# Referral persona landing page test: a design | SourceX

Canonical: https://refer.sourcex.si/guides/the-referral-landing-page-test-for-different-decision-makers
Last updated: 2026-11-04

A partner growth manager can test role-specific versions of their own referral landing page, but only one role decides: the CEO. Versions written for finance leads, operations leads, counsel or board members must route those readers to the CEO and never invite them to decide or apply. Run the test in two stages: paraphrase interviews with each role before launch, then a live comparison judged on qualified applications from companies whose CEO asked for the introduction, not on visits or clicks. No results for this design are published.

This design covers a page on the partner's own website. SourceX's pages are not part of the test, and the partner never collects company records through the page or anywhere else.

## One decider, many readers of the same page

A landing page about data licensing attracts the people who handle the pieces: the finance lead curious about money, the operations lead who knows the systems, in-house counsel who worries about contracts, a board member who wonders whether to raise it. Each of them can help. None of them can say yes for the company: SourceX's screening page says opening a licensing discussion and signing for the company fall to the CEO.

That creates the hypothesis worth testing. A page written only for the CEO may lose the colleagues who would have carried it to the CEO. Role-specific versions may bring more CEOs to the conversation, or they may invite helpers to act on their own and produce introductions the CEO never agreed to. Only a test separates the two.

**Hypothesis, written before launch.** A role-routed set of pages will produce at least as many qualified applications as a single CEO page, with no rise in applications started without the CEO's yes.

## Rules every version must follow before it goes live

These apply to the control page and every variant alike. They come from SourceX's published terms and the sample Channel Partner Agreement, not from this test.

| Rule | Source | What it means on the page |
|---|---|---|
| No public claim to be SourceX's partner without written authorization | Sample agreement, section 4.5 | Ask SourceX in writing before naming it as your partner; never imply endorsement or preferential treatment |
| No suggestion that an estimate, interest, a signature or an upload assures a purchase or a reward | Sample agreement, section 4.6 | No payout figures, no "you will be paid" |
| The partner cannot commit SourceX to anything | Sample agreement, section 7.1 | No prices, timelines or acceptance promised on SourceX's behalf |
| The reward is paid separately and never reduces the company's payout | Terms, section 4 | Say who could be paid, and that the company's payout is the same either way |
| A click alone decides nothing about credit | Terms, section 2 | Never tell readers that following your link secures anything |
| Fit is SourceX's decision | Screening page | "May be a fit; SourceX decides after its review", never "your company qualifies" |

The terms sentence behind the fourth row says the reward "is paid separately by SourceX and is never deducted from the supplier’s agreed compensation." Where the page states location criteria, it quotes SourceX's sentence in full: Companies in the U.S., Canada and the EU qualify, as do companies in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia when at least 70% of the data is in English. How to tell clients about a reward is discussed in [do I have to disclose a referral fee](/questions/do-i-have-to-disclose-a-referral-fee); whether any disclosure law applies where you publish is a question for your own adviser.

## Control page and role-routed set

**Control (A):** one page written for the CEO, linked from all your outreach. **Treatment (B):** the same CEO page plus four short routing pages, each linked from outreach aimed at that role.

| Version | Reader | What the reader came to learn | What the version tells them | The only action offered |
|---|---|---|---|---|
| CEO | The company's chief executive | Whether this could concern the company | The published criteria, who could be paid, that SourceX decides fit | Ask the partner for an introduction, apply directly, or decline |
| Finance | Finance lead | How money moves | The reward comes from SourceX separately; the company's payout is unchanged; no amounts promised | Forward a two-line note to the CEO |
| Operations | Head of operations or IT | Which records and how much work | The kinds of records SourceX describes; that the partner wants none of them; that the company decides scope | Forward the note to the CEO |
| Counsel | In-house legal | Rights and contracts | SourceX checks rights in its own review; the company's counsel advises the company; the page gives no legal advice | Forward the note to the CEO |
| Board | A director or investor | Whether to push the company | The CEO decides; what a seat reveals stays with that company and may prompt one question to its CEO | Raise it with the CEO once, or not at all |

No non-CEO version offers an application button, an approval path, a form asking about systems, headcount or files, or a way for the reader to request the introduction on the company's behalf. The two-line note is the same everywhere: what the page is about, and that the CEO decides whether to look.

## Stage one: the role-specific comprehension study

Before any version meets real traffic, test whether each role reads its version correctly. Digital.gov's plain-language guidance describes paraphrase testing: a participant reads a section and explains it in their own words, the facilitator does not correct them, and six to nine interviews are suggested (our paraphrase of a search excerpt).

- **Who takes part.** Six to nine people per role, recruited with consent from colleagues and contacts who hold that role somewhere and are not being prospected. Nobody is asked about their own company.
- **What they read.** The version for their role, and the CEO version for the CEO readers.
- **Five questions after reading.** Who decides whether the company talks to SourceX? Who could be paid, and by whom? Does this page say your company qualifies? What would you do next? Does SourceX stand behind this partner?
- **Scoring.** Each answer is marked correct, partly correct or misread, in the participant's own words.

**Launch gate, set before the first interview.** A version goes live only when every reader in its role names the CEO as the person who decides, and none reads it as saying a company qualifies or as an endorsement by SourceX. A version that fails is rewritten and tested again with new readers. These counts are design checks, not results to publish.

