How to segment partners by real referral behavior: a recency, frequency and fit grid that replaces job titles, with an unproven row for anyone under three companies

Short answer

Segment partners by what their dated records show, not by job title, firm size or the network they claim. Three measures do the work: recency (days since their last proposed company), frequency (unique companies proposed in the last 180 days) and fit (the share of their companies that passed the published fit facts at a metadata screen). Leave reward out, because it lags by months and SourceX publishes no conversion rates. Anyone with fewer than three companies stays unproven until the records can carry a label.

Visual briefFull-size diagram
A two-by-two grid with proposal frequency across and fit rate up, holding Illustrative counts of steady sources, busy misfits, quiet experts and fading partners, with separate strips for dormant and unproven partners and a consent flag that overrides any cell.
Illustrative bench of 40 partners sorted with this page's rules; counts and names are invented.

This playbook is for a head of channel partnerships deciding where scarce support hours go across a bench of introducers. On 26 October 2026 the program facts were checked against SourceX's terms (Version 2026-10-22.2, since replaced by version 2026-11-04), its who-qualifies page and its home page. The grid, its cut-offs and the forty-partner bench are this page's proposal with invented figures.

What a title tells you, and what a proposal record tells you

A title tells you who is placed to know the right CEOs. SourceX's home page lists the roles best placed to refer: M&A advisors and business brokers, private equity and operating partners, fractional CFOs and accountants, IT managed-service providers, wind-down and restructuring advisors, and operators and executives. That list is a good recruiting filter. It is a poor management tool, because two fractional CFOs with the same client base can behave in opposite ways once they hold a partner link.

A proposal record tells you what someone actually did: which company they proposed, when, whether it passed the fit screen and whether its CEO knew. Activation, in this grid, is a partner's first proposed company that reaches your screen. Referral quality is the share of a partner's unique companies that pass the fit facts, read with the reason for each miss.

Customer analytics offers a precedent for grouping by behavior. Fader, Hardie and Lee's paper on RFM and customer lifetime value puts it plainly: "Measures of recency, frequency, and monetary value are sufficient statistics for an individual customer's purchasing history." Their iso-value curves exist to group customers "who have different behavioral histories but similar future valuations." This grid borrows the idea and changes one letter. Monetary value is replaced by fit, because a reward is rare, arrives months late and depends on SourceX's review, while fit is the part of quality a partner controls.

MeasureDefinitionWhy it is in the grid
RecencyDays since the partner's last proposed companySeparates partners who have gone quiet from those who are merely occasional
FrequencyUnique companies proposed in the last 180 days, counted once eachSets how much review and support a partner will use
FitShare of the last six unique companies, or all if fewer, that passed the fit screenThe quality signal the partner can improve
RewardNot usedLagging, rare and decided by conditions outside the partner's hands
Title, firm size, claimed networkNot usedUseful for recruiting, silent about behavior

The activity-and-fit segmentation framework

Apply the rules in order, each quarter, from dated records. A proposed company passes the fit screen only when it meets the published facts: a peak of 50 or more full-time employees (contractors do not count), records the company made itself over a span of years, the rights to license those records, and a qualifying location. 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. Partners answer these from what they know and what the CEO tells them; nobody asks for files, samples or logins.

  1. Evidence gate. Fewer than three unique companies ever: unproven. No other label yet.
  2. Recency. No proposal in the last 180 days: dormant, whatever the fit history.
  3. Frequency. Three or more unique companies in the last 180 days is frequent; one or two is occasional.
  4. Fit. At least two-thirds of the last six unique companies passed the screen is high; less is low.
  5. Overlay flags. A consent flag, any introduced CEO saying they did not know or agree, pauses the partner and overrides the segment. A duplicate-prone flag marks two or more proposals in 180 days that name a company already in your claim log. A concentration flag marks any partner behind more than a quarter of the network's fit-passing companies.
SegmentRule from the recordsSupport you giveSupport you stop
Steady sourcesFrequent and high fitFirst review slot; a quarterly 20-minute call about dated events they are seeingGeneric training and newsletters
Quiet expertsOccasional and high fitOne note when criteria change or a dated event hits their sectorVolume targets of any kind
Busy misfitsFrequent and low fitA session on their own declined rows; one proposal a month until two pass in a rowPraising their activity in partner updates
FadingOccasional and low fitOne review conversation, then let the pattern decideRecurring reminders
DormantNothing proposed in 180 daysOne changed-criteria note, then archiveEvery recurring email
UnprovenFewer than three unique companies everA plan for their first three companies, three open at mostAny label until the third company

