Most revenue forecasting software is quick to set up for a reason. It asks for a starting revenue, a growth rate and a churn rate, which are the three numbers the forecast was supposed to tell you. Then it draws a confident line, and that line is only your own assumptions carried forward.
A subscription brand does not have to forecast that way, because the schedule of upcoming charges already exists. A forecasting tool is a point solution, and it is only as good as the foundation it reads from. This guide covers how to tell which kind of tool you are looking at, what it should do for a subscription brand, and how to backtest it.
How Revenue Forecasting Software Works, Top-Down or Bottom-Up
Revenue forecasting software for a subscription business is either top-down or bottom-up by design. A top-down tool asks you to type a starting revenue, a growth rate and a churn rate. A bottom-up tool reads the schedule of what is about to bill, nets it for failures and churn, and models new subscribers separately.
How can you tell a forecasting tool is top-down?
The setup screen is the first tell. A tool that produces a forecast before it has connected to your billing app has not read your schedule, so the line comes from the rates you typed. The second tell shows up in use, where nudging the growth rate by a point moves the whole forecast and nothing else in the tool changes. A team in that position often ends up reforecasting every month.
A tool that reads the schedule
A bottom-up tool connects to the subscription app first, and everything else hangs off the charges already scheduled there, at the price each subscriber is billed. The build itself is covered in how to forecast subscription revenue, so this guide stays on judging the software. Unit demand is a separate job, covered in demand forecasting.
Why the difference shows up in the first miss
Both kinds of tool miss, and the difference shows up on the day they do. Take a quarter where a top-down tool forecasts $1.512M and a bottom-up build forecasts $1.354M, the figures from that forecasting guide. The quarter closes at $1.230M, so the top-down tool missed by $282K and can only offer a growth rate to adjust. The bottom-up build missed by $124K and shows where.
| Layer | Forecast | Actual | Gap |
|---|---|---|---|
| Committed base | $1.152M | $1.080M | -$72K |
| New subscribers | $202K | $150K | -$52K |
| Bottom-up total | $1.354M | $1.230M | -$124K |
| Top-down total | $1.512M | $1.230M | -$282K |
Illustrative figures, built on the quarter in the forecasting guide.
The committed base came in light because more rebills failed than assumed. New subscribers came in light because fewer of them survived to the first full-price rebill.
What Forecasting Software Should Do for a Subscription Brand
A subscription brand needs forecasting software that starts from scheduled charges at the price actually billed. It should net the base for failures and churn separately and model new subscribers on their own survival. It should also tie the result to cash timing and let you test a price or churn change first.
Start from scheduled charges at the price actually billed
The base is every charge scheduled in the quarter, at what each subscriber pays after intro pricing and recurring discounts. Revenue forecasting software that starts from list price overstates the base. The committed base is the first input, covered in your committed rebill revenue as a forward number.
Net the base for failures and churn, separately
A declined card and a cancellation are different events with different fixes, so a single churn rate hides which one moved. Subscription revenue forecasting works better with a rebill rate and a cancellation rate as two inputs, which is the split in voluntary vs involuntary churn.
Model new subscribers with their own survival
Intro pricing makes a first order partial revenue until the first full-price rebill, so the software needs a survival curve to that point. Adds modeled by channel and offer show which one is failing, where a blended rate cannot.
Does the software tie the forecast to cash timing?
Subscription forecasting that stops at a quarterly total cannot show the month the account runs short. A purchase order pays a deposit, a balance and freight on different dates, which is the question behind fund peak season from committed revenue.
Can you test a price or churn change before you make it?
Changing a price or adding a save offer changes churn, and the effect arrives months later. Software that lets you change one input and watch revenue and contribution move shows whether a change pays. The method is in model the revenue impact of a churn or price change before you make it.
Backtest Revenue Forecasting Software Before You Rely on It
A backtest runs the software on a past quarter and compares its forecast with what happened. It only counts when the inputs are frozen as they stood on the first day of that quarter, because a growth rate chosen after the fact can fit any history.
How do you backtest forecasting software?
In practice, the finance lead picks the last two closed quarters and asks each vendor to forecast them from the data as it stood on day one. Revenue forecasting software that is bottom-up can do that if it keeps the schedule as it stood then. A top-down tool can only repeat the growth rate its user types in.
Note: Check that the backtest's actuals are the same revenue the subscription revenue forecast predicts. Net revenue compared with a gross figure shows a miss that never happened, and a KPI master settles which one counts.
Which layer moved when it missed?
The useful demo request is last quarter's miss, opened up. In a bottom-up tool each layer ties back to its platform, so a slipped rebill rate can be checked against the billing app. That traceability is lineage, and a top-down tool can only offer a new growth rate. Why forecasts drift is covered in why your revenue forecast drifts.
Demo questions that expose the design
These questions show whether a tool can support a backtest.
| Question | Why it matters | What to listen for |
|---|---|---|
| What does it need before showing a forecast? | Typed rates mean top-down. | Scheduled charges from billing. |
| What counts as a new subscriber? | Each definition changes the adds. | A written definition, changed once. |
| How does it treat intro pricing? | First orders are partial revenue. | A survival curve by channel and offer. |
| Can it show last quarter's miss by layer? | An unsplit miss cannot be fixed. | Base, adds and timing tied to sources. |
How Saras iQ Helps
Saras iQ assembles the forecast from committed rebills, modeled adds, and expected churn across sources, so the forecast is built on the actual base. For a team weighing revenue forecasting software, iQ holds the schedule, the retention curves and the definitions any bottom-up tool would need.
The forecast assembled from the actual base
In iQ, subscription records from Recharge, Skio, Loop, or Stay AI sit in one normalized dataset with the source platform on every record. Next scheduled charge, recurring revenue and cadence give the committed base, and the cohort view supplies retention for the churn assumptions. Fields a platform does not provide stay empty and are never inferred. Through the iQ MCP server, Claude can run forecasting workflows on that certified layer. A KPI master, lineage and default definitions sit behind it.
Where Saras iQ Fits
The business decides what counts as a new subscriber first, and a Saras consultant writes it into the context layer with you. A Saras analyst then tests it against your historical questions before go-live. Scheduled is not collected, since a failed payment can still take some of it away.
AG1's COO describes the starting point most brands share.
"We had a wealth of data but lacked infrastructure. Saras helped us transform our data strategy."
Ben Smith, COO & Advisor, AG1
The context layer behind iQ comes with three capabilities for a forecast you can check.
- KPI master for revenue. Forecast and actual need one definition, and the master checks for a second.
- Lineage to each platform. A miss gets picked apart layer by layer, and each figure traces to its source.
- Default definitions for what counts as new. First-ever subscribers, reactivations and converting buyers can each count, so defaults settle it.
Conclusion
A forecast you can explain is worth more than one that looks precise. Before you sign for a tool, check whether it starts from the schedule or from a typed growth rate, and backtest it on a closed quarter. Revenue forecasting software earns its place when a miss can be traced to a layer.
If you want your scheduled charges, retention curves and definitions assembled into one forecast across your tools, talk to our data consultants at Saras Analytics. They can set it up with Saras iQ.

.png)
.webp)







.jpg)



























































































