It Came Back Fine. That's the Problem.
The file lands in your inbox. You open it. The logo is the right size, the colors are yours, the photo of the van is the good one. Nothing is wrong with it. You look at it for a few seconds, feel a small flat nothing, and type "looks great, let's run it," because you cannot put your finger on what would make it better.
Two weeks later it has spent six hundred dollars and produced two leads, one of which was a wrong number.
Then comes the post-mortem, and it is always about the ad. The headline was weak. The color should have been brighter. Maybe it should have been video. Maybe the designer just isn't very good. Sometimes that last one is even true. But most of the time the ad was decided long before anyone opened a design file, in a two-line email that asked somebody to make something without ever saying what it was supposed to do.
That email is your brief. You wrote one whether you meant to or not.
"Make Me an Ad" Is a Wish, Not a Brief
Here is what actually gets sent, most of the time. A logo file. A folder of photos pulled off the website. A link to a competitor's ad with the words "something like this." If the client is feeling generous, one more line: make it pop.
Now look at what that email is really asking for. It does not say who the ad is for. It does not say what the person should walk away thinking. It does not say what stops that person from buying, or what proof exists that they should. It does not say where the ad will appear or what happens after the click.
Those are five business decisions, and the email has just handed all five to somebody who has never answered your phone, never lost a quote to the cheaper guy down the road, and never read your reviews.
They will make those decisions anyway. That is the part owners miss. Silence in a brief is not neutral, and it does not get flagged back to you. It gets filled in, quietly, by whoever is doing the work, using the only information they have, which is your website and their own instincts about what an ad looks like.
There is a harder truth underneath this, and it applies to the best creative people in the business as much as the cheapest. Nobody knows which ad will win. Anyone producing ad creative is making an educated guess and then letting the market grade it. A brief does not remove the guess. It narrows it. Without one, the guess is "what should we say to somebody, somewhere, about this business." With one, the guess is "what is the strongest way to say this one thing to this one person who is worried about that one thing." The second guess is wrong far less often, and it is wrong in ways you can learn from.
You Are the Creative Strategist Whether You Want the Job or Not
The advertising industry splits creative work into two jobs that sound identical and are not.
One person decides what should be made and why. Which customer, which message, which angle, which hook is worth trying next, and what the results say to change. In performance marketing that role is usually called a creative strategist, and it barely existed fifteen years ago.
The other person makes sure the thing that gets made is actually good. The craft, the taste, the look, the standard. That is the creative director's job, and it has been around for as long as advertising has had teams.
Your business has neither of these people on payroll. When you hire a freelancer, an editor, a design service, or a small agency to produce creative, you are almost always buying the second job. You are buying execution. The first job does not disappear because nobody was hired to do it. It lands on you, by default, usually without anyone saying so out loud.
That sounds like a burden. It is closer to leverage. Deciding who the ad is for and what it should say is the cheapest task in the entire production, it takes about twenty minutes, and you can do it better than anyone you could possibly hire, because it runs on information they do not have.
Half of a Good Brief Is Information Only You Have
This is the part that cannot be outsourced at any price, and it is worth being specific about what it consists of.
The objection you hear on nearly every job. You know it without looking it up. It is price, or timing, or "I need to talk to my wife," or "we had someone out last year and they never came back." Whatever it is, you have heard it a few hundred times and your ad has probably never once addressed it.
The words your customers actually use. People do not say "premium climate control solutions." They say the upstairs bedroom has been ninety degrees since June and nobody will call them back. That sentence is worth more in an ad than anything you could compose, because it is already the language running through the next customer's head.
The real reason people pick you. Not the reason on your homepage. The actual one, which is usually smaller and more specific and slightly embarrassing in how ordinary it sounds. You answer the phone. You show up when you say you will. You are the only one in the county who works on that brand of boiler.
What a customer is worth to you. This decides how hard the ad has to work and how much you can afford to pay for the click. A brief that does not know this cannot be graded.
If you do not have these on the tip of your tongue, they are sitting in your business already, and an hour of reading will get you all four. Go through your Google reviews, and pay attention to the three-star ones, because those are the people who nearly did not buy and will tell you exactly why. Read the last twenty emails from customers. Look at the quotes you lost and be honest about the pattern. Read the comments on your own ads, and on competitors' ads, which are often more useful. If your category has an active corner of Reddit, read that too, because anonymity makes people blunt in a way review platforms never do.
An hour of that will hand you more usable material than a week of sitting in a room trying to think of ideas.
The Seven Lines
Here is the whole document. It fits on one page, and if it does not fit on one page it is not a brief, it is a novel nobody will read.
1. Who exactly. One person, not a demographic. "Homeowners 35 to 65" is not a person. "A homeowner who just found a damp patch on the bedroom ceiling and is scared the whole roof is going" is a person, and you can talk to them.
2. Where it runs, and how it will actually be seen. An Instagram feed on a phone with the sound off is a completely different ad from a YouTube pre-roll on a television, which is different again from a display banner sitting beside an article. Naming the placement is not a technicality. It decides length, whether words have to carry the message, and how fast the point has to arrive.
3. The one thing. If they remember a single sentence from this ad, what is it. If you cannot get it down to one, the ad will try to say four things and land none of them.
4. The objection it has to beat. Every ad is arguing with something, even when it pretends not to be. Name what it is arguing with so the argument can be won on purpose.
