When the marketed pipe thins, a large share of the year's deals are found before a broker is hired. The transaction record says which owners are likely to sell and where deals in their submarket actually clear. Starboard uses it to put those owners in front of your desk, and then to underwrite whatever arrives, marketed or sourced, in minutes and cited.
Institutional deals rarely go to market with a price on them. What arrives is a whisper number, a guidance range, a call for offers with a reserve nobody writes down. Those expectations are set by people looking backward at the last set of comps, so they lag the market in both directions. In 2021 brokers guided low to spark bidding wars and the prints ran past the number. Since 2023 reserves have sat above where buyers will pay, and a seller anchored above the market gets one of two outcomes: no trade, or a pulled process. The gap between expectation and market shows up first as silence. It is in the volume long before it is in a print.
That silence is where the sourcing year is decided. When the marketed market goes quiet, the deals that still happen are the ones a buyer went and found before a broker was hired. The MSCI Real Capital Analytics transaction record tells you where to look. It holds two halves of one object: the probability an asset trades inside a given window, read from volume and hold periods by submarket and property type, and the distribution of where it clears given that it does, read from repeat sales. Together they are what Starboard means by transactability.
Where the market is still moving, and who will still sell
Falling volume means fewer marketed deals, and more of the year's trades coming from owners who were approached directly. When sellers hold their reserve and bidders hold their number, processes get pulled, owners stay put, and the count of closed sales thins. Every trade is the outcome of a private go/no-go decision, which makes the volume series the closest available reading of transaction probability at the market level. Divide it by the stock of assets that could have traded and the count becomes a rate, the share of owners in a submarket who actually sold in a given year; that denominator, turnover, is what sharpens a count into a probability. Tracked by submarket and property type, it tells your desk where the market is still moving and where it is frozen. In the frozen places the hold-period record says which owners are still likely to sell, so you can reach them before a process exists.
Read the collapse after the 2021 peak as the expectation gap in data. Reserves stayed where the 2021 prints had put them, buyers' numbers moved, and the result was fewer trades, skewed toward owners who had to sell. Those are the owners a sourcing desk wants to reach first, and what they walked away with is the second half.
Where it clears, given that it does
You already run a back-of-envelope: thirty minutes with the rent roll, the T-12 and a cap rate tells you whether the deal can work. The weak point is the benchmark: the whisper and the guidance lag the market, so in a falling market you pass too early on deals that will clear well below guidance, and in a rising one you chase deals that will run past it. The clearing distribution replaces the whisper with where deals in that submarket and property type have actually printed. This is the second thing Starboard is built for: getting to the analysis quickly. When the OM lands, or the owner you sourced agrees to speak, Starboard screens the deal on arrival and underwrites it in your templates, comps and risks included, every number cited to the cell, checked against the clearing distribution for that submarket and property type and against your own deal history. The same screen on every deal, marketed or sourced, in minutes, so your desk sees more and decides faster.
Repeat sales are the cleanest window into seller outcomes: the same asset, two prints, one owner in between. Annualize each seller's price change from their own purchase to their own exit, and the outcomes of a given exit year form a wide, moving distribution. It is a record of what sellers walked away with relative to what they paid, before debt, capex and income, so it is a price change and nothing more. Its value to you is the shape. A cap-rate headline or a broker's guidance is one summary of that shape; the shape itself is what you are bidding into. Drag through the exit years below and watch it travel.
The figure is evidence of a regime shift, and two things follow. Guidance set against 2021 comps describes a market that no longer exists, and a reserve set against it produces a pulled process. So check your number against the shape for your submarket and property type: where sellers there have actually exited, how wide the spread runs, and how large the loss tail is. Guidance is one point in that range.
Finding the owners who will sell
The two figures are the market-level halves of the pair. The transactability score brings the first question down to a single owner: will this one sell inside my window, and how much stands between you and a closed deal. In the product it arrives as a ranked queue, and the queue is how you find deals. You set the criteria your mandate already reads in, submarket, property type, size, vintage, and every owner that matches comes back ranked by likelihood to sell inside your window, with the four components behind the score, so your desk calls the right owners before a process exists.
Each component comes from a different evidence class. Asset friction: environmental exposure, liens, deferred capex, zoning, researched from public records. The owner's clock: a hold-period model trained on 600,000 repeat holds across the global transaction record; owners past their expected hold sell at a higher rate than those inside it. Market hotness: per-metro, per-type absorption momentum, which is Figure 1, localized. And owner selling probability: a classifier over transaction history and the owner's own stock performance, which applies where the owner is a REIT.
What your desk gets, concretely: the owners in your criteria, ordered by how likely each is to transact inside the window, with the evidence behind every score laid out component by component. The ranking moves as the record does: an owner who crosses their expected hold climbs, and so does a submarket whose volume turns. The test of a score like this is outcomes: whether the owners it ranks highest go on to sell inside the window more often than the rest of the queue. That is the same test we apply to the REIT model, and it is the number to ask for on the call.
The deals that are not on the market yet
In a thin market a large share of the year's deals are found. The marketed pipe carries the assets whose owners have already decided; everything else is behind a door, with an owner who is past the hold they planned, or in a fund near the end of its life, or in a submarket where volume has just turned. Reaching that owner before the announcement is the oldest arithmetic in the business, and the queue does it at the scale of a mandate: every owner in your criteria, ranked by likelihood to sell, with the evidence, so the first call goes to the owner most likely to take it. That is deal flow your desk would otherwise never see.
The same ranking gives you a byproduct: the confident pass. The score says whether an owner is likely to sell, and nothing about how, so the pass it supports is narrow: the owners the record says are unlikely to sell this quarter are owners your desk can leave alone, decided before the first call. That frees the week for the finds.
Once a deal is in hand, from either door, the run is the same: minutes in your templates, every number cited, checked against the clearing record and your own deal history. The queue and the screen both read against that history and your passes, so they reflect your buy box as your desk actually runs it.
Bring us a submarket
Every figure in this piece is built on licensed MSCI Real Capital Analytics data, so we show it on a call. Bring us a submarket and a property type. On that call we'll pull up its volume series, which is its transaction probability over time; its clearing distribution, where sellers there have actually exited and how the loss tail has moved; and the owners in it who are past their expected hold, ranked. Then the owners in it who are likely to sell and are not on the market. Bring a live OM as well, and we'll run it while you watch.
Source: MSCI/RCA transactions via Snowflake, cached extract, through 2026-07-24. 694,279 US closed sales; 162,569 usable single-asset repeat sales (≥95% interest conveyed, holds of 1–30 years, annualized change within ±30%). Figures are aggregates of licensed MSCI Real Capital Analytics data; prepared by Starboard Research. Header photograph: the Chicago Loop across Lake Michigan from the Museum Campus, by Jeff Brown on Unsplash.