Customer-to-Referral Rate (CFR): Definition, Formula, and Limits
Most businesses can say how many leads came in last quarter. Far fewer can say how many of the customers they finished work for introduced someone new. Customer-to-Referral Rate gives that question a denominator.
In one line
Customer-to-Referral Rate (CFR) is the percentage of eligible customers in a completed-experience cohort who make at least one verified, attributable introduction of a prospective new customer within a stated period. A share, a click, a stated willingness to recommend, and a completed sale are not the same thing as an introduction. A CFR report names the referring customer, the referral event, the observation window, and the limits of what the business can verify.
The definition
CFR-90 = (unique eligible customers with at least one verified referral introduction within 90 days of completion ÷ unique eligible customers whose experience was completed in the cohort) × 100
The 90 is the observation window in days, written CFR-W in general: CFR-90, CFR-180, CFR-365. Ninety days is a working default, not a validated norm. A business with a long service cycle may need a longer window, and the window is always reported beside the result.
The clock starts when the experience is complete: the service was performed or the purchase was fulfilled, and the customer can judge the outcome. It does not start at a lead, a quote, a marketing contact, an email open, or a review request.
The denominator counts distinct customers, not invoices, jobs, orders, or email addresses. A customer with four completed jobs enters the cohort once, under one rule written in advance, such as the first completed job in the period.
The numerator counts distinct referring customers, not referrals. One customer who introduces five people is one referring customer. The five prospects are counted separately, as referred leads.
An illustration, not data. Three customers finish on the same day. The customer who introduces two prospects inside the window is counted once. A shared link that nobody acts on is not an introduction.
The letters CFR have other meanings in law, medicine, and software. Where that could confuse a reader, write the name out.
What makes an introduction verified
An introduction counts when there is lawful, checkable evidence of a real, identifiable prospective buyer, and the existing customer who sent them can be named with reasonable confidence. Four common forms:
- The prospect says so. A new inquiry names a specific existing customer as the person who referred them, and the answer is stored with its date.
- The customer makes the introduction. An existing customer submits an introduction, the referred person has agreed to be contacted, and the resulting lead is recorded once.
- A code or link, followed by an inquiry. An identified new prospect uses a customer's referral code or link to make a real contact request. The code being shared is not enough.
- A conversation, written down at the time. A staff member records a spoken introduction after talking with the prospect, in a structured field with a date, not reconstructed later.
Offline referrals count. A phone call, an email, or a conversation in person is a valid introduction when it is confirmed and recorded. A method that counted only tracked links would miss much of how people recommend a business.
What does not count
| Often counted as a referral | Why it is not a CFR introduction |
|---|---|
| A link click, QR scan, or view of a referral page | No identifiable prospect made contact |
| A social share or forwarded email | A recommendation was attempted. Nobody was introduced yet |
| A survey answer such as "I would recommend you" | It is stated intent. CFR counts what a customer did |
| A "referral" session in web analytics | It is a visit from a link on another website, not a customer's personal recommendation |
| A public review | It is public evidence, measured by Customer-to-Review Rate |
| A lead from an affiliate or a paid partner | It is a commercial arrangement, reported separately from customer advocacy |
| A referred prospect who buys | It is an outcome of an introduction, tracked as a referred customer |
| A staff guess, a test record, or a customer referring themselves | There is no evidence of a real introduction |
These can each have their own counter. None of them enters the CFR numerator.
A worked example
A business completes work for 200 unique eligible customers in a cohort, and every one of them has had the full 90 days. Twelve of them made at least one verified introduction inside their own 90 days. CFR-90 is 12 ÷ 200 = 6%.
Those 12 customers introduced 19 different prospects. Nineteen is the count of referred leads. It is not 19 referring customers. Five of the 19 went on to buy. Five is the count of referred customers, and it is not part of CFR.
The numbers are illustrative. They are not a client result and not a benchmark, and this page does not publish one.
Timing and data states
An introduction counts for a cohort only if it happens on or after the customer's completion date and inside the window. A referral made before the work was finished, or long after the window closed, is real. It belongs in another view, not in this number.
Every reported CFR carries one of three states:
- Verified. Every customer in the cohort has had the full window, and the referral channels the business agreed to track were observed for all of it.
- Provisional. Some customers are still inside their window. Forty days after the first completion, CFR-90 is provisional, and it is never compared with a finished cohort as if the two were equal.
