The Sibling-Unit Expansion Playbook: Won Division to a Ranked Intro Slate

This playbook takes one closed-won business unit at a large company and turns it into a ranked slate of the sibling divisions and subsidiaries that could buy the same thing, dated so you know when it was checked. It is written for founders selling their own product, AEs, SDR leads, and partnerships teams who want to expand into other business units of an existing customer without paying for a hierarchy data subscription. Budget about two working days. The output is one row per sibling with a buying class, a named likely owner, a contract route, and a specific intro ask for your current champion.

What a sibling-unit slate is and why it is worth two days

A sibling-unit slate is a dated, ranked list of the other divisions and subsidiaries of a customer you already serve. Each sibling is classed by whether it can buy, matched to a likely owner, and given a contract route. It is worth building because published benchmarks show expansion deals closing faster, with fewer people involved, than new logos. Be aware that those benchmarks cover more than sibling units.

Three terms of art, defined once:

• Sibling unit: any business unit, division, or legal subsidiary under the same ultimate parent as the unit you closed.
• Champion: the person inside the won unit who wants you to succeed and has standing to introduce you.
• Whitespace mapping: the usual vendor definition is a grid of products against buying units, filled in from your own CRM and billing data. This guide fills the rows your CRM has never seen.

The case for the work comes from one large benchmark. The Ebsta x Pavilion GTM benchmarks covered 655,000 opportunities across 387 companies:

Read the table carefully. "Expansion" in that study blends upsell and cross-sell into accounts that are already buying. It is not limited to new business units. Part of the expansion advantage comes from the smaller stakeholder count, and a sibling that controls its own budget may go back to new-logo dynamics: its own owner, its own procurement, and its own eight people. The route that keeps the stakeholder count low is usually the shared-service route, where the buyer you already know buys on the sibling's behalf.

The second reason is multi-threading, which means having relationships with several people in the buying organisation instead of one. Gong Labs analysed 1.8 million opportunities. Deals that closed had twice as many buyer contacts as deals that did not, and multi-threading raised win rates by an average of 130% in deals over $50K. Gong also reports that large strategic deals won by reps include an average of 17 contacts. Every sibling introduction adds threads at the account level, which protects the renewal of the unit you already won as well as opening a new one.

Every row on a finished slate carries these fields:

• Unit name and legal entity
• Parent and reportable segment
• Buying class: buys alone, buys through a shared service, or out of reach
• Likely owner: name, title, source, date checked
• Contract route: affiliate users, affiliate order form, or new paper
• Fit score, path score, and rank
• The specific intro ask

Where the family tree comes from: free sources and their blind spots

Four free public sources together give you a usable family tree: Exhibit 21 in a US filer's 10-K, the ASC 280 segment note, GLEIF Level 2 parent records, and UK Companies House PSC records. Each one has a blind spot built into its rules, so cross-check them rather than trusting any single source.

Exhibit 21: the list of subsidiaries

Exhibit 21 is the list of subsidiaries attached to a US public company's annual report. What it proves: the listed entities exist and the filer consolidates them. How it lies: a filer may leave out subsidiaries if, considered in the aggregate as a single subsidiary, they would not be a "significant subsidiary" under Rule 1-02(w) of Regulation S-X. The threshold is 10%, applied through investment, asset, and income tests. Because the test runs on the omitted entities combined, a large group can drop many small ones. Those small, unlisted entities are often the exact siblings nobody in your CRM has seen.

ASC 280 segment note: how the company reports to itself

The segment note in the financial statements lists the reportable segments. A reportable segment is a part of the business whose results the chief operating decision maker (CODM) reviews separately. The CODM is whoever allocates resources across the company, often the CEO. Under ASU 2023-07, filers must disclose the CODM's title and position and the significant segment expenses regularly provided to the CODM, and the annual disclosures now extend to interim periods. What it proves: how the company organises its own reporting, and who sits at the top of it. How it lies: a segment can be purely a reporting construct with spend held at group level. That is the subject of the autonomy section below.

