When to repeat a session and when to change the topic: a repeat-demand rubric that reads only what attendees asked and requested, never what anyone knows about their companies, and ends in one of five calls

Short answer

Repeat a session when people who missed it are asking for the same content and the answers held up. Rewrite it when a question kept coming back after the session had answered it. Change the topic when the questions arriving have moved to a subject the session never covered, and pause when nobody asked for anything. Make the call from the session's question log and written requests alone: founders' company facts, registrants' headcounts and any guess about who might qualify stay out of it.

Visual briefFull-size diagram
An Illustrative bump chart of five question themes ranked by the questions each drew across three runs of one session: sale or acquisition climbs from fifth to third to first and ends in change topic; who pays whom goes from third to first to second and ends in rewrite the answer; who decides falls from first to fourth to fifth and ends in answered; records that fit and customer contracts stay in the middle.
Illustrative: the three runs scored in the worked example on this page. Ranks come from attendees' own questions; registrations and company facts play no part.

This page is for the community or platform lead at a fund who has run a session for founders on company data licensing and now has to choose the next one. The calendar decision goes wrong in both directions: a full room gets rerun though its answers failed, or a small room gets dropped though the people who missed it keep asking. Neither registrations nor the prospect of an introduction says which question the community still has.

Five calls after a session, and what each one commits you to

CallWhen it fitsWhat you produce next
Repeat as isPeople who missed it ask for the same content, and the answers held upThe same session on a new date, announced to the whole community
Repeat with a rewritten answerA theme kept coming back after the session answered itThe same topic with the failing segment rebuilt, plus a written correction to the last audience
Change the topicQuestions moved to a subject the session never coveredA new session on that subject; the old content becomes a written page
Answer in writingA few questions went unanswered and nobody asked for another sessionA written note to everyone who attended, posted where the community reads
PauseNobody asked for anythingNothing; the topic waits until someone asks

Every call except Pause has a visible output, and Pause is a legitimate result. A community that hears from its host only when there is something to answer reads the next invitation more carefully. For one partner's consent-first habits across many introductions, the repeat referrals case study is the closest live example.

The audience question decides whether a session is the right format

Write the session's question down before the invitation, in the community's words, for example "How does a company license its records, and who decides?" Then pick the lightest format that answers it.

  • A live session suits a question many people share and that invites follow-ups.
  • A written answer suits a single question with a published answer; a rerun adds nothing to it.
  • A private route suits any question that needs one company's facts. Answer the general rule in the room and send the rest to the company's own conversation with SourceX, as the guide on questions that need private company context sets out.
  • A pre-session survey can collect the question itself; the founder survey design asks only what founders want covered.

Whatever the format, a session about SourceX carries the same few facts each time: the company's CEO decides whether it discusses licensing, and once a company is acquired the acquirer's CEO decides for it; funding does not make a company eligible; the host may be paid a referral reward; and nobody collects company data for SourceX.

The repeat-demand topic selection rubric

Five signals, each scored 0, 1 or 2, all read from what attendees and would-be attendees asked or wrote. Keep the session's Q&A record and one folder of written requests; nothing else feeds the score.

SignalWhat you countScores 0Scores 1Scores 2
S1 Carry-overQuestions still unanswered when time ran outNone1 to 34 or more
S2 Return questionA theme asked again, by someone else, after the session answered itNoneOne themeTwo or more themes
S3 Absent demandWritten requests for the session from people who did not attendNone1 or 23 or more
S4 DriftShare of questions on a subject outside the session's stated scopeUnder 20%20% to 39%40% or more
S5 Rerun requestAttendees asking in writing for a rerun, or for a colleague to hear itNone1 or 23 or more

Read the calls in this order and stop at the first match:

  1. S4 scores 2: change the topic.
  2. S2 scores 1 or more: repeat with a rewritten answer.
  3. S3 and S5 together reach 3: repeat as is.
  4. S1 scores 1 or more: answer in writing.
  5. Anything else: pause.

Drift comes first because a better version of the old session still misses a community whose questions have moved. Return questions come before demand because rerunning an answer that failed spreads the failure to a new audience.

Left out on purpose. Registrations and attendance count the invitation, not the need. Introductions requested are recorded for the community's own reporting but never steer the topic, since a calendar chosen to produce introductions becomes a funnel. Attendees' companies, their size, systems or likely fit, never enter the score, and neither does any estimate of rewards. How to read a single event beyond its registrations is covered in the event quality guide; this rubric only compares runs.

Where the inputs sit. On Zoom, for example, webinar reports include a Q&A report that shows each question with who asked and who answered it; attendee, Q&A, poll and survey data appear only after the webinar ends; and when a poll is relaunched, the poll report keeps only its last run, so a rerun that needs a comparable poll should use a new one (Zoom support, paraphrased). A post-webinar survey can carry one open question, "What should the next session answer?", and its responses download as a report (Zoom support, paraphrased).

Methods note. Measured: counts of questions, themes and written requests from the host's own Q&A record, survey and request folder, coded by one person within a week and checked by a second. The thresholds were set on 27 October 2026 as starting points; they are not tested values, and no benchmark for them has been published. Not measured: what attendees understood beyond the return-question signal, anyone who never attends, and anything about companies.

