A good briefing describes a problem, not a solution. That sounds obvious and it is the mistake in almost every first draft: it already says what has to be made, so you are buying execution while paying for thinking.
The six questions a briefing answers
1. What is going on? The situation, with numbers. Not "we want to grow" but "our awareness among thirty-somethings fell from 34 to 21 percent in two years".
2. What has to change? One goal, measurable, with a deadline. Several goals means the agency has to pick one, and not necessarily the one you would have picked.
3. For whom? The audience described as people, not as a segment. What do they do today, what holds them back, and what do they already know about you?
4. Why would they choose you? Your positioning, and honestly: if you cannot write it yourself, that is your real assignment.
5. What is fixed and what is free? Brand rules, legal constraints, channels you already have, things that failed last year. This chapter saves everyone a round.
6. What is the budget and when is it due? Without a budget you get proposals you cannot compare and usually cannot afford.
Two things that improve a briefing more than anything else
Say how you will know it worked. One sentence, up front, with a number in it. It forces you to be precise and gives the agency something to hold its proposal against.
Say what explicitly does not have to be made. A briefing without limits produces proposals with five channels in them when you wanted two.
The practical test
Have someone who did not help write it read the briefing and summarise the assignment in one sentence. If that sentence is wrong, the agencies will not get what you meant either.
IKAg publishes a free briefing analysis at www.ikag.be/briefing that checks a briefing on these points.
See also: what belongs in a pitch briefing and how much work you may ask for.