Nobody Keeps a Copy of the Market Data They Pay For

You pay for a market-intelligence subscription every month, you set rates and underwrite deals off what it tells you, and you keep almost none of it. That is the part of the Green Street and StorTrack story worth an operator's attention, and it has nothing to do with AI.

Here is what happened. Green Street acquired StorTrack on July 15, announced out of Newport Beach, terms not disclosed. StorTrack was founded in 2014, sits in Southfield, Michigan, and is the benchmark for unit-level pricing, supply, and development data in this industry. The deal also pulled in ListSelfStorage.com and RVParkIQ.com. Then on August 11, Green Street made its Model Context Protocol server generally available, piping its research, market data, market forecasts, sales comps, and single-property valuation model directly into corporate instances of Claude, ChatGPT, and Gemini. Green Street's own framing is that a market underwrite that used to take days now takes a conversation.

StorTrack's unit-level rate data is not on the dataset list in that August launch. I would not build a plan around it staying off.

Everyone buys the same feed. Not everyone buys the same question.

The instinct is to read this as a data-access story, and it isn't. Anybody can buy StorTrack. Optimize is priced for operators and has been for a decade. The private equity group underwriting a portfolio two counties over is looking at the same scraped rates you are.

What changed is the cost of asking a complicated question of that data. Their analyst types it in plain language and gets a set of markets ranked against comps and a valuation model in one pass. Your version of the same question is an export, a spreadsheet, and a call to a regional manager. Same facts on both sides of the table, very different number of hours between the question and the answer, which in a competitive bid is the only part that matters.

That gap does not close by buying more data, and it does not close by buying the same MCP connector. It closes when the market feed stops being the only thing in the room. A ranked market list is an outsider's view of your own submarket. You have something the outsider does not, which is what actually happened at your counters, and the feed is only worth what your own record can do to it.

Published rates are not rates

Here is the part I would be more worried about than the speed gap.

A market feed scrapes what facilities publish. Take a 10x10 non-climate in a submarket where three competitors advertise between $129 and $139 and you sit at $149. The feed says you are priced above market, and it is correct about the advertised number. What the feed does not know is that two of those three are running six weeks free on new move-ins, that your in-place average on that same unit size is $171 across tenants past their first year, and that your annualized increases stick at a rate you can measure and they can't. Published rate is an acquisition signal. In-place rate is the business.

Feed the published number into an AI pricing loop with nothing else attached and you have not automated a good decision. You have automated a bad comparison and removed the friction that used to catch it. That is not an argument against pricing tools. It is an argument that a market feed is one input to a rate decision and never the decision, and that the thing which makes it mean something is your own in-place history sitting beside it.

Keep the receipts

The unglamorous consequence of an acquisition like this is that terms change. Not maliciously. Coverage gets rationalized, methodology gets harmonized with the parent platform, packaging gets rebuilt, pricing tiers move. StorTrack's founder said clients keep the same tools and support they count on today, and I expect that is sincere and true for a while. "For a while" is the operative part. Vendors get integrated after they get bought, because that is the point of buying them.

So a question worth asking yourself before the next renewal. If the feed changed shape tomorrow, could you still say what happened to advertised rates in your submarkets over the last eighteen months?

For most operators the answer is no, because they query the feed live and archive nothing. The subscription gets treated as a window rather than a source. Institutions do the opposite. They pull it, store it, and build a history they own, which is why they can talk in trend lines while you are stuck talking about last week.

Fixing that is dull and cheap. Pull the market data you are already paying for on a schedule, land it in your own store, timestamp it, keep it. Most feeds worth having sell an API for exactly this, StorTrack included. It buys two things. A rate history nobody can revoke, and the ability to check whether the feed's version of your market has ever matched what actually happened at your counter. The second is worth more than it sounds. Nothing calibrates trust in a data source like eighteen months of watching it be wrong about your own submarket in a consistent direction.

The trade-off, since there always is one. You now own a data pipeline, which means somebody has to keep it running and somebody has to notice when it quietly stops. That is a real cost. If you run a small cluster in one metro and you know every competitor by name, I don't recommend it - you already hold the history in your head and the pipeline is overhead. Past the point where you stop knowing the market by driving it, the archive is worth more than the subscription.

The vendors will keep consolidating and keep bolting AI front doors onto data you already buy. None of that changes what is worth doing this quarter, which is getting your own record into a shape that can sit next to a market feed and argue with it.