Munger's perhaps most cited piece of advice is deceptively simple: invert, always invert. To solve a hard problem, don't just ask "how do I achieve X?" Ask "what would guarantee I fail to achieve X?", then avoid those things. The path forward is often clearest when viewed from the opposite direction.
Most CRE acquisitions teams don't use this principle nearly as far as they could. They build investment theses, then sit back and wait for the deals that brokers send them. The filter runs inbound.
This is not wrong. It's just half the job. The reason only half the job has been done until now is simply that it was infeasible to take in the observable universe of all transactions and filter out what matches your firm's buy box. Until now.
Munger's inversion, applied to deal sourcing, asks a different question: if your criteria defines exactly what you want to buy, why isn't it also telling you exactly where to look? Every filter you've constructed is not just a tool to validate an incoming deal but rather a description of the universe of properties whose owners you should be calling just before those deals ever hit the market.
The firms winning the best deals are not necessarily seeing more opportunities. They're using the same criteria everyone else has, but running it in both directions.
The framework for inverting inbound filters into outbound sourcing tools has three layers, each corresponding to a different dimension of what your criteria are expressing.
Layer 1: Fit Criteria as a Property Universe Query
Your fit criteria (size range, submarket, asset class, vintage, occupancy threshold) is a precise database query waiting to be run proactively. Most teams only run it reactively, against deals that have been sent to them. The inversion is to run it against the entire MSCI universe of owned assets.
If your mandate is 20,000–35,000 SF light industrial, infill markets, sub-5% vacancy, built 2000 or later, that's not just a filter. It's a description of roughly 1,200 assets across markets like Dallas–Fort Worth, Atlanta, and Phoenix. You don't need a broker to tell you those buildings exist. You can know their ownership structure, their loan maturity schedule, and their historical transaction behavior today.
The key shift is treating your fit criteria not as a yes/no gate on arriving deals, but as a property-level SQL query you run on the market continuously. The output is not a ranked deal list. It's an outreach list which the agent uses as a starting point for investigation, looking into a multitude of criteria including the transactability score, which we'll get into next.
Layer 2: Transactability as a Timing Signal
Transactability scoring, which quantifies seller friction, asset friction, and market friction, is typically used to evaluate whether a deal in front of you is worth pursuing. Inverted, it becomes a timing model for which owners are likely to transact in the next 90–180 days, before they've decided to sell.
Loan maturity is the clearest signal. A fund with $40M of floating-rate debt maturing in nine months, on an asset that hasn't been refinanced or listed, is most likely a motivated seller candidate whether or not they're officially "on the market." MSCI Real Capital Analytics often surfaces this data. Most acquisitions teams look at it when evaluating an OM. The inverting move is to run a loan maturity screen on your target property universe every day, and surface motivated owners before a broker does.
The asymmetry is significant. When a high-transactability signal appears on an asset in your fit universe, you usually have a 2–4 month window before the owner decides to formally list, engage a broker, or take another path. That window is the competitive advantage.
Layer 3: Learning-to-Rank as a Feedback Loop
Learning-to-rank (LTR) models are typically framed as tools for personalizing what a user sees. Inverted, they are something more valuable: a mechanism for continuously refining which outbound targets are worth prioritizing, based on your team's revealed preferences.
Every time your acquisitions team engages with a deal (views it, saves it, contacts a broker, pursues it, or passes), the LTR model learns something about what your firm actually buys versus what it says it buys. Stated preferences are criteria you've written down. Revealed preferences are the pattern your behavior encodes over time.
The gap between the two is real at most firms. The mandate says Class A; the last six deals the team pursued were B+ buildings at a discounted basis. Nobody wrote that down anywhere, but the behavior encodes it, and the model notices before anyone says it out loud.
Nobody wrote the right-hand column down. The pattern is encoded in which deals the team viewed, saved, pursued, and passed on, and the ranking model weights it accordingly.
The Compound Effect: Why All Three Layers Together
Each layer of the framework, run in isolation, provides an incremental edge. Run together, they produce a qualitatively different capability: a continuously updated, intelligence-ranked outbound sourcing pipeline that requires no additional research overhead from your team.
The mechanics are straightforward. Layer 1 defines the universe: roughly 800 to 1,500 assets depending on mandate breadth. Layer 2 filters that universe down to the 80 to 150 owners who show elevated transaction motivation signals right now. Layer 3 ranks those 80 to 150 by predicted relevance to your team's actual buying behavior, surfacing the 20 to 30 worth a personal call this week.
The result is not a list of deals. It's a weekly conversation priority list, generated from your own criteria, running in the direction you haven't been using it.
Consider the arithmetic of the traditional approach. An acquisitions team at a $1–2B fund receives perhaps 200 deals per quarter from broker relationships. Of those, perhaps 15 to 20 meet basic fit criteria. Of those, 2 to 3 close. The close rate on reviewed deals is 10 to 15 percent, not because the team is bad at underwriting, but because the deal selection entering the funnel is random relative to seller motivation.
Invert the filter, and the funnel changes shape. Instead of 200 random inbound deals with a 10 percent close rate, you're pursuing 20 proactively identified deals with measurably elevated seller motivation, and a close rate that reflects that motivation.
A middling transactability score pursued proactively beats a perfect fit score pursued reactively. Fit tells you what you want. Transactability tells you what will close.
Layer 1 finds ≈1,200 assets that fit the box. One Layer 2 screen marks ≈600 of them as outreach candidates. No broker, no OM, just the same criteria running in the other direction.
The competitive moat compounds over time. The LTR model improves as your team accumulates behavioral data. The transactability signals update weekly as loan maturities approach, occupancy shifts, and market conditions evolve.
The Practical Starting Point
The inversion doesn't require a technology overhaul. It requires a reframing of what your criteria is for.
The first step is to run your existing fit criteria as a universe query rather than a deal filter. Export that universe, with loan maturity dates, ownership type, and occupancy history. That list exists today, in the data you're likely already paying for.
The second step is to identify the top 10 percent of that universe by transactability signals right now. Those are your calls this month. Not because someone sent you an OM, but because the math says they're likely to transact.
The third step is to track every deal your team engages with, regardless of outcome. The behavioral data you're generating every week is the raw material for a model that will eventually rank your entire target universe by predicted relevance. The sooner you start collecting it, the sooner that ranking becomes useful.
Munger's inversion is not a platitude. In CRE deal sourcing, it is a specific, executable instruction: take the criteria you use to filter what comes to you, and run it in the other direction. The deals worth finding are already in the data. They just haven't been sent to you yet.
Data insights courtesy of MSCI Real Capital Analytics. Counts and examples are illustrative.