Private equity and portfolios
Pooling competitors' records after a roll-up: antitrust limits
By SourceX Editorial · Reviewed by Noah Loul ·
Short answer
Pooling competitors' records after a roll-up can raise antitrust information exchange concerns when the combined data reveals current or future prices, wages, capacity or bids of identifiable companies, or reaches parties outside common control. A practical rule: pool only historical, aggregated, de-identified records for a written purpose, and get antitrust counsel's approval before anything leaves the group.
Key takeaways
- The most sensitive categories are current and future prices, wages and benefits, capacity, bids and customer-specific terms.
- Aggregation, aging, de-identification, exclusion and firewalls are the main handling tools, and counsel decides how far each must go.
- Before closing, buyer and target remain independent, so competitively sensitive information usually goes only to a clean team.
- Common ownership changes the analysis but does not settle it; minority stakes, co-investors and outside licensees need separate review.
- Published safe harbors are not a shortcut; guidance in this area has changed, so confirm current guidance with counsel.
Why does a roll-up create information exchange questions?#
A roll-up creates information exchange questions because it brings together companies that recently competed for the same customers, workers and contracts, and some of them may still compete or remain partly owned by others. Antitrust laws, including the Sherman Act, are concerned with competitors coordinating on price, wages or output, and sharing sensitive information can be evidence of that coordination or a channel for it.
The questions arise at three points. Before closing, during diligence and integration planning. After closing, when the platform wants benchmarks, shared pricing tools or a combined dataset across add-ons. And when pooled records are licensed or shown to an outside party, such as an AI developer, a benchmarking vendor or a pricing software provider that also serves the group's competitors.
Which categories of information are most sensitive?#
The most sensitive categories are those a competitor could use to adjust its own prices, pay or output: current and future prices, wages and benefits, capacity, bids and customer-specific terms. Age, granularity and whether the source company can be identified all change how sensitive a given record is.
| Category | Where it shows up in roll-up records | Why it is sensitive | Common handling rule |
|---|---|---|---|
| Current and future prices | Price books, estimates, quotes, rate cards, discount approvals | Can align pricing between competitors | Exclude current and planned prices; use aged history only if counsel agrees |
| Wages and benefits | Technician pay, bonus plans, offer letters, recruiter fees | Can align pay and hiring terms in labor markets | Exclude from pools or aggregate across many sources |
| Capacity and output | Crew schedules, truck counts, warehouse space, production plans | Can support supply restriction or market allocation | Share only historical, aggregated figures |
| Bids and pipeline | Open RFP responses, win and loss notes, backlog | Reveals where and how a company will compete | Exclude open bids; firewall closed bid history |
| Customer-specific terms | Contract pricing, margins by account, renewal dates | Enables customer allocation or targeting | Remove customer names and account-level terms |
| Costs and margins | Job costing, supplier pricing, margin reports | Can signal pricing floors | Aggregate and age before any shared use |
| Strategic plans | Expansion targets, planned price changes, territory maps | Forward-looking and company-specific | Keep within the owning company |
- 01Your work records
- 02Personal details removed
- 03You approve the release
- 04Licensed to AI developers
Handling rules: aggregation, aging, firewalls and exclusion#
Handling rules reduce sensitivity by changing what a recipient can learn from pooled records. No single rule is enough on its own, and the right combination depends on the market, the number of contributing sources and who will see the result.
| Rule | What it means | Where it helps | Limits |
|---|---|---|---|
| Aggregation | Combine data from enough sources that no one company's figures can be backed out | Benchmarks and pooled datasets | Small markets or a dominant source can still be identifiable |
| Aging | Use historical records only, never current or forward-looking data | Price and wage history | How old is old enough depends on the market and counsel's view |
| De-identification | Remove company, site, customer and employee identifiers | Licensed record sets | Details in free text can still reveal the source |
| Exclusion | Drop fields or whole record families | Price books, open bids, pay data | Reduces the usefulness of some records |
| Firewalls and clean teams | Restrict raw data to a small group, often outside advisers | Diligence and integration planning | Needs written protocols and discipline |
| Purpose limitation | Written limits on how pooled data may be used | Internal benchmarks and licenses | Only as strong as its monitoring and enforcement |
Do information exchange safe harbors still help?#
Information exchange safe harbors help less than many deal teams remember, and they should not be treated as a shortcut. Some practitioners still apply numeric aggregation and aging tests from older agency guidance, but guidance in this area has changed, and counsel should confirm what current guidance and case law say before relying on any threshold.
