# Referral program conversion uncertainty memo | SourceX

Canonical: https://refer.sourcex.si/guides/a-referral-program-with-strong-fit-but-uncertain-conversion
Last updated: 2026-11-04

Strong fit means your clients look like the companies a program describes. It says almost nothing about how many will convert, because the decisive steps happen outside your firm: the CEO's interest, SourceX's duplicate and rights review, the program's current needs and the Acquisition Partner's final written acceptance. SourceX publishes no conversion rates. So size your firm's part by the cost it can absorb if nothing converts, learn the rate cheaply through a small dated pilot, and write down in advance which result means expand, maintain or decline.

The mistake to avoid is treating a long list of good-looking clients as a forecast. A list measures your screen. Conversion measures everything after it.

## Fit is the part you can see; conversion is decided elsewhere

Trace one strong-fit client through the steps that follow your screen and mark who decides each.

| Step after a strong-fit screen | Who decides | Can your firm know it beforehand? | How the firm finds out |
|---|---|---|---|
| The CEO wants the conversation | The client's CEO | Partly; you can ask | A permissioned ask that is easy to decline |
| The company is new to SourceX | SourceX's records; any earlier lead, account or email counts as a relationship | No | Duplicate review |
| The records match a current program | SourceX, in review | No; criteria follow the specific program | SourceX's review |
| The rights to license hold up | SourceX's rights review, with the company | Partly, through metadata questions | The rights review |
| Final written acceptance | The Acquisition Partner | No | SourceX status |
| Supplier paid, reward payable | SourceX, under the payment conditions in the terms | No | A payment record |

One step of six sits even partly with your firm. That is the whole case for treating conversion as uncertain rather than merely unknown: no amount of extra screening will move the other five into your column.

## The client problem, and the firm's real part in it

The client's question is whether records it already keeps could be licensed for AI training and evaluation, and on what terms. Your firm cannot answer that, and should not try. What it can do is recognize a likely fit, ask the CEO whether they want a conversation, make the introduction and step back. SourceX handles the agreement, rights review, data preparation, delivery and payout with the company directly. Partners never ask for, receive or handle raw client records or credentials; every question stays at metadata level: which records, roughly how many, how many years, which system, who owns them.

The firm's exposure is therefore narrow but real: the hours spent, the client's view of its judgment if the conversation goes nowhere, and any professional rule on referral compensation. CPA firms in particular should sort clients before fees; the [attest versus non-attest comparison](/compare/cpa-referral-fee-attest-client) explains why.

## Why a strong-fit book invites an optimistic forecast

In their Harvard Business Review article "Delusions of Success", Dan Lovallo and Daniel Kahneman argue that cognitive biases such as anchoring and competitor neglect, combined with organizational pressure, push managers toward overly optimistic forecasts. Their remedy is the outside view: look at a class of similar projects, lay out a rough distribution of how they turned out, and place the current project within it.

A firm looking at fourteen clients that all pass the fit screen is taking the inside view, built from what it knows about each client. The outside view would need outcomes from comparable introductions, and none are published for this program. The honest substitute is a reference class the firm builds itself: its first few introductions, logged with every ending, including the dull ones.

## The scenario-based feasibility memo

Write the memo before the first ask. It has seven parts, and none of them is a revenue line.

1. **Client problem**, in one sentence the client would recognize.
2. **Firm's role**, including what it will never do: handle records, promise acceptance, quote a payout.
3. **Fit evidence**: how many clients pass the four published fit facts at metadata level, as a count.
4. **Unknowns**: the step table above, with who resolves each.
5. **Participation model** and the hours budget, set as if income were $0.
6. **Stopping scenarios**: where introductions might stop, and the firm's response to each.
7. **Review date and decision rule**: which observed result leads to which decision.

The scenarios are defined by where the process stops, not by odds. Nothing in the memo multiplies a reward by a probability.

| Where introductions stop | What the firm spent | Effect on client trust | What the firm learns | Response |
|---|---|---|---|---|
| At the CEO: declines or no answer | Ask time only | Small, if the ask was easy to decline | The screen finds companies, not interest | Decline, or stay respond-only |
| In SourceX review: duplicate, rights or program fit | Asks, follow-ups, a few check-ins | Neutral, if expectations were set early | Which fit fact or rights pattern fails in your book | Narrow the screen and maintain |
| At least one reaches acceptance and payment | All of the above, plus months of waiting | Positive, if the client valued the process | The path works for this kind of client | Consider expanding |

