# Referral form usability test, no test entries | SourceX

Canonical: https://refer.sourcex.si/guides/test-the-referral-form-before-sending-more-traffic
Last updated: 2026-11-04

You cannot test SourceX's form by filling it in. Partners never submit test or dummy introductions, invented companies or any real company's details to SourceX. What you can test is the path you control: the email you send, the partner link inside it and the instructions that tell a CEO who fills in what. Walk that path with colleagues in five tasks, send every question your materials cannot answer to SourceX, and judge the change on verified applications from companies whose CEO agreed, not on clicks.

Sending more companies down a path that confuses them wastes the CEO's goodwill faster than any other mistake a partner can make. The form itself belongs to SourceX and changes on SourceX's schedule. Your instructions are the part that breaks most often, and the part you can fix today.

## Your part of the path and SourceX's part

Five segments sit between a CEO's yes and an application SourceX can verify. You write one of them, share two, and can only ask about the last two.

| Segment | Who controls it | How you test it | Never |
|---|---|---|---|
| Your introduction email | You | A colleague reads it aloud and says what they would do next | Promise fit, timing or an amount |
| Your partner link | SourceX issues it; you choose where it goes | Copy it exactly from the dashboard; ask SourceX where it lands | Click through to generate visits under your own tag |
| Who fills in what | The company decides; the CEO decides whether to proceed | Your instructions say it in one line, and a colleague repeats it back | Offer to complete any part of the application |
| The application screens | SourceX | Read the public first screen; ask SourceX about the later screens | Type anything into the live form |
| What the CEO sees after submitting | SourceX | Ask SourceX what the applicant receives and who follows up | Guess, then describe your guess as fact |

Two program facts frame the whole exercise. The company applies itself, and SourceX's terms decide credit from dated records, so a link click on its own decides nothing ([what creating a link does and does not do](/questions/does-creating-a-referral-link-make-an-introduction-eligible-for-a-reward)). Partners can also submit an introduction from the dashboard, according to [how it works](/how-it-works), but only after signing the Channel Partner Agreement. Neither route needs you to touch the company's records.

## What a company meets on the application's first screen

SourceX's public application page, read on 2026-10-27, tells an applicant the following on its opening screen. Read it yourself before you write a single instruction, and reread it before each new round, because SourceX can change it.

- **Length.** Three screens of about seven minutes check fit; account, systems and the agreement come after. The step bar reads step 1 of 7: Company, Account, Company details, Data overview, Systems, Privacy, Review & submit.
- **What is not needed.** No files, no passwords or API tokens, and nothing is licensed or transferred by applying. The page also says applying does not promise an offer or payment (our paraphrase).
- **Who is asked.** A role field gives "Owner, CEO" as examples, and legal details for the agreement are asked once, with nothing signed by applying.
- **Location.** The country field notes that the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia need at least 70% of records in English.
- **Two features to mention in your instructions.** Answers can be saved for later, and an optional tool drafts answers from a pasted website. That tool is for the company. A partner who pastes a client's website into it has sent company details to SourceX.

The first screen is all you can see without entering anything. Everything after it is a question for SourceX, which is exactly how the test treats it.

## The five-task path walkthrough

Recruit colleagues, not clients. Give each one your draft email, the instructions you plan to send and the public application page on screen. Ask them to think aloud while they work through five tasks, and stay quiet. Nielsen Norman Group describes a basic think-aloud study as three things: representative users, realistic tasks and a facilitator who keeps quiet, and it warns that untrained prompts can change what people do (paraphrase).

| Task | What you ask the colleague | A pass sounds like | Failure to log | Fix you own |
|---|---|---|---|---|
| 1. Start | "Using only my email, show me where you would begin." | Finds the link and knows the company applies itself | Looks for an attachment or expects a call first | Put the link and one "you apply yourself" line first |
| 2. Who | "Who at your company fills this in, and who decides?" | The CEO decides whether to go ahead; others may help gather facts; the partner fills in nothing | Names the partner, or a manager, as the decider | One sentence naming the CEO as the person who decides |
| 3. Prepare | "What would you get ready before starting?" | A list of the systems that hold records; no files or logins | Plans to gather files, samples or passwords | Quote the page's "no files" line in your instructions |
| 4. Commit | "What does applying commit you to, and does anyone get paid?" | Applying licenses nothing; the partner may receive a reward from SourceX that does not come out of the company's payment | Thinks applying is a sale, or is surprised the partner could be paid | State your possible reward plainly, beside the link |
| 5. After | "What happens after you submit, and how would you say no?" | SourceX reviews; the partner promises no timing or result; ignoring or declining ends it | Expects the partner to chase, or an answer within days | Describe the next step only as far as SourceX has told you |

Task 4 rests on a published rule, so quote it rather than paraphrase it: "The referral reward is paid separately by SourceX and is never deducted from the supplier's agreed compensation" (terms, section 4). The company receives the same amount whether or not a partner introduced it. How to word the disclosure itself is covered in [whether you have to disclose a referral fee](/questions/do-i-have-to-disclose-a-referral-fee).

