Revenue Forecasting for Subscription Businesses

Subscription revenue is theoretically easier to forecast than one-time revenue because you start each month with a base of recurring customers. In practice, most SaaS companies still get their forecasts wrong by a wide margin. The error usually comes not from the math but from the assumptions — overestimating new sales, underestimating churn, or ignoring expansion and contraction entirely. Here is how to build a forecast that is honest, actionable, and useful for planning.

The Building Blocks of a Subscription Forecast

A subscription revenue forecast has four components, each of which must be estimated separately:

1. Starting MRR: Your current monthly recurring revenue. This is a known number, not an estimate.
2. New MRR: Revenue from customers acquired during the forecast period.
3. Expansion MRR: Additional revenue from existing customers (upgrades, add-ons, seat expansion).
4. Lost MRR: Revenue lost through churn (cancellations) and contraction (downgrades).

Forecasted MRR for any month = Starting MRR + New MRR + Expansion MRR - Churned MRR - Contraction MRR

Each subsequent month uses the previous month's ending MRR as its starting point. This creates a rolling forecast that compounds over time — which means small errors in assumptions compound too.

Forecasting New MRR

New MRR is typically the hardest component to forecast because it depends on marketing, sales execution, market conditions, and pipeline conversion — all of which are variable.

Method 1: Pipeline-Based Forecasting

If you have a sales team, build the forecast from the pipeline:

• Count each deal in the pipeline
• Multiply by stage-specific win rates (e.g., demo completed = 30% close probability, proposal sent = 50%, negotiation = 75%)
• Multiply by expected deal value
• Sum across all deals to get expected new MRR

This method works well for the current quarter but becomes unreliable further out because the pipeline has not yet been built. For quarters 2-4, layer in assumptions about pipeline generation rate based on historical trends.

Method 2: Growth-Rate Extrapolation

For self-serve products or when detailed pipeline data is not available, use historical growth rates:

• Calculate the average new MRR per month over the last 3-6 months
• Apply a growth or decay rate based on trends (is new MRR accelerating or decelerating?)
• Adjust for known upcoming changes (new marketing campaigns, product launches, seasonal patterns)

What to Watch Out For

Teams consistently overestimate new MRR. This is part optimism, part incentive alignment (the sales plan always assumes the best case). Counter this by:

• Tracking forecast accuracy: compare what you forecasted last quarter to what actually happened
• Using at least two forecast scenarios (base case and conservative case)
• Separating pipeline-backed forecast from aspirational targets

Forecasting Churn

Churn is more predictable than new business because it comes from a known population (your existing customers). The key is using the right churn rate and understanding its drivers.

Which Churn Rate to Use

Use your trailing 3-month or 6-month average gross churn rate as the baseline. Do not use a single month — churn is noisy month-to-month, and a single bad month (or good month) will distort the forecast.

Distinguish between:

• Voluntary churn: Customers who choose to leave. Influenced by product satisfaction, competitive alternatives, and price-value perception.
• Involuntary churn: Customers who leave due to payment failures, expired cards, or administrative issues. Often fixable with better dunning processes.

Churn Adjustments

Adjust the baseline churn rate for known factors:

• Contract renewals: If you have annual contracts, large renewal cohorts create predictable churn risk windows. If 20% of your ARR renews in Q4 and your renewal rate is 85%, you can estimate the Q4 churn impact precisely.
• At-risk accounts: Your customer success team likely knows which accounts are at risk. Factor in their win-back probability.
• Product changes: Major pricing changes, feature removals, or platform migrations can temporarily increase churn. If you are planning one, adjust upward.
• Seasonal patterns: Some businesses see higher churn at year-end (budget resets) or in summer (reduced usage in certain industries).

Forecasting Expansion and Contraction

Expansion MRR (upgrades, seat additions, usage growth) is often the most overlooked forecast component, which is unfortunate because it can be the difference between a flat forecast and a growing one.

Expansion Patterns

Look at your historical expansion data:

• What percentage of existing customers expand each month?
• What is the average expansion amount per expanding customer?
• Is expansion accelerating or decelerating?
• Is expansion concentrated in certain customer segments or products?

