[object Object]

The CRO asks for a single number in the Monday call: the rolling-quarter forecast. You build a forecast in two hours from the template, demo it Friday, and on Monday three regional managers say their numbers are off. Forecasts feel simple in the wizard and brutal in production because the hierarchy, the rollup, the snapshot timing, and the adjustment policy are four separate decisions that each break trust if wrong.

Hierarchy Is the First Real Choice

Forecasts roll up a hierarchy of users or territories. Pick one and commit. Mixing user-based and territory-based hierarchies in the same forecast leads to double counting on shared accounts and orphan numbers when reps move teams.

User hierarchy: Manager column on systemuser drives the tree.
Territory hierarchy: Territory parent on territory drives the tree.

If your business reorganizes more than once a year, territory is more stable. If your reps own a fixed book of business, user is simpler.

Forecast Columns That Earn Their Place

Default forecast columns are Pipeline, Best Case, Committed, Won, and Quota. Each is computed from a forecast category on the opportunity. Custom columns add work for everyone and rarely add signal. If you must add a column, define it as a calculated field on opportunity first, prove it survives a quarter without complaint, then promote it to the forecast.

Pipeline    = Open opps in period
Best Case   = Opps with category Best Case
Committed   = Opps with category Committed
Won         = Closed-won in period
Quota       = Annual target divided by period weight

The forecast category column on opportunity is the lever. Train reps to update it weekly, or the rollup is built on dust.

Snapshot Cadence

The platform takes a snapshot of the forecast on a schedule. The default is every six hours. For a Monday morning forecast meeting, configure the snapshot to run Sunday at 10pm and again Monday at 7am. Anything more frequent than every two hours starts to load the async server without giving leaders a meaningfully different number.

The snapshot is what powers the trend chart. If you change the snapshot frequency mid-quarter, the trend line bends because the sampling rate changed, not because the pipeline did. Pick a cadence and freeze it for the quarter.

Adjustment Rules

Adjustments are notes a manager types over a child rollup to override the rep number. They are necessary, and they are the single most contested feature in any forecast meeting. Decide the policy before launch:

1. Adjustments are visible to the rep below.
2. Adjustments require a comment.
3. Adjustments expire at quarter close.

Without these three rules, adjustments become a passive aggressive note channel. With them, they become a real tool.

Multi-Currency Forecasts

Forecasts respect the org base currency for the rollup. Reps see their local currency in the line items but the manager rollup converts to base. The conversion uses the exchange rate stored on the opportunity at close, not the live rate. If your business operates across volatile currencies, document this rule and refresh exchange rates weekly through the Currency entity update.

async function refreshExchangeRate(currencyId, rate) {
  await fetch(`/api/data/v9.2/transactioncurrencies(${currencyId})`, {
    method: "PATCH",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({ exchangerate: rate })
  });
}

Layered Forecasts for Predictive

The Sales Premium tier adds a Predictive forecast column powered by historic conversion rates and time-in-stage signals. Layer it next to your manual forecast for one quarter before relying on it. If the predictive column tracks within five percent of actuals, promote it to the primary view. If it does not, keep it as a sanity check.

Permissions and Visibility

Forecasts are scoped to the user and their direct hierarchy. A regional VP sees their tree and below. A skip-level manager who needs visibility further down requires the View All Forecasts privilege. Grant it sparingly. Wide visibility creates pressure to massage numbers upstream.

What to do this week

Lock the snapshot cadence, write down your adjustment policy in three lines, and remove every custom column from your forecast that nobody references in the Monday call. Run one full snapshot cycle and confirm the rollup matches a manual FetchXML pull within one percent.

[object Object]
Share