## Two ways to take part, and the option to say no

| Dimension | Model A: respond only | Model B: five-CEO pilot | Decline |
|---|---|---|---|
| What the firm does | Introduces only when a client raises data licensing, or a dated event arises such as a documented system migration | Asks five strong-fit CEOs within one quarter, then reviews | Nothing, and records why |
| Effort | Minutes per year until a client asks | A fixed hours budget, logged weekly | None |
| Terms to handle | Channel Partner Agreement signed before submitting; 90-day window; one partner per company | The same, for five companies at once | None |
| Trust risk | Lowest; the client started it | Moderate; five asks the clients did not request | None, though a client may later ask why no one mentioned it |
| Economics | Rare, unplanned rewards | Possible rewards months out; budget assumes $0 | None |
| Learns conversion? | Very slowly | Yes, for this kind of client, within two or three quarters | No |
| Right when | Fit is real but the firm has no capacity | Fit count is five or more and the firm would make the asks unpaid | Rules bar compensation and the firm will not introduce unpaid, or rights patterns fail |

The [canvas for whether to add a data referral program](/guides/should-your-firm-add-a-data-referral-program) asks the earlier question, whether to take part at all. This memo assumes the fit question has a yes and the conversion question has no answer yet.

## One firm, two memos (Illustrative)

Illustrative: an invented 30-person operations advisory firm serving regional logistics and claims companies. Fourteen clients pass the four fit facts at metadata level.

**Memo 1, Model B.** The firm budgets 24 hours over a quarter and asks five CEOs, chosen by fit facts, not by friendliness. Two quarters later: three CEOs agreed to talk, two companies submitted verified applications, one is still in SourceX's review, one was declined because its strongest records belonged to its customers, and nothing has been paid. Hours used: 19. Decision: **maintain**. The firm stays respond-only for the other nine clients and runs a second group of five only after the open case resolves. If that company later reached Closing and received a $250,000 payout, the reward would be $62,500 (Illustrative: USD 50,000 plus 5% of the payout, which exceeds 15% of it); the memo still records $0 until the payment conditions are met. A further qualifying transaction with that company first agreed within 12 months of that first Closing would pay 15% of its own payout, with no cap per company.

**Memo 2, a different practice group.** The same firm's accounting affiliate looks at its own book and finds most strong-fit clients are attest clients. Its ethics partner rules out compensation for them. The group's memo says **decline** for compensated introductions and allows an uncompensated introduction only where a client asks first.

## Four verdicts for a strong-fit book, each with its reason

| Decision | Choose it when | Reason to record |
|---|---|---|
| Pilot | Five or more clients pass the fit facts and the firm would make the asks unpaid | The firm needs its own reference class |
| Maintain | The pilot ended with open cases or mixed endings | Evidence is incomplete, not negative |
| Expand | At least one company reached final written acceptance, the endings show which screen works, hours stayed in budget and clients found it useful | The path works for this client type |
| Decline | CEOs consistently do not want the conversation, the rights pattern fails, or rules bar compensation and the firm will not do it unpaid | The cost is certain and the benefit is not |

For the cash side of the same decision, the [probability-free income planner](/guides/how-to-model-referral-income-when-no-deal-is-guaranteed) keeps every amount at zero until it is payable.

## Which clients belong in the fit count

Count a client only when, at metadata level, its full-time headcount peaked at 50 or more (contractors excluded), its operating history is documented, it owns or can license its records, and its CEO would want the conversation, since the CEO is the one who approves licensing talks and signs on the company's behalf. 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 a staffing firm, only its own full-time employees count; describe placed workers separately and introduce the firm anyway when its records are strong. Count a corporate group once: a parent and its subsidiaries are one referred company, and the USD 50,000 floor applies once, to that company's first qualifying transaction.

Even a client that passes every check above leaves three questions to SourceX's review: is there a current program its records match, do customer contracts restrict reuse, and what would a license cover?

## Where this memo can steer a firm wrong

- **Five is not a sample.** A pilot shows whether the path works for your clients, not a conversion rate.
- **Slow is not no.** A company still in review on the review date is an open case, not a failure.
- **Friendly selection.** Picking the five warmest clients flatters the pilot; pick by fit facts.
- **Moving criteria.** SourceX applies the criteria of the specific program, so a decline this year may not hold next year.
- **Unlogged goodwill.** Record any client who found the ask unwelcome; that cost never shows in a reward line.

For the basic mechanics of business referral rewards, see [what a referral program is](/questions/what-is-a-referral-program), and for consulting-specific routes, the [management consultants page](/for/management-consultants).

## 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.

If your memo says pilot, [register for a partner link](/register) and sign the agreement before the first ask, check each of the five against [who qualifies](/who-qualifies), and keep the [program terms](/terms) beside the memo's review date. Related methods sit in the [guides index](/guides).

## Sources

- https://refer.sourcex.si/terms
- https://refer.sourcex.si/who-qualifies
- https://hbr.org/2003/07/delusions-of-success-how-optimism-undermines-executives-decisions
