# Partner FAQ effectiveness test: before and after | SourceX

Canonical: https://refer.sourcex.si/guides/the-partner-faq-experiment-that-removes-repeated-questions
Last updated: 2026-11-04

An FAQ can stop the same questions arriving from every CEO you contact, but only an honest count shows whether it did. Log and code every question for at least three equal periods, publish a dated FAQ whose answers restate SourceX's terms and fit page in your own words, then log three more periods with the same codes. Keep the FAQ if covered questions fall, a control code stays level and verified applications per contacted company do not drop. No partner or SourceX results for this study are published yet.

This is a design for a partner's own outreach materials. The FAQ is yours; the rules in it are SourceX's. Every answer repeats what the [program terms](/terms), [who qualifies](/who-qualifies) and [how it works](/how-it-works) already say, and an answer that cannot point to one of those pages is cut.

## The hypothesis, and the measure that outranks the question count

**Hypothesis.** Publishing an FAQ with every outreach message lowers the number of covered questions per contacted company, without lowering the share of contacted companies that reach a verified application.

The two halves are not equal. An FAQ that cuts questions by putting CEOs off has failed, however quiet the inbox becomes. So the measure that decides is qualified, not convenient:

| Role | Measure | What never counts |
|---|---|---|
| Primary | Verified applications credited to your introduction, per contacted company, read at the same age for every company | Page views of the FAQ, link clicks, opens |
| Mechanism | Covered questions per contacted company per period | A question the FAQ does not answer |
| Control | Scheduling and logistics questions per contacted company per period | Anything the FAQ addresses |
| Unwanted effects | Misreadings, declines that cite the FAQ, questions about its wording, records sent to you | Silence, which is not agreement |

Verified applications come from the stage shown on your partner dashboard, which SourceX says displays each referral's stage and its reward status. Credit goes to the first valid introduction within the 90-day window when the company submits a verified application (terms, section 2), so read each company's outcome no earlier than 90 days after it was contacted.

## The question-frequency before-after study

| Element | Specification |
|---|---|
| Unit | A company whose CEO received your standard outreach message |
| Periods | Equal lengths, for example two weeks; at least three before the FAQ goes out and three after |
| The change | FAQ version 1, dated, linked from every outreach message from a fixed start date; nothing else in the message changes that day |
| Question log | Every question from a contacted CEO or their staff, by email, call note or chat, logged the day it arrives |
| Coding | One code per question from a codebook frozen before period one; the same company asking the same code twice in a period counts once |
| Covered codes | The eight questions the FAQ answers (below) |
| Control code | Scheduling and logistics: when to talk, who else joins, which time zone |
| New codes | Any question created by the FAQ's own wording, logged separately |
| Floor | At least five contacted companies in every period and twenty across each side; if a period falls short, every period is lengthened by the same rule, decided in advance |
| Second coder | A colleague blind-codes one question in five; disagreements are reported, not quietly fixed |

The three-and-three rule comes from published study-design guidance. Cochrane's Effective Practice and Organisation of Care group (EPOC) suggests leaving out interrupted time series studies that lack a clearly defined point when the intervention happened and at least three data points before and three after it (paraphrase). A partner's FAQ test is a small interrupted series: one dated change, counts on either side.

The control code is the check on everything else. If covered questions and scheduling questions both fall after publication, fewer conversations or a quiet season probably did it, not the FAQ.

## The FAQ: eight answers in the partner's words, each with its source

The wording below is Illustrative, written for this page; yours should sound like you. The source column is not optional. Before publishing, read each answer against its page on the day you date the FAQ.

| Covered code | The question as CEOs ask it | Answer, in the partner's words | Where SourceX says it |
|---|---|---|---|
| C1 | Who decides whether we talk to SourceX? | You do, as CEO. Neither I nor anyone else can make that decision for your company. | Who qualifies: "Under the program, opening a licensing discussion and signing for the company fall to the CEO." |
| C2 | Does your reward come out of what we would be paid? | No. My reward, if there is one, comes from SourceX and leaves your payment untouched. | Terms, section 4: "The referral reward is paid separately by SourceX and is never deducted from the supplier's agreed compensation." |
| C3 | Do you need to see our records? | No. I will never ask for files, samples or logins. What your company shares is decided between you and SourceX. | How it works: "You do not need to export, upload, or describe confidential records." |
| C4 | Could our company be a fit? | Possibly; SourceX decides in its review, under the criteria of the specific program. See the location sentence below the table. | Who qualifies: screening profile and "SourceX applies the criteria of the specific program, not one universal minimum." |
| C5 | Are you paid, and when? | Yes, I may be. Nothing is earned when you reply, meet or sign; a reward follows only a completed, paid licensing deal and SourceX's own conditions. | Terms, section 5: "A lead, meeting or signed agreement alone does not earn a payable reward." |
| C6 | What would SourceX pay us? | I cannot say, and I will not guess. Your company and SourceX agree commercial terms directly. | How it works: the company agrees commercial terms with SourceX |
| C7 | We have been in touch with SourceX before. Does that matter? | Please tell me. It does not change who decides for your company; it can mean no partner is credited for the introduction. | Terms, section 3: a relationship "includes any earlier lead, account or email correspondence between SourceX and the company, however it began." |
| C8 | What happens if we say yes? | I submit one introduction, your company applies itself, and SourceX reviews it. Whether to go ahead stays your decision. | Who qualifies: "The company applies itself at /apply"; how it works: the company controls whether to proceed |

**C4 in full.** The answer adds the screening facts from who-qualifies in plain words: roughly 50 or more full-time employees at the company's peak and several years in operation (some programs set higher thresholds, and smaller specialists can still be considered); documentation the company created itself; and the right to license it without breaching client confidentiality, employee privacy or contracts with others. Personal consumer data, scraped data, material that belongs to clients and one-off files are usually not a fit. On location it quotes the page word for word: 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. For how the English share is counted it links [the English-language guide](/guides/english-language-data-referral-eligibility) rather than explaining it.

