show progress compared to quota, Description.
The Date field is what puts a record into a period — usually Est. Close Date. It can live on a related entity, which is how Opportunity Product configurations work.
Calculated columns reference other columns by unique name, e.g. Closed + Commit + BestCase.
A calculated column can only be adjustable if the formula uses addition and subtraction only, references no adjustable columns, no Simple columns, no Prediction column and no Hierarchy related columns, and any calculated column it references is itself non-adjustable.
Data types are Currency, Integer and Decimal. If you mix currency and numeric columns the forecast will not activate until you pick a primary data type — the Activate step shows an error with a Fix issue link that opens the Layout warnings pane.
Field-level secured columns need read access granted to the # Dynamics 365 Sales Forecasting application user, or the column renders empty with no error.
Microsoft documents no maximum column count, but the grid is horizontally scrolled and a wide forecast is unusable on a laptop.

When to use / skip

You do not get to skip this — a forecast without a layout is not a forecast. The real question is how far past Auto-configure you go.

Take Auto-configure and stop, for most clients. It gives you the six category columns plus Quota, correctly mapped, which is what nine out of ten sales organisations actually look at. Add one calculated Total column and you are done.

Go further when the client forecasts on something the categories don't express: weighted pipeline, gap-to-quota, a coverage ratio, or a manually keyed "commit number" that sits alongside the system roll-up. Those are calculated and simple columns respectively, and they are worth the effort because they are usually the numbers the leadership meeting actually runs on.

Custom option sets instead of Forecast category are where projects go wrong. You lose Auto-configure, and Microsoft's own guidance is that you then need a Power Automate flow to keep your custom field in step with opportunity status. That is a permanent maintenance liability for what is usually a naming preference. Push back.

Configuration decisions

Auto-configure with the standard Forecast category, or a custom option set with the flow and mapping burden that comes with it.
Which columns are adjustable. This is a process decision, not a technical one — every adjustable column is a place a manager can override the system.
Whether to add a calculated Total, and whether that total is adjustable (which then locks its source columns out of being adjustable).
Which columns show Show progress compared to quota. Progress bars are what sellers respond to; putting them on the wrong column drives the wrong behaviour.
The primary data type, if the client wants both currency and unit-count columns in one grid.
Whether hierarchy-related columns earn their space — manager name and territory are handy for filtering, but they are text and they widen the grid.

Gotchas

Making a source column adjustable after you have used it inside an adjustable calculated formula is blocked. Decide the adjustability model before you build the formulas, not after.
Non-English characters in a column's Name stop the Unique name field auto-populating. Type the unique name yourself; formulas reference it, so a blank one breaks the calculated column later.
Field-level security on Est. Revenue is common at clients with sensitive deals, and the forecast silently shows nothing rather than failing. Check the forecasting application user's access early.
The Forecast category field on the opportunity form cannot be hidden with Visible by default — you have to remove it from the form. Clients who want the category driven purely by sales stage always ask for this.
Simple columns cannot be populated until the forecast is activated, so the Quota column looks broken throughout build. That is expected.
Column changes on an active forecast do not backfill instantly; give it the recalculation cycle before telling the client it is wrong.

Consultant notes

Demo the grid with progress bars against quota. It is the thing sales leadership reacts to, far more than the column mapping you spent the day on.
Get sign-off on the adjustable column list in writing. "Who can override the number" is a governance question and it will come back at UAT.
Keep unique names short and boring — you will be typing them into formulas, and renaming later is not free.
Before go-live, check one opportunity end to end: category set on the record, correct column in the grid, correct period from the close date.
If the client asks for weighted pipeline, build it as a calculated column over the category columns rather than trying to make Dynamics weight by probability. It is simpler and they can see the maths.

Worth another look if Auto-configure ever extends to custom rollup entities, or if the adjustable-formula rules are relaxed.