GLEIF Level 2: who owns whom, globally

GLEIF is the global registry of Legal Entity Identifiers (LEIs). Its Level 2 data has each entity report its "direct accounting consolidating parent" and its "ultimate accounting consolidating parent". A parent's record also shows its children's LEIs, so you can start at the ultimate parent and walk down the tree. How it lies: an entity can file a reporting exception instead of naming a parent. The allowed reasons include having no parent, opting out for exceptional reasons, and the parent having no LEI. Every one of those leaves a blank branch.

Companies House PSC: control in the UK

The free Companies House API covers companies, officers, and persons with significant control (PSC). PSC records show major shareholders (25%+) in any UK company, and the data may be reused free of charge, including commercially, as long as the source is acknowledged. How it lies: a public-data aggregator that republishes this dataset warns that "errors in the data must be assumed". The officer list causes more trouble than the PSC record itself, because a shared director looks like a family link when it is not.

What this guide did not verify: specific EU national registries, and US private-company sources beyond GLEIF. For a private US parent without an LEI, the cheapest reliable source is usually your champion's own org chart. Ask for it in the second conversation, not the first.

How to tell whether a sibling unit buys on its own

Class each sibling as buys alone, buys through a shared service, or out of reach, using the segment note, procurement evidence, and the titles people hold. No published procedure exists for this judgement. The signals below are partial, and each one can mislead.

The three classes:

• Buys alone: the unit controls its own budget and has its own decision-maker, usually a divisional IT, technology, or procurement lead.
• Buys through a shared service: a group procurement or shared-services function buys on the unit's behalf. The owner sits at group level, not in the division.
• Out of reach: you can find no budget signal and no reachable owner, or the unit does not fit the product.

The strongest public signal before any outreach is in the segment note. A filer may disclose that the CODM is regularly provided with only budgeted or forecasted expense information for a segment, or uses consolidated expense information. When you see that wording, start from the assumption of central control.

What the title data says about where owners sit

In Refolk's index of professional profiles, explicitly divisional IT titles are rare, and central buying titles are not:

The 16 US divisional or business-unit CIO and IT Director profiles sit at employers including Abbott, CBRE, WTW, Sysco, Tyler Technologies, and University Hospitals. All 3 UK profiles carry the exact title "Divisional IT Director", one of them at Bunzl Catering Supplies. On the central side, the most common title is Group Procurement Director, which is 9 of a 25-profile sample.

How to use these numbers: they are matches on headline titles only, not a census of how companies buy. They tell you not to read a missing divisional title as a missing owner. In the UK especially, the person who owns the decision for a sibling is more often a group function, or someone in the division with a generic title like "Head of Operations". The US figure is 5.3 times the UK figure, which suggests US groups are more likely to put a named technology owner in the division. Treat that as a reason to search, not as a conclusion.

This is the slowest part of the job: finding the real owner when the title does not announce it, across a dozen entities whose names you learned an hour ago. Refolk lets you ask for those people in plain English, by employer, unit, and function, and returns named profiles you can check against the family tree. That turns three hours of guessing at titles into reviewing a list.

Contract route: affiliate order under the existing MSA, or new paper

Before ranking anything, read the signed master services agreement (MSA) for three things: the definition of Affiliate, any cover-page authorisation for affiliate users, and a clause letting an affiliate sign its own order form. If those clauses exist, a sibling can buy without renegotiating the framework terms. If they do not, the sibling needs new paper.

The clearest public reference is the Common Paper Cloud Service Agreement, a standard SaaS contract template. It offers two routes. In the first, if the Cover Page authorises it, individuals from the customer's Affiliates may use the customer's account as Users under the customer's agreement. In the second, an affiliate that signs a separate Order Form with the provider creates a separate agreement between the provider and that affiliate, and the customer is not liable for it. Common Paper's own annotation says the point is that affiliates can enter their own Order Form "without having to renegotiate the Framework Terms."

