A principal who screens fifty OMs in a quarter has fifty opinions about fit. Each opinion is informed by the principal's experience, the firm's strategy, the deals already in the pipeline, and the time pressure of the moment. These opinions are usually right. They are also untraceable. The deal that got declined six months ago for being "too small" cannot be reconstructed when a similar deal arrives, because "too small" was a judgment, not a record.
The shift from judgment to scored fit does not eliminate the principal's role. It separates the mechanical part of fit, which can be quantified, from the strategic part, which cannot. The mechanical part is whether the deal matches the criteria the firm has already articulated. The strategic part is whether those criteria are still right and whether this specific deal warrants an exception. A score handles the first. The principal still does the second.
The point of scoring is not to make decisions. It is to make decisions defensible.
What Scoring Replaces
Without a scoring system, every screening decision is a single sentence: "pass" or "no thanks." The rationale lives in the principal's head. When a broker asks why, the answer is reconstructed on the spot. When the firm reviews its own screening discipline at year-end, the data does not exist.
Without Scoring | With Scoring |
|---|---|
Pass or pursue, no rationale | Component scores with weighted composite |
Inconsistent across screeners | Uniform criteria applied to every deal |
No record of why | Logged rationale per dimension |
Hard to explain to brokers | Defensible feedback at the field level |
Patterns invisible | Decline reasons aggregable across deals |
Scoring does not slow screening down. It captures the rationale that was already happening in the principal's head and makes it queryable.
Component Scores Versus Composite Scores
A score that collapses to a single number is harder to use than a score that decomposes into components. A 68% composite tells the principal that the deal is borderline. A 68% composite with an attached profile of 95% on geography, 90% on size, 60% on returns, and 30% on tenancy tells the principal exactly what the deal is and why it scored where it did.
Component | What It Measures |
|---|---|
Geography fit | MSA, submarket, exclusion zones |
Asset fit | Type, subtype, vintage, condition |
Size fit | Deal size, unit count, equity check |
Return fit | IRR, multiple, yield-on-cost vs. targets |
Tenancy fit | Credit, mix, concentration |
Business plan fit | Strategy alignment with firm capability |
Process fit | Marketed vs. off-market, broker, timeline |
Each component is scored independently. The composite is a weighted average. The weights are firm-specific. A multifamily fund might weight size and geography heavily and tenancy lightly. A net-lease fund weights tenancy and credit heavily and treats size more flexibly.
Hard Filters and the Score Ceiling
Hard filters interact with scoring in a specific way: they cap the maximum score, but they also short-circuit the calculation. A deal that fails a hard filter does not score at all. It is logged with the failed filter and routed to auto-decline. There is no need to compute the rest of the components.
Hard Filter Outcome | Result |
|---|---|
All hard filters passed | Score the soft components |
One hard filter failed | Auto-decline with logged reason |
Multiple hard filters failed | Auto-decline with logged reasons |
This matters because the firm wants to spend zero principal time on deals that fail hard filters. The system declines them, logs the reason, and produces a courtesy response to the broker without any human in the loop.
What a Borderline Score Surfaces
The deals that earn the most value from scoring are the ones in the 60% to 80% range. These are the deals that get inconsistent treatment without a system. Some weeks a 70% deal gets a meeting; other weeks the same deal gets ignored. The inconsistency is not a discipline failure; it is a workflow failure caused by triage variance.
A scoring system with explicit thresholds removes this variance. A deal scoring 72% with the third-best score in the queue this week gets the same treatment as a deal scoring 72% with the first-best score next week. The principal's time is allocated to the top of the queue, and the queue is constructed from scores, not from inbox order.
Score Band | Default Routing |
|---|---|
90% plus | Top of queue, scheduled review |
75 to 89% | Standard review |
60 to 74% | Review if capacity, batched |
Below 60% | Auto-decline with rationale |
The thresholds are configurable. A firm in a slow quarter lowers the auto-decline cutoff. A firm with full pipeline raises it. The change is a single parameter, not a policy.
What a Score Is Not
A score is not a recommendation. It is not a probability of closing. It is not a statement that the deal is good or bad in absolute terms. It is a measure of how closely the deal matches the firm's stated criteria, weighted by the firm's stated priorities, calculated from the extracted data.
This distinction matters because firms misuse scoring when they treat it as predictive. A deal that scores 95% can still be a bad deal. A deal that scores 55% can still be a great deal. The score answers a narrower question: does this deal match what the firm has said it wants to buy? When the score is high and the deal is bad, the buy box is wrong. When the score is low and the deal is great, the buy box is too narrow. Both are correctable.
What the Score Is | What the Score Is Not |
|---|---|
Match to stated criteria | Predicted return |
Sort key for principal attention | Investment recommendation |
Defensible record of decision | Substitute for diligence |
Quantified version of the buy box | Quantified version of judgment |
The principal's job is to interpret high-fit deals on their merits and to interrogate low-fit deals where the firm wants to make exceptions.
Defensible Feedback to Brokers
The feedback a firm gives a broker shapes the next deal that broker sends. Vague feedback ("not a fit right now") produces vague follow-up. Specific feedback ("size below threshold; returns 40 basis points light; tenant concentration too high") produces submissions that hit closer to the firm's criteria.
A scoring system makes specific feedback the default, not the exception. The decline message can include the components that scored low and the buy box criteria they failed against. Brokers who get specific feedback over time submit better-fit deals. Brokers who get vague feedback either disappear or keep submitting noise.
The firm also benefits from reading its own decline reasons in aggregate. If 60% of deals decline on geography, the firm has a broker network problem in addition to a deal flow problem. If 40% decline on returns, the market has shifted relative to the firm's stated targets. Each pattern is a decision input that did not exist before scoring.
What "Done" Looks Like
A scoring system that produces defensible screening decisions meets the following criteria:
Every deal that passes hard filters receives component scores and a composite.
Every component is tied to a buy box criterion with documented weight.
Every screening decision references the score and the rationale.
Borderline deals are routed to consistent review thresholds.
Decline reasons are queryable across the firm's deal flow.
Broker feedback is generated from the same scoring data.
If a screening decision cannot be reconstructed from the scoring record alone, the system is incomplete.
Conclusion
Scoring does not replace investment judgment. It replaces the part of judgment that was always mechanical pattern-matching against criteria the firm could already state. The firms that scored every deal for a year know more about their own decision-making than the firms relying on memory and inbox order. They can defend declines, give specific broker feedback, and identify the patterns in their own flow that are pulling the buy box away from where the firm thinks it is. The score is not the decision. It is the evidence.