**What the FAQ never contains.** A new criterion, a reward figure, a promise of fit, a description of how SourceX prices offers, or anything about a specific company. Its footer carries the date, the terms version (2026-10-22.2 on the day this page was checked) and the updated date shown on who-qualifies, plus links to all three pages and to SourceX's own [FAQ](/faq). Before any of this goes out, you have signed the Channel Partner Agreement, which section 1 of the terms requires before you submit an introduction.

## Measuring understanding and the effects nobody wants

A falling question count is not the same as understanding. Three further lines show whether the FAQ informed people or only quieted them:

- **Misreadings.** A CEO who writes "so we qualify" or "SourceX has approved us" read the FAQ as a verdict. One is a stop rule; see below.
- **Shift in kind.** Good FAQs move questions from the basics to specifics: from "who decides?" to "how do we count our English share?". Log new codes and read them alongside the totals.
- **Declines that cite the FAQ.** A CEO who says the FAQ made it sound complicated or risky is an unwanted effect. Record that it happened, not the private reasons behind it.

Questions that need someone else's answer, such as a company asking whether a particular contract restricts its records, go to a backlog rather than into the FAQ; [the founder-question backlog](/guides/a-backlog-for-founder-questions-your-platform-team-cannot-answer) and [the question owner board](/guides/how-to-track-questions-waiting-on-the-company) show two ways to run one.

## Minimum duration and stop rules

Three periods on each side is the floor, not the target. Do not publish the FAQ partway through a period, and do not change your outreach message, your list rule or your sending volume at the same time; any of those would arrive with the FAQ and take its credit.

**Stop, withdraw or restart when:**

- SourceX publishes a new terms version or changes who-qualifies. Withdraw the FAQ the same day, re-date it against the new pages, and start a new after-series. An answer that no longer matches a live page is worse than no answer.
- A CEO reads the FAQ as a promise of fit or payment. Withdraw, rewrite the answer that caused it, re-date it and restart the after-series.
- Anyone sends you files or records because of the FAQ. Do not open or keep them, tell the company you do not need them, and fix the wording before the FAQ goes out again.
- Contact volume in an after-period changes by more than half compared with the before-periods. Keep counting, but report that period separately.
- Never stop because the early after-periods look good.

## Reading the before and after columns

| Count | Before, periods 1 to 3 | After, periods 1 to 3 |
|---|---|---|
| Companies contacted | Not yet collected | Not yet collected |
| Covered questions per contacted company | Not yet collected | Not yet collected |
| Control questions per contacted company | Not yet collected | Not yet collected |
| New questions about FAQ wording | Not applicable | Not yet collected |
| Misreadings | Not yet collected | Not yet collected |
| Verified applications per contacted company | Not yet collected | Not yet collected |
| Records sent to the partner | Not yet collected | Not yet collected |

| What the counts show | Reading |
|---|---|
| Covered questions drop at the publication point and stay down, control questions stay level, applications do not fall | Keep the FAQ, and re-date it whenever SourceX's pages change |
| Covered and control questions fall together | Do not credit the FAQ; something else changed |
| Covered questions fall and applications fall too | The FAQ may be putting CEOs off; rewrite before keeping it |
| Covered questions do not move | Nobody reads it, or it answers the wrong questions; check which codes persisted |
| New wording questions or misreadings appear | Rewrite those answers, re-date, and run a fresh after-series |

Read the primary measure as counts first. With the numbers most partners have, the application guard can catch a collapse but not a small decline, and [the five-replies page](/guides/how-to-avoid-declaring-a-winner-from-five-replies) explains why small totals cannot settle a close call. Say that limit in the write-up rather than claiming the FAQ was neutral.

## Where a before-after count can mislead

- **No concurrent comparison.** Anything else that changed at the publication date, a conference, a news story or an update to SourceX's pages, moves the counts too. The control code and the three-and-three periods narrow the doubt; they do not remove it. A stronger version alternates outreach batches with and without the FAQ link, using the [hold-constant approach](/guides/how-to-compare-two-outreach-angles-without-changing-everything) for everything else.
- **Coding drift.** Whoever codes learns what the FAQ covers and may start filing questions differently. The frozen codebook and the blind second coder exist for this.
- **Some questions are good.** A CEO who asks three careful questions may be the most serious prospect you have. The aim is to remove repetition, not curiosity.
- **The FAQ ages.** SourceX may publish new versions of its terms (section 11). A dated FAQ shows when it was last true; an undated one cannot.

## 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.

Start the question log this week and draft the FAQ only from what it shows: [register as a partner](/register) if you have not, check every answer against [who qualifies](/who-qualifies) and the [program terms](/terms) on the day you date it, and find related designs in the [guides hub](/guides).

## Sources

- https://refer.sourcex.si/terms
- https://refer.sourcex.si/who-qualifies
- https://refer.sourcex.si/how-it-works
- https://epoc.cochrane.org/sites/epoc.cochrane.org/files/uploads/Resources-for-authors2017/what_study_designs_should_be_included_in_an_epoc_review.pdf
