The Data Center Vote Was Unanimous. The Audience for It Isn't.
Somewhere in a congressional district right now there is a constituent who has been to four county hearings about a proposed data center. She knows the acreage. She knows which substation it would tie into. She has a folder.
On September 16, the House passed the Ratepayer Protection Act by a vote of 417 to 3. The bill directs state regulators and unregulated utilities to consider whether the largest electricity customers — data centers drawing more than 100 megawatts — should cover the full incremental cost of the generation, transmission, and distribution upgrades built to serve them. The next day, an effort to move it through the Senate by unanimous consent was blocked. It has not moved since.
Her office is going to email her about that vote. It will run three sentences, it will go to the entire list, and it will tell her less than she knew in June.
That is not a writing problem.
A 417-3 vote is a fact about Congress, not a fact about your district
Three members voted no. That is as close to consensus as the House gets, which makes it easy to treat the vote as universal good news, and universal good news as newsletter material.
The country agrees with the bill. The country is not your list.
Pew released a survey on September 22, fielded July 20 through August 9 among 10,548 adults, finding that 60 percent of Americans would be not too or not at all comfortable with a new data center in their community. The level is striking. The movement is the story. In January, 30 percent said data centers were mostly bad for the quality of life in nearby areas. By August it was 49 percent. Among rural Americans, the share saying data centers harm the environment went from 32 percent to 50 percent in seven months.
Now put that 60 percent inside one district. Part of it is the woman with the folder. Most of it is people holding an opinion the way anyone holds an opinion about something happening two counties over.
Both groups are in the 60 percent. One of them is an audience.
The bill tells you who to write to, if you read it that way
Read the legislation as a targeting spec rather than a policy summary. It names a customer class, a load threshold, a regulator, and a rate base.
Every one of those is a geography.
A 100-megawatt threshold points at specific parcels. State regulators point at specific commissions and specific dockets. Cost recovery for grid upgrades points at specific utility service territories — and a service territory is a line on a map that can be matched against an address file.
This is not an issue where interest is spread thin and has to be inferred from attitudes. Interest here is concentrated, physical, and in most cases already sitting in a public proceeding with a docket number attached to it.
Which makes it a test. An office that can find the four thousand households with a stake will sound like it knows what is happening at home. An office that cannot will send three sentences to seven hundred thousand people and file it as outreach.
One email, three failures
The first failure hits the woman with the folder. She opens a message announcing that a bill passed and reads an office that does not appear to know what is going on in her own county. Nothing in it is wrong. It is pitched at someone who has not been paying attention and delivered to the person paying the most.
The second failure is everyone else, and nobody counts it. For a constituent forty miles away on a different utility, this is a note about a bill she had not heard of, concerning a facility she will never see, on a subject that does not touch her bill. Not offensive. Just noise — and noise gets charged to the next send, not this one.
The third failure is the expensive one. The office learns nothing.
An undifferentiated send produces an undifferentiated response. Nobody can look at the report afterward and say whether the people who opened it were the people who care. The open rate goes into the weekly summary, and the most valuable thing the issue could have surfaced — the name of every constituent in the district who is actually engaged on energy costs and siting — never gets written down.
Party is the weakest variable in the room
The instinct with an issue this loud is to reach for the political variable. On this one it explains close to nothing.
A bill that draws 417 votes does not divide an audience along familiar lines. The polling says the same thing more bluntly: discomfort with a nearby data center runs at 54 percent among Republicans and 68 percent among Democrats. A real gap. Nowhere near enough to sort a district on.
Three miles matters. Party doesn't.
A constituent near a queued project has a stake regardless of anything else in her record. A constituent forty miles away on another system does not acquire one by holding strong views.
Start with geography, then with the people who already wrote in
The objection at this point is that no office keeps a field called "lives near a proposed data center." True, and a smaller obstacle than it sounds.
Address geography belongs in every constituent record. If a program cannot resolve its file down to a county or a ZIP code, that gap is worth closing on its own terms. It predates this bill and it will outlast it, because nearly every segment an office will ever want to build runs through geography first. An office without it is not only unprepared for this issue. It is unprepared for the next one.
Then there is the signal most programs never convert into a segment: prior constituent-initiated contact. The people who wrote in about utility bills. The casework involving a utility. The correspondence about a zoning matter. Those constituents raised their hand before there was a vote to announce, and pulling them into a list requires appending nothing at all.
The public record does the rest, and on this issue it is unusually good. Interconnection queue requests are published and tracked. Independent project trackers cover proposed and under-construction facilities at the county level. Local opposition is organized and visible — one tracking effort counted 142 active groups across 24 states. A staffer with an afternoon and a browser can find out what is proposed in the district.
Most of the work here is research, not data acquisition.
We have built this segment before under a different name
On one Senate program running more than fifteen audience segments, agriculture content went to farmers and ranchers, and when the Senator secured school funding, parents of school-aged children got a dedicated email instead of a buried bullet in a general newsletter. The mechanism never changes: find the constituents the message is actually about, then write to them instead of around them.
Data centers happen to be an unusually clean case. The geography is concrete, the record is public, and the people who care have already identified themselves at four county hearings. Most issues are harder than this one.
Which is the argument for practicing on it.
The office that can find those four thousand constituents is running a constituent communications program. The office that sends to all seven hundred thousand has a mailing list. If you are not certain which one you have, that is worth twenty or thirty minutes of conversation.