Revenue Forecasting Models Explained: Which One Fits Your Business?

Revenue Forecasting Models Explained: Which One Fits Your Business?
Photo by micheile henderson on Unsplash

“How should I model revenue?” is one of the most common questions I get from founders. The answer depends entirely on how your business makes money.

Here’s a breakdown of the most common revenue models and how to forecast each one.

Subscription Revenue (SaaS)

The most common model for software companies. Customers pay monthly or annually for ongoing access.

The Components

Starting MRR/ARR: What you’re beginning with.

New MRR: New customers × Average contract value

Expansion MRR: Existing customers upgrading, adding seats, or moving to higher tiers.

Contraction MRR: Existing customers downgrading.

Churned MRR: Lost customers × Their contract value

The Formula

Ending MRR = Starting MRR + New MRR + Expansion MRR - Contraction MRR - Churned MRR

Tips for Accuracy

  • Model customer counts and ACV separately. It’s easier to estimate “we’ll add 20 customers at $500/month” than “$10K in new MRR.”
  • Track gross vs. net churn. Gross churn is just losses. Net churn includes expansion revenue, which can make churn negative (a good thing).
  • Account for seasonality. B2B sales often slow in December. B2C subscriptions might spike during holidays.

Example

Starting MRR: $50,000 New customers: 25 at $400 average = $10,000 Expansion: $2,000 (upsells) Contraction: $500 Churn: 3% = $1,500

Ending MRR: $50,000 + $10,000 + $2,000 - $500 - $1,500 = $60,000

Transaction-Based Revenue

Marketplaces, payment processors, and platforms that take a cut of transactions.

The Components

Transaction volume: Number of transactions processed

Average transaction value: How much each transaction is worth

Take rate: Your percentage of each transaction

The Formula

Revenue = Transaction Volume × Average Transaction Value × Take Rate

Tips for Accuracy

  • Understand your volume drivers. Is volume driven by number of users? Usage per user? Seasonal patterns?
  • Watch for ATV changes. As you scale, average transaction value might increase (bigger customers) or decrease (smaller customers discovering you).
  • Model take rate changes. Volume pricing, enterprise negotiations, and competitive pressure can all affect your take rate.

Example

Monthly transactions: 10,000 Average transaction value: $150 Take rate: 3%

Revenue: 10,000 × $150 × 0.03 = $45,000

Usage-Based Revenue

Cloud infrastructure, APIs, and consumption-based products.

The Components

Active users/accounts: Who’s using the product

Usage per account: How much they consume

Price per unit: What you charge per unit of consumption

The Formula

Revenue = Users × Usage per User × Price per Unit

Tips for Accuracy

  • Segment your users. Power users and casual users have very different usage patterns. Model them separately.
  • Plan for usage growth. As customers get value, they typically use more. But there’s usually a ceiling.
  • Account for tiering. Many usage models have volume discounts. Your effective price per unit decreases as usage increases.

Example

Active accounts: 500 Average monthly usage: 10,000 API calls Price: $0.001 per call

Revenue: 500 × 10,000 × $0.001 = $5,000

One-Time Revenue

Product sales, implementation fees, professional services.

The Components

Units sold: Number of products or projects

Average price: Price per unit

The Formula

Revenue = Units Sold × Average Price

Tips for Accuracy

  • Forecast units based on pipeline. One-time revenue is harder to predict than recurring. Anchor projections in your actual sales pipeline.
  • Be conservative. Without recurring revenue, each period starts from zero. Build in buffer for deals that slip.
  • Model payment timing. Large one-time deals often have milestone-based payments. Model when cash actually arrives.

Hybrid Models

Many businesses combine models. A SaaS company might have:

  • Monthly subscription (recurring)
  • Implementation fees (one-time)
  • Add-on purchases (transaction)
  • Overage charges (usage)

How to Handle Hybrids

Model each revenue stream separately, then combine:

  1. Subscription revenue: The core forecast
  2. One-time revenue: Implementation, setup, training
  3. Usage/transaction revenue: Overages, add-ons, marketplace fees

This gives you visibility into the composition of revenue, not just the total.

Choosing Your Model

The right model depends on how you charge, not how you want to charge. Ask:

  1. What does the customer pay for? Access? Transactions? Usage?
  2. How often do they pay? Monthly? Annually? Per transaction?
  3. What drives changes in their spending? More seats? More usage? More transactions?

Your forecast should mirror the actual mechanics of how money flows into your business.

From Model to Forecast

Once you’ve chosen your model:

  1. Identify the drivers. What inputs determine each component?
  2. Set assumptions for each driver. How will they change over time?
  3. Document your logic. Why do you expect these changes?
  4. Build scenarios. What if drivers perform 50% better or worse?

The goal isn’t a perfect number. It’s a framework for thinking about how revenue will evolve, and what levers you can pull to change it.


Profitual supports multiple revenue model types out of the box: subscription, transaction, usage, and hybrid. Build your forecast using the model that matches your business. Get started.

Ready to simplify your forecasting?

Join 500+ businesses using Profitual to build investor-ready projections.

Start Free Trial