ARR forecast: bridge new business, expansion and churn

An ARR forecast estimates the recurring annual run rate a subscription business expects to have at a future date. Build it as opening ARR plus new business and expansion, less contraction and churn, with each movement tied to a customer or opportunity and an effective date. Keep the expected forecast distinct from the contracted subscriptions already in force.

Define the balance before projecting it

Choose a forecast date, a reporting currency and a written recurring-revenue policy. For this example, a fictional US software company measures active recurring subscriptions at December 31, 2026, in USD. Implementation fees and other nonrecurring services stay outside ARR.

For flat monthly subscriptions under this policy, annualize the recurring monthly charge by multiplying by 12. This relationship is described in Stripe's explanation of ARR and MRR. Contracts with variable usage, future price steps or temporary discounts need an explicit policy before they can join the same balance.

Forecast closing ARR = Opening ARR + New-business ARR + Expansion ARR - Contraction ARR - Churn ARR

This is a balance at a date. It does not measure the sales recognized over the intervening quarter. Keep the period-based sales projection separate, even when both reports start from the same contracts. A subscription beginning near the forecast date contributes its annualized active charge under the stated ARR policy; it has not delivered a year of services by that date.

Start from the exports you already have

Freeze a September 30, 2026, subscription export for opening ARR. Keep subscription ID, customer ID, recurring charge, billing interval, currency, service dates and contract status. Reconcile its included rows to the finance-approved opening balance.

Bring a renewal list with renewal date, recurring dollars exposed, cancellation notices and expected changes. Add recurring opportunities from the pipeline, including expected service start, recurring contract value and outcome evidence. An opportunity's total deal amount may include services, so do not automatically treat that amount as ARR.

Give every movement an event ID and classify it once. A partial downgrade is contraction; a complete cancellation is churn. Expansion is only the additional recurring amount above the customer's opening subscription. A new customer's initial subscription belongs in new business.

Keep reactivations separate if your company reports them separately, and add that bridge line consistently. The example below has none. Save the forecast version and preparation date so later actuals can be compared with the assumptions known at the time.

Calibrate the assumptions from renewal history

The company uses these illustrative historical observations from comparable, completed cohorts. Loss columns show positive magnitudes. All amounts are annualized recurring dollars, and the historical rates are calculated from the displayed exposures.

These are dollar-weighted outcomes. The last row means activated recurring dollars divided by the eligible pipeline dollars frozen at the historical cutoff. It is not a count of deals won, and it is not an extra factor to apply to an already weighted pipeline.

Keep the forecast horizon and opportunity stages comparable. A rate measured from a late-stage cohort cannot defend an early-stage opportunity with a different expected start date. Override a historical rate when signed evidence warrants it, retaining the original assumption and the reason for the change.

Work the quarter-end bridge

The current exposures and assumptions produce these expected movements. Historical rates above are carried at full precision; displayed percentages use one decimal place.

The resulting bridge groups customer rows for readability. Each existing subscription belongs to exactly one opening group. Stable accounts have no expected movement during this horizon.

The company expects closing ARR of $348,000: $360,000 + $12,000 - $12,000 - $60,000 + $48,000. Expected retained ARR in the at-risk group is a portfolio estimate; it does not assert that each subscription will partially survive.

Reproduce the result in Excel and prove it

Place the bridge headers in row 1, the customer groups in rows 2 through 6 and the total in row 7. Columns B through F are opening ARR and the movement amounts; column G is closing ARR. Enter in G2 and fill down:

=B2+C2-D2-E2+F2

The closing total in G7 is:

=SUM(G2:G6)

This independent bridge check must return zero:

=SUM(G2:G6)-(SUM(B2:B6)+SUM(C2:C6)-SUM(D2:D6)-SUM(E2:E6)+SUM(F2:F6))

If movement rows cover several periods, sum only those effective after the opening snapshot and on or before the forecast date. On a separate Movements sheet, put effective dates in C, movement types in D and positive movement amounts in E. The new-business subtotal is:

=SUMIFS(Movements!E2:E1000,Movements!C2:C1000,">"&DATE(2026,9,30),Movements!C2:C1000,"<="&DATE(2026,12,31),Movements!D2:D1000,"New business")

Extend the ranges to include every exported row. Microsoft documents summing values that meet multiple criteria; the same selection rule should govern the customer detail and the company bridge.

Arithmetic agreement alone cannot find a duplicated subscription included on both sides. Also check unique event IDs, missing customer IDs, movement dates and the opening subscription total. Compare retained customer movements with the cohort retention reconciliation, which covers that diagnostic in detail.

Covirage is not affiliated with Microsoft or Stripe.

Turn uncertainty into an accountable review

If the forecast's new-business activations move beyond December 31, remove the displayed $48,000 addition. Closing ARR becomes $300,000: $348,000 - $48,000. The change is timing, so sales must supply evidence about service starts rather than changing renewal assumptions to restore the original total.

At the next review, replace estimates with effective subscription changes and preserve the original forecast. Finance reconciles the balance, sales explains pipeline timing, and customer success explains renewal outcomes. Keep the approved annual target distinct from the latest estimate; budget versus forecast explains the reporting distinction.

Where it goes wrong

• Total contract value becomes ARR. Separate recurring charges from implementation, services and other nonrecurring items before annualizing.
• Renewal dollars are added twice. A same-value renewal preserves opening ARR. Only its incremental increase belongs in expansion.
• Churn hits the wrong population. Apply losses to exposed recurring dollars, with dated account evidence, rather than every opening subscription indiscriminately.
• Weighted pipeline is weighted again. Store the unweighted recurring exposure and the applied rate so each expected addition can be rebuilt.
• Late starts enter the wrong balance. A signed deal with service beginning after the forecast date cannot enter active closing ARR under this article's policy.

Check your ARR forecast on your own file in Covirage

Bring a sample of your own subscription, renewal and recurring pipeline exports; Covirage maps the columns, and its tools compute and check the ARR movements against your supplied assumptions. The external AI model explains the result with every figure traced to its rows. Create a free Covirage account, then set up your business and bring your file.

See Covirage for SaaS sales teams for the supporting industry context.