A firm's view of its own deal flow is filtered. The principal sees the deals that made it through triage, the deals their analysts thought were worth a second look, the deals that fit a template the firm has already closed against. The principal does not see the deals that were declined at step one, because by definition those deals never reached step two.
This filtering is rational. No firm has time to evaluate every OM in full. The cost of the filter is invisible, which is what makes it dangerous. The firm cannot count the deals it should have pursued and did not, because those deals never appeared on the firm's radar.
The best deals tend to be the deals that require reading to evaluate. They are also the deals most likely to be filtered out by surface-level triage.
The Selection Mechanism
Triage is a selection process. The selection criteria are usually informal and unwritten.
Selection Heuristic | Effect |
|---|---|
Familiar broker name | OM gets opened |
Cover-page price within range | OM gets read |
Recognizable market | Forwarded to principal |
Known asset type | Reviewed in detail |
Cover-page cap rate near target | Modeled |
A deal that fails any of these heuristics often does not progress. The firm calls this efficiency. The selection bias is built into the process.
The deals most likely to fail surface heuristics are not the bad ones. They are the ones whose value is hidden by structure: a broker the firm has not transacted with, a market the firm has not closed in recently, an asset type adjacent to the firm's core, a cover-page metric distorted by the deal's specific structure (assumable debt, tenant credit upside, deferred maintenance discount).
Where the Best Deals Hide
The deals that close at the highest returns tend to share certain characteristics. They are also disproportionately the deals that get missed in surface triage.
Characteristic | Why It Hides | Why It Outperforms |
|---|---|---|
Off-market or pre-market | No glossy OM | Less competition |
Distressed or repositioning | Headline numbers look bad | Embedded value-add |
Atypical structure | Doesn't match templates | Mispricing relative to standard |
Newer broker relationship | Not on top-tier mental list | Broker eager to prove value |
Adjacent asset type | Outside stated buy box | Cross-sector dislocation |
Tier-two market | Not in default geography | Stronger fundamentals than tier-one entry pricing |
Each of these deals requires reading the OM to recognize. A glance at the cover page produces the wrong impression. A model run on extracted financials produces a different impression than the asking metrics suggest. The work to surface them exists. It usually does not happen because the firm's triage layer disqualifies them first.
The Counterfactual Problem
The cost of selection bias is impossible to measure directly. The firm cannot point to the deals it missed because those deals never entered the system. What the firm can measure is downstream: deals that closed where the firm was not in the running, deals that traded at attractive returns where the firm received the OM but never engaged, brokers who stopped submitting after a year of getting no engagement on submissions the firm now wishes it had reviewed.
Counterfactual Signal | Where It Appears |
|---|---|
Closed deal report shows attractive returns on a deal you saw | The OM was in your inbox; you have no record of it |
Broker stops submitting after low engagement | The broker concluded the firm was not serious |
Competitor closes in your stated buy box | The OM was sent to your firm too; you passed at triage |
Year-end review shows fewer pursued deals than expected | The funnel narrowed at intake without anyone noticing |
These signals are weak individually. Aggregated, they describe a firm operating on a smaller subset of its actual deal flow than it realizes.
The Triage Bias as a Hiring Issue
The most consequential filter is usually the most junior person in the workflow. An analyst with eighteen months of experience is making first-pass calls on whether a deal is worth the principal's time. Their pattern recognition is necessarily limited to the deals they have already worked on. Their bias toward familiar templates is not a personal failing; it is a structural feature of where they sit in the process.
A firm that wants to broaden its deal flow has two options. The first is to hire more senior triagers, which is expensive and does not scale. The second is to remove the triage layer and replace it with structured extraction and scoring, which produces consistent first-pass output regardless of who is staffing the inbox.
Option | Tradeoff |
|---|---|
More senior triage | Expensive; bottleneck at hiring |
Structured intake | Capital expenditure; consistent output |
Status quo | Selection bias compounds with team experience |
The status quo is usually selected by default rather than by decision.
What Reading Every OM Produces
A firm that extracts and scores every inbound OM operates on a different dataset. The dataset is unfiltered by junior triage. The buy box scoring is consistent across submissions. The deals the principal sees are sorted by fit, not by who happened to open the inbox.
The first quarter of operating this way usually reveals patterns the firm was not expecting:
A broker the firm thought was inactive has been sending consistent fits that were getting deleted.
A market the firm "isn't in" has been producing 25% of high-fit submissions.
An asset subtype the buy box doesn't list has been showing up at attractive entry yields.
The principal is spending two-thirds of triage time on deals that score below the firm's threshold.
Each of these is a discovery. Each of them requires a buy box update or a workflow change. The firm that did not extract every OM did not have access to any of them.
The Compounding Coverage Effect
The advantage of reading every OM is not that the firm catches one missed deal. It is that the firm builds a database of what it has seen, which informs every subsequent decision.
Coverage Decision | Effect on Year Two |
|---|---|
Extracted every OM in year one | Buy box calibrated to actual flow; broker relationships data-driven |
Extracted only escalated OMs | Buy box reflects what got through triage; broker relationships anecdotal |
The firm with full extraction enters year two with a clearer picture of its own market position. The firm with partial extraction enters year two with the same biases as year one, plus another year of accumulated drift.
What "Done" Looks Like
A screening process that surfaces missed deals meets the following criteria:
Every inbound OM is extracted to the same field standard.
Every extracted OM is scored against the buy box.
The principal works from a queue ranked by fit, not by triage decisions.
Decline patterns are reviewable and feed buy box updates.
Counterfactual signals (broker drop-off, competitor closes) are tracked.
The firm can answer "what did we see and what did we pass on" with data, not memory.
If the firm cannot retrieve a record of every OM that arrived in the last twelve months and the rationale for each decision, the selection bias is operating without check.
Conclusion
Selection bias in deal screening is structural and invisible. It cannot be solved by discipline or by hiring senior people, because it operates at the moment of triage, before any senior judgment is applied. It can be solved by removing the triage layer entirely and replacing it with structured extraction and scoring. The cost of the bias is the deals the firm never sees, and those deals are disproportionately the ones with the most embedded value. The firms that close this gap will see more, learn more, and over multiple cycles outperform the firms still operating on the deal flow that happens to make it through their inbox.