The safer habit is to start from the purpose. Write down why the pooled data is needed, who will see it, and what decisions it will inform. A benchmark meant to improve dispatch efficiency needs different fields from a dataset meant to train a scheduling model, and neither needs current price books.
Before closing: gun-jumping and clean teams#
Before closing, the buyer and the target remain independent competitors, so the target should not share competitively sensitive information beyond what diligence genuinely requires, and the parties should not coordinate conduct. Where premerger notification under the Hart-Scott-Rodino Act applies, its waiting period adds further limits on acting as one company too early.
Clean teams are the usual tool. A small group, often outside advisers plus employees with no pricing or hiring role, reviews sensitive data under a written protocol and reports only aggregated conclusions to the deal team.
- Name clean team members, and exclude anyone who sets prices, wages or bids.
- Define which data categories go only to the clean team.
- Report conclusions in aggregated or summarized form.
- Log what was shared, with whom and when.
- Return or destroy clean team materials if the deal does not close.
After closing: what changes and what does not#
After closing, sharing among companies under common ownership and control is often analyzed differently from sharing between independent competitors, but the details depend on the ownership structure. Wholly owned add-ons, majority stakes with minority holders, companies held by different funds of the same sponsor and co-investments can each be treated differently.
Outside parties change the picture again. A pooled dataset licensed to an AI developer, a benchmarking firm or a pricing tool vendor that also serves the group's competitors can become a channel between competitors. Keep licensed sets historical and de-identified, restrict re-identification and onward disclosure in the contract, and let counsel decide whether a given category belongs in the license at all.
Illustrative: a home services roll-up reviewing a pooled dataset#
Illustrative: a fictional HVAC and plumbing roll-up, Tarnbrook Comfort Partners, owns several companies in neighboring metro areas, running on ServiceTitan or Housecall Pro. Most add-ons are wholly owned; one is a minority investment alongside its founder. The platform wants to pool estimates, completed jobs and technician pay records for an internal benchmark and a possible license.
Holdco counsel removes the minority-owned company from the pool pending separate review. Technician pay, current price books and open estimates are excluded. Completed job records stay, with prices aged, customer and company names removed and service areas generalized. The internal benchmark shows aggregated ranges to the platform finance team only, the license draft bars re-identification and use for pricing recommendations, and each company's signer approves its own contribution.
How SourceX handles competitively sensitive fields#
SourceX handles competitively sensitive fields as a scope question within the Rights and Preparation steps of the SourceX five-step transaction. Price, wage, capacity and bid fields are flagged during the rights review, and preparation removes, ages or aggregates them as the supplier's counsel directs.
SourceX does not give antitrust advice. Each entity's counsel and signer decide what is included, and the SourceX Evidence Packet records the permitted use, the privacy record and what was excluded, so the decision trail is clear if anyone asks later.
Frequently asked questions
Does common ownership remove antitrust risk from sharing between add-ons?
Not entirely. Sharing among companies under common control is often analyzed differently from sharing between independent competitors, but partial ownership, co-investors, separate funds and pre-closing periods complicate the picture. Outside recipients of pooled data raise separate questions. Counsel should map the ownership structure before any pooling begins.
Can the platform build internal benchmarking dashboards from add-on data?
Often, with care. Many groups limit dashboards to aggregated, historical figures, restrict who sees company-level detail, and keep pay and current pricing out. Where an add-on is not wholly owned or still competes independently, counsel may require tighter limits or leave that company out.
Are historical price records acceptable to include in a license?
There is no blanket answer. Sensitivity generally falls as records age, are aggregated and lose identifiers, but how much is enough depends on the market, the number of sources and who receives them. Counsel decides, and preparation then applies the agreed rules consistently.
Why are wage and pay records treated so carefully?
Companies that compete for workers can harm labor markets by aligning pay, much as sellers can harm customers by aligning prices. Technician, driver or recruiter pay data shared among companies hiring from the same pool can raise concerns, so many roll-ups leave pay records out of any pooled set.
Who should sign off before a pooled license goes out?
Antitrust counsel for the group, each contributing company's authorized signer and, where relevant, the sponsor. Their decisions should be documented: which companies contributed, which categories were excluded or aged, and what use restrictions bind the recipient.
Related resources
See if your company qualifies
A short company assessment. No data uploads are needed.