- Unavailable. The business cannot reliably connect a referral to the customer who made it and to a real prospect. Unavailable is not zero. A 0% is reported only when a finished cohort was fully observed and nobody made a verified introduction.
What CFR does and does not tell you
- It measures the referrals a business can see. Some word of mouth is never recorded. No method can state the true total without the data.
- It is not a trust score. A higher rate is consistent with trust and relevance, and it is also shaped by incentives, purchase size, service cycles, the category, and how well referrals are captured.
- It does not show cause. A rise in CFR does not prove that a particular email or program produced it, and it carries no promise about revenue or rankings.
- It does not compare cleanly across industries. Referral timing differs by category. Ranking businesses against each other needs matched windows and enough customers in each cohort.
- Eligibility rules are written before the count. A customer is never removed from the cohort for being unhappy, inactive, or unlikely to refer.
- One lead claimed by two customers is flagged, not counted twice. The rule for competing claims is set before reporting, and the uncertainty is shown.
Consent and incentives
A customer cannot give consent on a friend's behalf. The sound pattern is one where the referred person makes contact themselves, or has agreed to be contacted before any outreach.
A reward for a referral is a material connection, and the Federal Trade Commission's endorsement guidance explains when one must be disclosed. Paid affiliates are reported separately from ordinary customer advocacy.
Referral rewards are kept apart from review requests. A reward tied to a review is prohibited by the major review platforms.
Law, healthcare, finance, and other regulated fields have their own rules on referral fees, privacy, and client confidentiality. Those rules decide what can be measured and offered there.
What happens after the introduction
CFR stops at the introduction. Whether the introduction turns into business is measured separately, each measure with its own cohort and window.
| Measure | What it counts | What it answers |
|---|---|---|
| Customer-to-Referral Rate (CFR-W) | Unique referring customers ÷ unique eligible customers in a completion cohort, within W days | How many customers introduce someone |
| Verified referred leads | Distinct prospects introduced in a period, with duplicates removed | How much work entered the pipeline |
| Referral lead conversion rate | Referred leads who became paying customers within a stated window ÷ all verified referred leads in that intake cohort | Whether introductions become customers |
| Referral-sourced revenue | Collected revenue from converted referred customers, with the attribution method stated | What the introductions were worth |
The conversion rate uses referred leads as its denominator, not completed customers, so it matures on its own schedule. A sale two years after the introduction is not moved back into a short-window figure. Revenue is labeled as what it is: collected sales, recognized revenue, or gross profit.
A lead's quality never changes the CFR after the fact. An introduction that sales did not want is still an introduction.
Where it sits in the Marketing Helix
In the Marketing Helix, the purchase is not the end of the customer's motion. The Post-Purchase Helix describes how one customer's experience reaches the next customer. It can become public evidence, which Customer-to-Review Rate and Time-to-Review measure. It can also become a personal introduction, which CFR measures.
Neither is automatic, and one does not lead to the other. Many customers do neither. A customer can refer without ever reviewing, and review without ever referring. The referring customer keeps moving on their own path, and the person they introduce starts a separate one.
Prior usage and attribution
Referral rates are an established measure with more than one meaning. Shopify's July 2026 guide gives two formulas under the same name: the number of customers who made a referral divided by total customers, and the number of referred customers divided by total new customers. Some software glossaries call the first one Customer Referral Rate and abbreviate it CRR, the letters this site uses for Customer-to-Review Rate.
The Marketing Helix is publishing one specific reporting definition: a completed-customer cohort, distinct referring customers, verified introductions, and a fixed window. It does not claim to have originated referral rates, the measurement of customer advocacy, or the initials.
Sources
- Shopify (July 13, 2026). Referral Rate: What It Is and How to Improve It.
- Shopify (July 27, 2026). Customer Referral Strategies: How to Get Customer Referrals.
- Alexander Jarvis. What Is Customer Referral Rate in SaaS? How to Improve It.
- Google Analytics Help. Default channel group, which defines the Referral channel as visits arriving through links on other sites and apps.
- Federal Trade Commission. FTC's Endorsement Guides: What People Are Asking.
- Federal Trade Commission. Consumer Reviews and Testimonials Rule: Questions and Answers.
How many completed customers leave a public review?
Customer-to-Review Rate measures the share of a completed-customer cohort that publishes a verified review inside a fixed window. Read the Customer-to-Review Rate definition →