Revenue estimate methodology, how Monetisation calculates every range

Why we show the method

Any tool can print a large number labelled "untapped revenue". Without the arithmetic behind it, that number is impossible to check and easy to dismiss. We think an estimate is only useful if you can see how it was built and disagree with it.

So every opportunity in Monetisation carries the same five things: a range, the inputs, the assumption, a confidence level and a plain-language method line. Every figure is labelled "Estimate". None is a guarantee of results.

The three values in every range

The dashboard counter "Untapped monetisation" adds up the expected values of open opportunities. It is labelled as an estimate too, and every item inside it can be opened.

Ranges are not statistical confidence intervals around a forecast. They are scenarios built from the assumption's plausible spread. That is a deliberate choice: most pricing questions do not have enough history for formal intervals before a test, and pretending otherwise would overstate precision.

Inputs

Each estimate uses only the data needed for it. Depending on the opportunity, inputs can include:

• Billing data from Stripe, through a restricted read-only key: subscriptions, prices, intervals, coupons, churn, expansion
• Traffic data from Google Analytics 4 and Search Console: visits, landing pages, queries, conversion
• Your public website: pricing page, plan names, feature lists, paywall, ad slots, affiliate links
• Product usage from a read-only database connection or CSV: usage per account
• CRM data from HubSpot or Salesforce: deal sizes and discounts
• Ad platform data from Google Ads and Meta Ads: spend and acquisition cost
• CSV exports of anything else, such as affiliate or ad network reports

The method line names which inputs were used and for what period, for example "your last 6 months of Team subscriptions".

The assumption library

Every estimate rests on at least one assumption about behaviour: how many customers leave after a price rise, how many choose annual when it is the default, what share of visitors click an affiliate link. Monetisation keeps these in an assumption library, one entry per opportunity type, more than 40 in total.

Each entry has:

• A central value and a plausible spread used for low and high
• The source of the default: general pricing practice, simple economics, or your own data
• The rule for replacing it: when your data is enough, your observed value replaces the default

You can see the assumption on every card. If you think it is wrong for your business, the range shows what happens across its spread, and your experiment results feed back into later estimates.

Confidence levels

AI crawler licensing, for example, usually stays at low confidence because amounts in that market are rarely public. A coupon that is still active after its campaign is usually high confidence, because the leakage is a fact in your billing.

Public pages only versus Stripe connected

The sample audit on the homepage reads public pages only. It knows your pricing page and business type, not your customers. Its estimates use typical values for your business type and are marked low confidence. Connecting Stripe changes the basis.

A worked example

Take a hypothetical SaaS with Stripe connected. The opportunity is a price rise on the Team plan for existing and new customers.

• Inputs: 214 Team accounts at 59 USD per month, monthly churn on Team 3 percent over the last 6 months
• Proposal: 12 percent increase, from 59 USD to about 66 USD
• Assumption: extra churn from the increase, central value 4 percent of affected accounts, spread 1 to 9 percent

The method line reads: "12 percent price increase on 214 Team accounts, 4 percent assumed extra churn, based on your last 6 months." Confidence is medium: the inputs are your own, the churn response is a general assumption until tested.

From estimate to measured result

An estimate decides what to try first. An experiment tells you what happened. "Run experiment" creates a card with a hypothesis, a metric, a baseline period and a test window, delivered in one of three ways:

• Snippet variants with a holdout (Growth and Scale): visitors are assigned to a variant or a holdout group that keeps the original page, and conversion and revenue per visitor are compared.
• Dated intervention: you make the change in your own billing. Monetisation fits the baseline trend from before the change and compares the period after with what the trend predicts.
• Playbook: a checklist and a measurement window for changes that are not a price.

Results show lift, an interval and one of four verdicts.

Losses are shown like wins. A dated intervention is weaker evidence than a holdout, because other things change at the same time, and the result says so.

What an estimate is not

An estimate is not a promise, a forecast of your total revenue, or financial advice. It is a structured, transparent guess designed to help you choose which change to test. Monetisation is read-only and never changes your prices or subscriptions, so every action stays your decision. The security page explains how data is accessed.

For where estimates are used day to day, see pricing optimization software, revenue optimization software and the arithmetic of annual plan pricing.

See an estimate on your own site

Run a sample audit from the homepage. In about a minute you get the first three opportunities with their ranges and method lines. No card is needed. Get started on a plan to connect Stripe and see estimates built from your own data.