## Stage two: the live comparison

Digital.gov places controlled comparisons near the end of testing because they show whether content works but not why (our paraphrase), which is why stage one comes first.

| Element | Specification |
|---|---|
| Allocation | Alternate fortnights between A and B, starting with a version picked by a coin flip logged before launch |
| Held constant | The same outreach channels and sending rhythm in both periods |
| Your experiment record | Your contact form asks only for a name, role, email and an optional question, and notes which page version it was sent from; this is your own record for the test, never something SourceX uses for credit |
| Primary measure | Qualified applications per arm: applications from companies whose CEO asked you for the introduction during that arm's fortnights, verified by SourceX, credited to your introduction and not shown as closed on the reading date |
| Reading date | 120 days after the last fortnight ends, which lets every 90-day attribution window close before anything is counted |
| Never counted | Page visits, clicks, time on page or form starts |

Your dashboard shows each referral's stage and reward status, as SourceX's [how it works](/how-it-works) page describes; copy those words exactly when you count.

| Guard measure | Why it is watched |
|---|---|
| Applications started by someone other than the CEO, without the CEO's yes | The main harm a role page could cause |
| Readers who ask "so we qualify?" | A sign the page implies acceptance |
| Records, files or system details sent to you | A protocol failure; the target is zero |
| Complaints or requests not to be contacted | Trust cost of the extra pages |
| Notes forwarded to the CEO, where the reader tells you | Whether routing happens at all |

## Duration, floor and stop rules for the page test

Run the live stage for at least two full quarters, and compare rates only when each arm has at least 15 introduction requests from CEOs. Below that, publish counts per arm and no percentages.

**Stop rules.**

- Someone other than the CEO starts an application from a role page: pause that page and rewrite its action.
- A reader sends you records or system details: pause the page until the wording is fixed, and keep nothing that arrived.
- SourceX has not authorized, or withdraws authorization for, any public description of you as its partner: take the wording down.
- SourceX changes its terms or screening page: reissue every version with a date and restart both the comprehension check and the live window.
- One arm looks ahead early: keep going. Stopping on a good streak is how false winners are made.

## What the results could and could not tell you

| Count on the reading date | A: CEO page | B: role-routed set |
|---|---|---|
| CEO introduction requests | To be counted | To be counted |
| Qualified applications (primary) | To be counted | To be counted |
| Applications without the CEO's yes | To be counted | To be counted |
| "Do we qualify?" questions | To be counted | To be counted |
| Records sent to the partner | To be counted | To be counted |
| Complaints | To be counted | To be counted |

**Decision rule, written before launch.** Decide in advance what size of gap would make you change your page, and report the gap with its uncertainty rather than as one number.

| Result | Decision |
|---|---|
| B ahead beyond your threshold, with no applications lacking the CEO's yes | Keep the role-routed set |
| B ahead, but any application started without the CEO's yes | Do not adopt B until its pages are fixed and retested |
| A ahead, or an interval that includes zero | Keep the single CEO page; the extra pages did not earn their place |
| Records sent to the partner in either arm | Fix the wording before reading anything else |

A result says something about your audience and your pages, not about partners in general. A site-level version of this kind of test, run on SourceX's own headline, is described in [the maximum-payout headline study](/guides/what-happens-when-a-referral-page-removes-the-maximum-payout-headline); checking the application path itself is covered in [testing the referral form before sending more traffic](/guides/test-the-referral-form-before-sending-more-traffic), and testing message length in [the long and short introduction experiment](/guides/how-to-test-long-and-short-company-introductions).

## Where a role-routed page can mislead

- **Helpers become deciders.** However carefully the page routes, a keen finance lead may act alone. The guard measure exists because wording cannot fully prevent it.
- **Comprehension is not behavior.** Readers who explain a page correctly in an interview may still skim it at work.
- **Role labels are guesses.** Many readers hold two roles, and in an owner-run company the owner is usually the CEO. Keep the CEO version one click away on every page.
- **More pages, more to keep true.** Five versions must all change when SourceX's terms change. A stale page is worse than a missing one.
- **Endorsement creep.** Logos, badges or "official" language imply a relationship the agreement does not grant without written authorization. Leave them off.

## Next step

> **Refer a company. No cap on rewards.** Default reward on a company's first qualifying transaction: the greater of $50,000 plus 5% and 15% of Actual Supplier Payout, with no cap; further qualifying transactions first agreed within 12 months of the first Closing pay 15%. Eligibility and payment conditions apply, including the USD 100,000 Minimum Transaction, and a signed Schedule A or addendum can differ. A lead, meeting or signature alone does not trigger payment.

Before any version goes live, [register as a partner](/register) and sign the Channel Partner Agreement, take every criterion from the current [who qualifies](/who-qualifies) page and keep a dated copy of the [program terms](/terms) with each page version you test. More experiment designs are gathered in the [guides hub](/guides).

## Sources

- https://refer.sourcex.si/who-qualifies
- https://refer.sourcex.si/terms
- https://sourcex.si/documents/partner/SourceX_Channel_Partner_Agreement_Sample.pdf
- https://digital.gov/guides/plain-language/test/paraphrase-testing
- https://digital.gov/guides/plain-language/test/comparative-studies
