Ask most people where a policy is decided and they will point to a room. The committee hearing, the cabinet meeting, the floor vote, the moment a minister signs. There’s a high probability that is where you would point too. It is a reasonable answer, and most of the time it is the wrong one.
The vote is where a decision is recorded. It is rarely where the decision is made.
Consider the arithmetic of Washington. In the 118th Congress (2023 and 2024), members of the House and Senate introduced more than 16,200 bills, according to the official résumés of congressional activity. Fewer than one in fifty became law. Put another way, for every bill that reached the President’s desk and was signed, more than fifty others were introduced and quietly died. Most of them never came anywhere near a vote. Nobody voted them down. They simply never became the thing people were talking about.
So what decided their fate, if not the vote?
Money matters less than it seems
The instinctive answer is resources: whoever spends the most wins. The best evidence we have says otherwise. Frank Baumgartner, Jeffrey Berry, Marie Hojnacki, David Kimball and Beth Leech spent years following 98 policy issues through Washington for their book Lobbying and Policy Change. They found that about 60 percent of lobbying campaigns failed to change policy at all, despite millions of dollars spent trying. Resources explained less than five percent of the difference between the campaigns that succeeded and the ones that did not. What they found instead was a system with a heavy bias towards the status quo.
That finding should unsettle anyone who believes influence can be bought late. If money barely moves the outcome, and the default is that nothing changes, then the side that wins is the side that has done the slower, less visible work of making change look sensible, safe and already half agreed before anyone is asked to commit.
Which raises the obvious question: when does that work happen?
The window is short
John Kingdon, whose Agendas, Alternatives, and Public Policies remains the most useful book ever written on how issues reach a government’s desk, described policy ideas floating in a "primeval soup" of specialists, staff and advocates, bumping into each other and being reshaped over years. A decision becomes possible only when three streams line up: a problem people recognise, a workable solution, and the political will to act. When they do, a window opens. Windows do not stay open long.
Most people read Kingdon as a theory about luck. We read it as a warning about timing. By the time the window opens, your case must already be finished, tested and familiar to the people who will decide. The advocates who win when a window opens are almost always the ones whose idea was already in circulation, already costed, already in the vocabulary of the officials drafting the options paper.
If you are writing your first briefing note on the day the window opens, you are not early. You are a visitor.
How the question is framed
In 1981, Amos Tversky and Daniel Kahneman published a short paper in Science that should be required reading for anyone who works near a government. They gave 152 students a choice between two programmes to fight a disease expected to kill 600 people. Described in terms of lives saved, 72 percent chose the certain option. They gave 155 other students exactly the same choice, with the same numbers, described in terms of lives lost. Now only 22 percent chose it.
Same facts. Same options. Different words, and the majority flipped.
Now think about how a policy question actually arrives in front of a decision-maker: as a one-page summary written by someone. Is the measure a subsidy or an investment? A restriction or a protection? A cost to the treasury or a saving on future spending? Whoever writes that first page has already decided a good part of the outcome, and they usually write it weeks before the meeting.
If you are arguing about the answer, you have probably already lost the argument about the question.
Consultations
Many institutions wait for the formal invitation to speak. In the United States, that invitation is the public comment period. Executive Order 12866 established 60 days as the standard length. But by the time a proposed rule is published for comment, the agency has already defined the problem, chosen its approach and drafted the text. The comment period is real and it matters. It is also, by design, a response to a decision that has largely been framed.
And volume is not persuasion. When the Federal Communications Commission consulted on net neutrality in 2017, it received more than 22 million comments. The New York Attorney General later found that nearly 18 million of them were fake. Millions of voices, and almost none of them were people.
India offers a different lesson. The Pre-Legislative Consultation Policy of 2014 asks ministries to publish draft bills for at least 30 days before they go to Parliament. Researchers found that more than 76 percent of bills introduced between June 2014 and May 2019 had no public consultation at all, and that 17 of the 29 bills listed for the 2021 Winter Session had none either. If your plan in Delhi depends on the formal window, you may find there is no window.
In both capitals, the official moment to speak arrives after the most important choices are already on the page.
What the early work looks like
We start with a question that sounds simple and rarely is: what does the person deciding need in order to say yes? Not what you want them to believe, but what they will need to tell their own minister, their own party, their own constituents. A decision-maker who agrees with you but cannot explain the decision to their people will not make it.
Then comes evidence, produced before anyone asks for it. A credible number, carefully sourced, in a form an official can lift straight into a brief. Officials under time pressure use the evidence that is in front of them, and the evidence in front of them is the evidence someone bothered to put there early.
Then the quiet conversations. Not pressure, which tends to harden positions, but testing. Which frame lands? Which objection keeps coming up? Who else needs to be comfortable? Every objection you hear privately in week three is one you can answer before it is raised publicly in week ten.
And finally, voices that are not yours. An argument made only by the people who benefit from it is an argument with a discount attached. The same argument made by a farmer, a mayor, a university economist or a former official carries its full weight.
None of this is glamorous. None of it makes the news. That is the point.
The best argument in the room is the one the decision-maker already believes is their own.
When to start
Earlier than feels necessary, and long before anything is formally on the table. The work of public affairs is mostly the work of making a decision feel obvious by the time it is taken: the problem already named in your terms, the evidence already in the file, the objections already answered, the allies already standing where they can be seen.
By the time the vote is called, you should already know how it ends.