Forty partners, sorted by what they did (Illustrative)

Illustrative: an invented network of 40 enabled partners, six months of records, 60 proposals naming 57 different companies.

The underperforming cohort. Three of the six busy misfits are MSP account leads recruited through one webinar. Twelve of their proposals were client companies with 20 to 40 full-time staff, or businesses whose useful records sit in their own clients' systems, which SourceX's guidance treats as usually not a fit. They are not careless; nobody showed them the screen. They get one session and a one-a-month cap.

The duplicate-risk case. Two account leads at the same MSP proposed the same 300-person manufacturer six days apart. One partner is credited per company, decided by SourceX from dated records, so the second proposal could only create confusion for the CEO. The second partner gets the duplicate-prone flag and the firm gets a shared claim log. Had the manufacturer emailed SourceX before either proposal, it would be a duplicate for both, and no partner would be credited.

Concentration. The five steady sources produced 15 of the 31 fit-passing companies. No single partner crosses the quarter line, but the network leans on five people for nearly half of its fit-passing companies.

SegmentPartnersUnique companies in 180 daysPassed fit screenTypical last proposal
Steady sources519153 weeks ago
Busy misfits62252 weeks ago
Quiet experts4657 weeks ago
Fading34111 weeks ago
Dormant700Over 26 weeks ago
Unproven1595Not used for this segment

The same forty partners by job title (Illustrative)

Sort the same bench by role and every role spreads across at least four segments. No profession is a segment. MSP account leads include one steady source and three busy misfits; fractional CFOs and accountants include two steady sources and three partners still unproven. A support plan built on titles would train the steady sources they do not need and miss the misfits who do.

The behavioral view also exposes what a title view hides: seven dormant partners spread over every role, still receiving the same monthly email as everyone else. Profession is still useful for recruiting, and for comparing whole roles with care, which the partner role quality study sets out.

Role (Illustrative)PartnersSteadyBusy misfitQuiet expertFadingDormantUnproven
Fractional CFOs and accountants9211113
MSP account leads11130124
M&A advisors and brokers6111012
Operators and executives8011123
PE operating partners6101013

Support hours, review capacity, concentration and overlap by segment

Segments earn their keep when hours run short. This bench sends about 10 proposals a month, comfortably inside a team that can review 20. When arrivals approach capacity, cut from the bottom of the priority order, never from the top: busy misfits' caps tighten first and steady sources keep their slot. The arithmetic for that moment is in when a partner network grows faster than review capacity.

Overlap needs its own rule because segments are per partner and companies are not. Keep one claim log for the whole bench, and treat a corporate group as one row: a parent and its subsidiaries count as one referred company, so the USD 50,000 floor on a first qualifying transaction applies once across the group and no cap applies to what follows. Whether an earlier introduction of one company in a group changes credit for a sibling introduced later is a question the terms leave to SourceX, which answers it while checking for duplicates, so write it in the log as open rather than guessing.