What the question log may hold, and who is invited back

  • Themes, not companies. Code each question by theme and strip company detail while you code: "could our warehouse scan history count?" becomes one tally under records that fit, with no name and no system.
  • Askers, then no names. Count distinct askers from the named report, then work from a copy with names removed. Questions sent without a name count as one asker each and stay unnamed.
  • Nothing imported. Board packs, CRM notes and what the fund knows about a founder's company never join the log. The rubric reads the session, not the portfolio.
  • Open invitations. A rerun is announced to the whole community, and the people who asked for it are told personally. Nobody is registered automatically, and declining needs no reply.
  • The host's interest, stated. Every invitation says the fund or its platform team may be paid a referral reward by SourceX if a company it introduces licenses data.
  • Nothing to SourceX. No list, names or themes tied to people leave the host. A company with a real next step takes it itself: its CEO applies to SourceX or asks for an introduction.

Three runs of one session, scored (Illustrative)

Illustrative. Invented counts for a generic fund community. The session was "How company data licensing works: what fits, who decides, who pays whom", and its stated scope was those three subjects, with customer contracts counted under fit.

ThemeRun 1Run 2Run 3
Who decides931
Records that fit763
Who pays whom576
Customer contracts312
Sale or acquisition (outside scope)159
All questions252221
SignalRun 1Run 2Run 3
S1 Carry-over2 (5 unanswered)1 (2 unanswered)1 (3 unanswered)
S2 Return question01 (who pays whom)1 (who pays whom)
S3 Absent demand1 (2 requests)1 (1 request)0
S4 Drift0 (1 of 25, 4%)1 (5 of 22, 23%)2 (9 of 21, 43%)
S5 Rerun request2 (4 requests)1 (2 requests)0
CallRepeat as isRepeat with a rewritten answerChange the topic

After run 1. Absent demand and rerun requests together reached 3 with no drift and no return question, so the host ran it again for people who missed it.

After run 2. Who pays whom drew seven questions, five of them after the segment that answered it. The host rebuilt that segment around one diagram: SourceX pays the company under the company's own agreement, and any referral reward goes to the partner separately. Section 4 of the terms says the reward is "paid separately by SourceX and is never deducted from the supplier's agreed compensation", and SourceX pays it from its own share, so a company receives the same amount with or without a partner. A short written correction went to run 2's attendees.

After run 3. Nine of 21 questions were about a pending sale or acquisition, which the session never covered. The host changed the topic: a new session on licensing questions during a sale, built on the published rule that once a company is acquired the acquirer's CEO decides for it, while before closing the company's own CEO still decides. The basics session became a written page. Who pays whom still drew six questions, two after the rebuilt answer, so the new session's pre-read repeats the diagram.

Measuring what the next run has to fix

  • Return-question rate per theme. Questions on a theme asked after the session answered it, divided by all questions on that theme. In the example, who pays whom went from 5 of 7 in run 2 to 2 of 6 in run 3. A falling rate means the rewrite is working.
  • Requests by kind, each over attendees. Rerun requests, written-answer requests, private-conversation requests and introduction requests, the last recorded and never targeted.
  • Corrections asked for. Each time an attendee points out a wrong or outdated statement, log it and answer in writing. The target is zero, and a nonzero count beats a hidden one.
  • What is not a measure. Seats filled, registrations and introductions say nothing about whether the community's question was answered. For more topic ideas drawn from client questions, see CPA firm client webinar topics and the guides library.

Ways the rubric can mislead you

  • Small numbers. With about 20 questions a run, one question moves a share by five points. Read the counts beside every percentage.
  • Loud askers. One person asking six questions can make a theme. Count distinct askers per theme as well as questions.
  • Unnamed questions. Questions sent without a name cannot be deduplicated, so a theme may be one persistent person.
  • The uninterested are invisible. People with no interest never ask, so the rubric can show that a topic went unasked, never that it is unwanted.
  • The facts move. Section 11 of the terms lists five versions published in October 2026 alone, from 2026-10-12 to 2026-10-22.2. A rerun with unchanged slides can repeat a superseded rule, so check the current version before every run.
  • The host's interest. A fund that can be paid for introductions has a reason to favor topics that produce them. That is why introductions are excluded from the score and why a second person checks the coding.
  • Demand outside the room. Whether company records stay in demand for AI training is beyond any community's control. No session should suggest that a given company's records will be bought.

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 the next run goes on the calendar, register as a partner if your platform team will make introductions, since the Channel Partner Agreement is signed before any is submitted, use who qualifies as the only source for the session's fit slide, and check every slide against the current program terms.

Choosing session topics is community work, and weighing SourceX as a possible investment is a separate exercise; anyone doing the latter can request investor materials through the SourceX investor overview.

This article is for information only and is not an offer to sell or a solicitation of an offer to buy securities. Any offering by SourceX is made only to verified accredited investors through its offering documents.

  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

Should a session that filled every seat be run again?

Not on that evidence alone. A full room shows the invitation worked; it says nothing about whether the answers held up or whether people who missed it are asking. Score the questions and written requests first, and rerun only if the rubric says so.

Can the topic choice use what the fund knows about which portfolio companies might fit?

No. The rubric reads only what attendees and would-be attendees asked or requested. Board-pack knowledge about a founder's systems or headcount stays out, both because it was shared for another purpose and because a topic chosen around likely introductions turns a community calendar into a referral funnel.

How soon after a session should it be scored?

Within a week, once the Q&A and survey data are available after the session ends. One person codes the questions by theme and a second person checks the coding, so the call does not rest on one reading.

Sources

  1. refer.sourcex.si/terms
  2. refer.sourcex.si/who-qualifies
  3. support.zoom.com/hc/en/article
  4. support.zoom.com/hc/en/article
  5. sourcex.si/investors

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