**Rules for every session.** Readers read and talk; nobody types into the live application. Use no real company name, not even a pretend one that resembles a client. Log each task as pass, partial or fail, with the reader's own words for anything that went wrong. Keep the log in your files only; it contains no company information, because none was used.

**How many readers.** Nielsen's argument for usability work is to test with no more than five users and run several small rounds, fixing problems between them, with more users when the audience splits into distinct groups (paraphrase). A CEO and the office manager who helps them read instructions differently, so a round can be three readers playing the CEO and two playing the helper. Fix what failed, then run a second round on the new draft.

## Questions only SourceX can answer

Send these by email, so the answer arrives dated and in writing, and add each answer to your instruction file with that date.

1. Which page does my partner link open, and does it carry the company straight to the application?
2. What does an applicant see on the screens after the first, in outline?
3. What email or message does the CEO receive after submitting, and from whom?
4. If a helper starts the application, how does the CEO take it over or approve it?
5. Has anything on the application changed since the date on my instructions?

If SourceX's answer contradicts your draft, SourceX's answer wins and the draft changes. Never fill a gap by describing what you expect the screens to say.

## What to measure: understanding, qualified applications and side effects

The walkthrough measures understanding. The real test comes afterwards, when CEOs follow the revised path, and it is judged on outcomes, not on clicks or visits.

| Measure | Definition | Read when |
|---|---|---|
| Task pass count (understanding) | Readers passing each task, per round, out of readers tested | After each round |
| Primary: qualified applications | Of the CEOs who agreed in writing to be introduced, how many companies submitted an application that SourceX verified and credited to your introduction, as your dashboard shows it | Day 45 and day 90 after each introduction |
| Clarifying questions | Questions each introduced CEO sent you about the path | Continuously |
| Partner-filled applications | Applications in which anyone outside the company entered company details | Continuously; the only acceptable count is zero |
| Unwanted effects | A CEO who felt pushed, misread a promise, or applied though a published fit fact clearly failed | Continuously, in the CEO's words where offered |

Compare the next set of introductions sent with the revised instructions against the last set sent with the old ones: at least ten CEOs in each set, or six months, whichever comes later. Counts that small show a direction, not a proof. Nielsen makes the same point about think-aloud testing itself: it "doesn't lend itself to detailed statistics, unless you run a huge, expensive study". Your dashboard stages are SourceX's words; copy them exactly ([where a partner sees the current stage](/questions/where-can-a-partner-see-the-current-stage-of-an-introduction)).

**Stop rules, fixed before the first round.**

- Stop the session at once if anyone starts typing into the live application.
- Stop sending the revised instructions if any CEO reports that they promised fit, an amount or a timeline, and rewrite before the next introduction.
- Pause and rerun the walkthrough when SourceX tells you the application or the terms changed.
- Stop the comparison if you notice you chose which CEOs got which version by how promising they seemed. Assign by date order instead.

## Reading the result and choosing the change

No partner's results are published here and none exist yet. A result is read against rules written in advance.

| What you see | What it means | What to change |
|---|---|---|
| A task fails for two or more readers in a round | The instructions, not the reader, are unclear | Rewrite that line and rerun the task |
| Task 2 fails even after rewriting | Readers still think someone other than the CEO decides | Move the CEO line to the first sentence of the email |
| Qualified applications rise, clarifying questions fall, no unwanted effects | The revision helps on the outcome that matters | Keep it, dated, and retest when SourceX changes the form |
| Qualified applications flat, understanding better | Clearer, not yet proven better | Keep it if no side effects appeared; keep counting |
| Applications rise but more clearly fail a fit fact | The instructions pulled in companies that should have ruled themselves out | Restore the fit line and the [fit checker](/tools/company-fit-checker) link |

Report the whole record: the dated instruction versions, every reader's task results and every introduction in both sets, including CEOs who never applied. A site-wide version of this logic, applied to SourceX's own headline, is designed in [what happens when a referral page removes the maximum-payout headline](/guides/what-happens-when-a-referral-page-removes-the-maximum-payout-headline); this page stays with materials a partner writes.

## Which companies belong on the path at all

A smooth path is wasted on a company that fails the published fit facts. Before any link goes out, the CEO should know the screen: at least 50 people on the full-time payroll when the company was at its largest, with contractors left out; years of records the company created itself; the right to license them; and a CEO who wants the conversation. 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.

The partner's discovery stays at metadata level: which kinds of records, roughly how many years, which systems, who runs them. Whether the records match a current program, what the rights review finds and what a license would cover stay unknown until SourceX reviews the company. A reply to an interested owner that sets these expectations is drafted in [a partner's first-response checklist](/guides/a-partner-s-first-response-checklist-for-an-interested-owner).

## 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.

Once two rounds of readers pass all five tasks, [register as a partner](/register) and sign the Channel Partner Agreement, check each company against [who qualifies](/who-qualifies), and keep the [program terms](/terms) version your instructions quote beside the draft. More partner methods sit in the [guides hub](/guides).

## Sources

- https://refer.sourcex.si/terms
- https://refer.sourcex.si/how-it-works
- https://refer.sourcex.si/who-qualifies
- https://sourcex.si/apply
- https://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/
- https://www.nngroup.com/articles/thinking-aloud-the-1-usability-tool/