What is not established publicly, so check it in your own paper:

• The control threshold. The Common Paper definition of Affiliate is based on control, but the exact percentage could not be verified for this guide. Read your own MSA's definition and check whether partly owned entities qualify.
• Pooled volume discounts. Nobody has published whether affiliate orders count toward the parent's volume tiers. Your own agreement decides.
• How common the clause is. Common Paper's help centre labels the affiliate option "Affiliates (uncommon)". Do not assume your agreement has it until you have read the text.

The two-day procedure, step by step

Seven steps take one won unit to a sent intro ask in roughly 11.5 working hours. No source documents this sequence. I put the order together from the sources above, and the timings are my estimates for a group with a dozen or so plausible siblings.

Who does what: the AE or founder owns steps 1, 2, 3, 6, and 7. An SDR can run steps 4 and 5 once the entity list exists. Legal or the deal desk joins for step 3 only. Two people working in parallel can finish in under two days, because steps 4 and 5 do not have to wait for step 3.

Use this as the working sheet. One row per entity, filled in left to right as the steps progress:

Slate date:
Ultimate parent:
Unit / legal entity:
Jurisdiction:
Direct parent (source):
Reportable segment:
CODM expense view (consolidated / by segment):
Buying class (alone / shared service / out of reach):
Evidence for class (source, date):
Likely owner (name, title):
Owner source and date checked:
Contract route (affiliate users / affiliate order form / new paper):
Signing entity / paying entity:
Fit (1-3):
Path (1-3):
Rank:
Intro ask (who, to whom, why now):

For step 5 on a large US group with real divisional owners, a query like this returns named people sorted by unit, which you then check against the segment map:

Ranking the slate: fit against path

Rank each sibling on two scores: how well it fits the product, and how easy the path to a signature is. Ask first about units that score high on both. Units with a hard path only earn effort when the fit is strong.

A simple rubric that holds up in practice. This is my scoring, not a published standard:

• Fit 3: same use case as the won unit, with public evidence (hiring for the role your product serves, or the same segment).
• Fit 2: adjacent use case, or the same segment with no direct evidence.
• Fit 1: plausible only in theory.
• Path 3: buys through a shared service you already sell to, or the cover page authorises affiliate users.
• Path 2: buys alone, and an affiliate order form is allowed under the existing MSA.
• Path 1: needs new paper.

When two units tie, rank the one the champion already knows above the other. The best introduction is one that costs the champion nothing.

Forecast honestly. A "worth new paper" sibling has its own budget, its own owner, and its own procurement, and nothing published says it will win at the 45% expansion rate. Forecast it closer to new-logo dynamics until it proves otherwise.

The champion intro ask

Ask your champion for one introduction to one named peer at the top-ranked unit. Give them a note they can forward without editing, and make it easy for them to decline. Ask for a peer first, not an executive.

The evidence here is thin, and I want to be plain about that. Nobody has published a response or meeting rate comparing champion-referred introductions with cold outreach to a colleague, and I found no published guidance on protecting the champion's standing. The nearest proxy is Gong's executive-timing finding, relayed by a secondary source: win rates drop approximately 6% when an executive is the first contact, and rise approximately 5% when executives are introduced around the third touchpoint.

That pulls against the segment note. ASC 280 points you to the CODM, because the CODM's title is the one named in the filing. The timing data says the CODM is the person you meet third, not first. Use the CODM title to understand who allocates budget. Use the peer introduction to get there.

How I protect the champion's standing. This is practice, not published evidence:

• Ask for one person, by name. "Anyone in the Industrial division" makes the champion do your research.
• Explain why that person and why now, using something public: a hire, a segment change, a shared project.
• Hand over a forwardable note so the champion does not have to describe your product in their own words.
• Offer an easy no. Make clear that "not the right time" is a complete answer.
• Report back within a day of the meeting. The champion's standing rises when an introduction they made goes well and they hear about it.
• If they decline, do not cold-email that person and mention the champion. Wait, or go through the group function.

Subject: One intro, only if it makes sense

Hi Priya,

Your team's rollout has gone well enough that I'm getting asked whether it would work elsewhere in the group. The one place I think it clearly fits is Daniel Osei's team in the Industrial division: they're hiring for the same role your team used us to support.

