THE LANGUAGE OF DATA SERIES | PART 3 OF 7
The Grey Market Was Built In
Full payloads plus honor system rights created a shadow economy in listing data. Customized payloads and access based delivery are how it ends.
By Saul Klein and John Reilly September 2026
For decades, MLS data feeds worked on a simple and, in hindsight, remarkable arrangement: send the recipient everything, and tell them on paper what they were allowed to use. This article, which stands alone, is about what that arrangement built, and about the redesign now replacing it.
The payload problem: when the pipe carried everything. In the era when feeds meant copying, the feed typically delivered the whole payload: every field, every record, every photo the recipient’s category could conceivably need, all at once. The sorting happened afterward, on paper. The license told the data consumer which parts of the payload it could use, display, and keep, and which parts it could not, and then everyone hoped. Send everything; trust the recipient to touch only what the grant covers. That is a system based on trust, and it worked at scale about as well as trust systems ever do.
The grey market. The trust based payload produced a predictable child, and the industry has a name for it. The grey market is not the black market of outright theft. It is the sprawl of MLS data circulating in ways nobody quite licensed and nobody can quite trace: the vendor feed that keeps running after the client relationship ends; the full payload delivered for one purpose and quietly repurposed for another; copies passed downstream to sub vendors and partners the license never named; over retained historical data reborn as someone’s analytics product; exported records laundered into tools and lists far from their source. No single act looks like stealing, and no single actor can be caught selling, because every recipient of a full payload became, in effect, a potential wholesaler the moment the copy landed on its servers. The grey market is what full payloads plus honor system rights inevitably build, and a meaningful share of the data products in this industry were built on it.
Customized payloads: the license enforced at delivery. The corrective is as simple to state as it was impossible to implement in the file transfer era: make the payload match the grant. Deliver only the fields, records, and media the license actually covers. The website display licensee receives the display fields and nothing riding along just in case. The valuation licensee receives what a valuation requires. The research licensee receives the defined project dataset and not one record more. When the payload is the license, violating your rights requires affirmative effort instead of passive drift, and the grey market loses its raw material at the source. Modern query based delivery makes this practical for the first time, because a pipe built on questions can shape every answer to the credential asking.
And where copying survives, price it as what it is. There remain legitimate cases for a recipient to hold a full copy; a portal serving millions of visitors cannot query the MLS for every page view. But a copy should never be priced like access, because it is not the same product. Access grants use while the data stays home under the owner’s enforcement. A copy surrenders possession, and with possession goes most of the owner’s practical control: data on someone else’s servers is governed by promises, not credentials. The party that takes a copy takes on what the owner gives up, and the price should say so. Access at one price. Replication, where permitted at all, at a premium reflecting the surrendered control. An MLS that charges the same for both is subsidizing the riskier product with the safer one.
Put the pieces together and the direction is unmistakable. Customized payloads shaped to each license. Access as the default mode of delivery. Replication as the priced exception rather than the free assumption. License defining payload, payload enforcing license. The grey market was a design decision, made unknowingly, one full payload at a time. The redesign now underway is the industry finally making the opposite decision on purpose.
This article stands on its own and is also part of the practitioner series “The Language of Data”. Copyright 2026 Saul Klein and John Reilly. All rights reserved. Commentary and general information, not legal advice.