SegmentReview priority when capacity bindsCheck-in rhythmProposal capWatch for
Steady sources1QuarterlyNoneConcentration flag
Quiet experts2Event-drivenNoneCriteria changes in their sector
Unproven3Fortnightly until the third companyThree openConsent flag on the first introduction
Busy misfits4MonthlyOne a monthDuplicate-prone flag
Fading5OnceOne a monthA pass that warrants a fresh look
DormantNot applicableOnce, then archiveNot applicableA new proposal, which restarts the rules

Moving between segments, and letting partners correct the record

Movement should need evidence. An unproven partner gets a label at their third unique company. A busy misfit becomes a steady source only after two consecutive passes and a high fit rate on the last six. A dormant partner who proposes again re-enters the rules with their history intact. Re-run the grid quarterly; monthly changes mostly measure noise.

Every fit result carries the criteria version it was screened under, because SourceX's criteria follow each program and can change, and a screen is valid on the date it was made. Show each partner their own row, let them correct facts such as a company logged against the wrong person, and keep each correction with its date. Segment names stay internal: never publish them, rank partners by them or attach a promise of income to them. A partner's own tracking of companies and outcomes fits the repeat partner worksheet, and scoring one partner in depth, with CEO feedback as a gate, is the job of the fit-trust-follow-through scorecard.

Where behavioral segments mislead

  • Small counts. Six companies make a fit rate; they do not make a reliable one. The evidence gate helps, and a single miss can still move a partner.
  • Self-fulfilling support. Steady sources get faster reviews, which keeps them steady. Check now and then whether a quiet expert would become steady with the same treatment.
  • Seasonality. Advisors around deal closings and accountants around year-end propose in bursts. A 180-day window smooths some of it.
  • A fit screen is not acceptance. Relevance to current data needs and rights are decided in SourceX's review, so high fit means careful proposals, not accepted ones.
  • Borrowed models carry their origins. The RFM paper models purchase streams, and the paper itself notes that scoring models attempt to predict behavior in the next period. This grid predicts nothing; it allocates next quarter's support and is rebuilt each quarter.
  • Activity is not a channel. Whether a partner counts as active is a separate definition, kept on the registered-enabled-active ladder; disputes over who introduced a company first follow the seven evidence checks. The guides hub collects the rest of the partner-network playbooks.

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 your steady sources and quiet experts are clear, make sure each has registered for a partner link and signed, that their next candidates are checked against who qualifies, and that everyone on the bench has read the duplicate rules in the program terms.

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Company signs and is paidThe company signs, delivers its data and is paid by SourceX under its agreement.
  4. Step 4You get your rewardthe greater of $50,000 plus 5% and 15% of the actual supplier payout, with no cap. A reward becomes payable after SourceX has paid the referred company under its signed agreement and the payout conditions in the program terms are met, including finance approval and your tax and payment details.

Common questions

Why not segment partners by profession, since some roles are better placed to refer?

Role is a sensible recruiting filter, and SourceX's home page names the roles best placed to refer. But two people in the same role behave very differently once enabled, and support decisions are about individuals. Use role to decide whom to invite and behavior to decide how to support them.

Should a partner's reward history decide their segment?

No. A reward arrives months after an introduction, only after a completed deal and the payment conditions in the terms, and it depends on things the partner does not control. Segment on proposals and fit, which the partner shapes and you can date.

How often should partners be re-segmented?

Quarterly, from dated records, with the criteria version noted beside every fit result. Moving partners monthly mostly measures noise, and annual reviews leave dormant partners on recurring emails for too long.

Sources

  1. refer.sourcex.si/terms
  2. refer.sourcex.si/who-qualifies
  3. refer.sourcex.si
  4. brucehardie.com/papers/rfm_clv_2005-02-16.pdf

Free resources

Know a company with valuable proprietary data?

Become a referral partner, get your link and introduce the company’s CEO (in an owner-run company, usually the owner). Referral partners can join from most countries. We can't work with partners in countries or regions under U.S. sanctions or on U.S. restricted-party lists. The companies you introduce should be based in the U.S., Canada and the EU, or in the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia with at least 70% of their data in English.

Refer a company →

I own a business

Explore licensing your company's data directly to SourceX for AI training and evaluation. Start a short assessment; no uploads needed.

Start an assessment