Apply the historical expansion rate to your forecasted existing customer base each month. If 3% of your customers expand each month with an average expansion of $200 MRR, and you have 500 customers, forecast $3,000 in monthly expansion MRR.

Contraction Patterns

Contraction (downgrades) follows similar patterns. Track the historical contraction rate and apply it forward. Contraction often increases during economic downturns as customers look to cut costs without fully canceling.

Building Scenarios

A single-number forecast is always wrong. The question is how wrong. Scenario planning gives you a range:

Base Case

Use trailing averages for all components without aggressive adjustments. This is your most likely outcome assuming current trends continue.

Optimistic Case

Assume new MRR 20% above base case, churn 20% below, and expansion 20% above. This represents the outcome if sales execution is strong, a new feature drives retention, or a marketing campaign outperforms.

Conservative Case

Assume new MRR 30% below base case, churn 30% above, and expansion flat. This represents the outcome if the market tightens, a competitive threat materializes, or execution slips.

Present all three scenarios to leadership and the board. The range communicates uncertainty honestly, and the distance between optimistic and conservative indicates how confident you can be in the forecast.

The Rolling Forecast Approach

Static annual forecasts go stale quickly. A rolling forecast updates every month:

1. Replace last month's forecast with actual results
2. Update assumptions for the remaining months based on new information
3. Add a new month to the end of the forecast period
4. Recalculate all three scenarios

This approach means your forecast is always based on the most current data and assumptions. It is more work than a set-it-and-forget-it annual plan, but it is far more accurate and useful for decision-making.

Tracking Forecast Accuracy

A forecast is only useful if you track how well it predicts reality. After each month closes:

• Compare forecasted MRR to actual MRR
• Break down the variance by component: was the miss due to new MRR, churn, expansion, or contraction?
• Investigate the root cause of large variances
• Adjust your methodology based on what you learn

Track forecast accuracy over time as a metric itself. If your forecast is consistently 15% too optimistic, apply a systematic correction until the root cause is addressed.

Common Forecasting Mistakes

Hockey Stick Assumptions

The forecast shows modest growth for the first two months, then suddenly accelerates in month three based on a new product launch, a marketing campaign, or the arrival of a new sales hire. These inflection points rarely happen on schedule. Use conservative ramp-up assumptions: new hires take 3-6 months to reach full productivity, new products take time to find market fit, and campaigns need optimization before they scale.

Ignoring Seasonality

Many subscription businesses have seasonal patterns they do not account for. B2B SaaS often sees slower new business in December and August. E-commerce subscriptions surge in Q4. Consumer apps may see seasonal usage patterns. Look at 12+ months of history to identify patterns and incorporate them into the forecast.

Using Gross New MRR Without Netting Out Churn

Excited about forecasting $50K in new MRR next month? Do not forget the $30K you will lose to churn and contraction. Net new MRR is what matters for financial planning, and it is usually much less exciting than the gross number.

Not Forecasting Cash, Just Revenue

MRR is recognized revenue, but cash flow timing can differ significantly, especially with annual contracts. If you collect annually upfront, cash is front-loaded and recognized revenue is spread. If you bill monthly, cash and revenue align. Make sure your cash flow forecast matches your billing model, not just your MRR forecast.

Tracking Revenue Metrics in clariBI

To track and monitor your revenue metrics in clariBI:

1. Connect your billing data to pull historical MRR, churn, expansion, and contraction data. See the data source connection guide.
2. Build a historical trends dashboard showing the trailing 12-month trend for each MRR component. Use the AI assistant to ask questions like "What is our average monthly churn rate over the last 6 months?"
3. Create an MRR components dashboard that breaks down starting MRR, new, expansion, churn, and contraction side by side. Update it monthly to compare against your planning assumptions.
4. Set up goal tracking to monitor MRR progress over time. clariBI's goal tracking feature lets you set milestones and track actuals against targets. See the goal tracking guide for setup.

The purpose of a revenue forecast is not to predict the future precisely — that is impossible. The purpose is to make the assumptions that drive your business explicit, test them against reality each month, and give leadership a reasonable range of outcomes for planning. A forecast that is honest about uncertainty and regularly updated with actual data is infinitely more useful than a precise-looking number that was set once and never revisited.