Would you be comfortable introducing me to Daniel? I've put a short note below you can forward as-is. If the timing's wrong or it's awkward for any reason, "not now" is a complete answer and I won't reach out to him myself.

Forwardable note:
"[Vendor] has been working with my team for [period] on [use case]. [One result]. Might be worth 20 minutes if you're dealing with the same thing."

Thanks,
Sam

How sibling-unit expansion goes wrong

Most failures come from treating a public source as complete, or treating a reporting structure as a buying structure. Each failure mode below has a check you can run before it costs you a champion's goodwill.

1. Exhibit 21 treated as complete. Small subsidiaries can be left out as long as they fall below the significance threshold in aggregate, so a missing name does not mean there is no entity. This hides exactly the siblings nobody has seen. Check: cross-check against the GLEIF Level 2 children of the ultimate parent.

2. GLEIF tree treated as complete. Parents without an LEI, and children that opt out of reporting, leave branches blank. A blank branch looks the same as no branch. Check: look at the reporting-exception record for each child. An exception that says "parent entities do not have an LEI" means there is a parent you have not found.

3. Officer list read as ownership. A director who sits on two boards looks like a sibling relationship but is not a control link. In one practitioner's words: "Do not read an officer list as an ownership list. Use the PSC record for control and the filings for the rest." Check: read the PSC record, and remember errors in that data must also be assumed.

4. Segment treated as a buying unit. A reportable segment can be a reporting construct with spend held centrally. Selling to the segment head then wastes a champion intro on someone who cannot sign. Check: does the note say the CODM uses only consolidated, budgeted, or forecasted expense data? If yes, class the segment as shared service until procurement evidence says otherwise.

5. No divisional title read as no owner. Only 3 UK profiles in Refolk's index carry divisional IT titles, against 196 central procurement or shared-services profiles. A missing "Divisional CIO" usually means the owner holds a generic title, or that buying is central. Check: look for separate procurement contacts for each entity, then search the division for people whose role matches your use case rather than your keyword.

6. Affiliate order assumed to inherit the parent's protections. Under the Common Paper standard, an affiliate's separate Order Form is a separate agreement and the parent is not liable for it. Finance assumes the parent's credit backs the deal, then the affiliate's payment terms surprise everyone. Check: confirm who signs and who pays before the order is drafted.

7. Expansion benchmark applied to a new business unit. The 45% Ebsta figure blends upsell and cross-sell into existing buyers. Promising it for a sibling with its own budget inflates the forecast. Check: forecast "buys alone" siblings closer to new-logo dynamics, and keep the expansion assumption for shared-service siblings only.

8. Champion asked to introduce a top executive first. The relayed Gong timing data shows a win-rate drop of about 6% when an executive is the first contact, and the request strains the champion. Check: reread the ask before sending. If it names a CODM or a C-level title, rewrite it for a peer.

A missing name in Exhibit 21 is the disclosure rule working as written, not proof the entity does not exist.

Keeping the slate current

A slate is only as good as the date at the top of it. Re-check it when a trigger fires, not on a fixed calendar: a new annual or quarterly filing, a changed segment note, a champion who moves roles, or a sibling that changes its procurement contact.

The mechanisms to watch:

• Segment notes now change mid-year. ASU 2023-07 extended the annual segment disclosures to interim periods, so read the latest 10-Q as well as the 10-K. A new reportable segment or a changed CODM expense view can move a sibling from one buying class to another.
• Exhibit 21 changes once a year. Diff the new list against your entity list. New names are new rows, and dropped names may be divestitures that should become out of reach.
• GLEIF records renew. A child that filed an exception last year may now name its parent, filling in a blank branch.
• Owners move. A named likely owner is a claim with a date on it. When the champion or an owner changes jobs, re-run step 5 for that unit. Refolk is useful here too, because the same plain-English query re-run against the unit shows who holds the role now.

Once the first introduction lands, go back to the slate and promote or demote the next units based on what the peer told you about how the group really buys. That conversation is the best evidence you will get, and it is free.