5. The proof you actually have. A specific review, a number, a photo of real work, a guarantee, a license, the years you have been doing this. "We are the best in the area" is not proof, it is a claim, and everyone makes it.
6. The action, and what happens after it. What should the person do, and what will they land on when they do it. An ad that promises a free quote and drops people on a homepage has wasted the money it took to earn the click.
7. What must not change. Claims you are not allowed to make, regulated language, the logo rules, the thing your industry body cares about. Say it at the start. Discovering it in round three costs everyone a week.
That is it. Seven lines, written once, reused with edits for every ad after this one.
Brief to a Number, Not a Vibe
"Make it feel premium" cannot be graded. Neither of you can ever be proven right about it, which means the review turns into a conversation about taste, and taste arguments are unwinnable and expensive.
Give the ad a job instead, and attach that job to something observable. This has to stop somebody scrolling in the first three seconds. This has to earn a click from a person who has never heard of us. This one is for people who already visited the site and did not call, so it can assume they know who we are.
The difference shows up when the ad underperforms. With a vibe, all you have is "I don't think it worked." With a job, you get a diagnosis. Plenty of people watched and nobody clicked, so the ad is interesting but the offer or the ask is weak. Almost nobody made it past the first second, so the opening frame is the entire problem and nothing after it matters yet. That is a note the next version can actually act on.
One caution. Do not import somebody else's benchmark percentages and treat them as law. The numbers agencies publish come from their own accounts, usually in categories nothing like yours, and a target lifted from a supplement brand will mislead a roofer. What matters is not the industry figure. It is that the brief named the job, so you can compare this ad against your own last one and know which direction you are moving.
The Four Sentences That Ruin a Brief
"Make it pop." This means "I will know it when I see it," which guarantees three rounds of revisions and a final ad that is whatever survived the arguing.
"Something like this competitor." You are copying the visible surface of a strategy you cannot see, for a business whose margins, customers and offer you do not know. You also have no idea whether that ad worked. It might be the one they are about to turn off.
"Can we get everything in?" Every message you add divides the attention available to the one before it. An ad with five points communicates none of them, which is a different problem from an ad with one weak point, and a much harder one to diagnose. We have written about this specifically for banner ads, where the temptation is strongest.
"Let's see what the team thinks." Group approval optimizes for the ad nobody objects to. The ad nobody objects to is also the ad nobody notices. Somebody has to own the decision, and in your business that somebody is you.
Consider what actually distinguishes ads that hold real spend. Naming the scent notes on a soap ad, because a screen cannot carry a smell and the buyer needs a way to imagine it. Putting a genuine customer review at the very top of the image instead of the bottom. Choosing one loud block of color because color is what interrupts a scroll, not copy. None of those are flourishes a designer thought of at the last minute. Every one is a decision that could have been written on a brief before the file was ever opened.
Brief the Idea, Not the Layout
An owner who writes "make me a carousel" has already skipped the only decision that mattered. Format is downstream of everything else. You choose it once you know who you are talking to, what you need to say, and where they will see it, and sometimes the honest answer is that the format you had in mind is the wrong container for the message. A carousel, for instance, is one idea told in sequence, not five separate ads stapled together, which is a mistake worth its own article.
There is a newer reason this matters more than it used to. Anyone can generate a competent ad now, in about a minute, for almost nothing. But generation models work by producing the most likely next thing, which means what comes back is an average of everything that already exists in that category. Average is precisely what your customer has trained themselves to scroll past. Cheap production has not made good ads more common. It has made mediocre ads infinitely available, which raises the value of the one input a machine cannot supply.
That input is the idea, and the idea comes out of the brief. The specific, slightly odd, true thing about your business that nobody else could claim. We have made the longer version of this argument elsewhere, and it has only got more true since.
One Brief Should Produce Three Different Ads
Here is a test worth running once. Give your brief to three different people and look at what comes back.
If you get three near-identical ads, you over-specified. You did not write a brief, you wrote a spec, and you have paid three people to be your hands. If you get three ads pointed at different customers saying different things, the brief was too loose and everyone filled the gaps themselves.
What you want is three ads with the same person in mind, the same single message, the same objection being argued with, and three genuinely different ways in. One leads with the review. One leads with the problem. One leads with the price nobody else will put in writing.
This matters for a mechanical reason as well as a creative one. Meta now reads the creative itself to decide who to show it to, and it groups near-identical variations together and treats them as effectively one ad. Five recolors of the same layout do not buy you five chances. They buy you one, and they waste the budget that funded the other four. A small set of genuinely different concepts beats a large set of variations, every time, which is why hunting for the one perfect ad is the wrong goal.
So write the brief tight where strategy lives and loose where craft lives. Lock the person, the one thing, the objection and the proof. Leave the format, the tone and the structure open, and let the people you are paying earn their fee in the space you left them.
How to Give Feedback Without Wrecking the Work
Most working relationships between an owner and a creative person do not break at the brief. They break in round two.
Here is feedback that destroys good work: "I don't like the blue." "My wife thought it was confusing." "Can we try something a bit different?" None of those can be acted on. They are reactions, and the person receiving them has no choice but to guess again, this time with less budget and less patience.
Here is feedback that improves it, and the difference is that every note points back at a line in the brief. This does not beat the price objection we agreed it had to beat. The one thing was next-day service, and I do not see it until three seconds in. This is written for a homeowner but it reads like it is talking to a builder. The proof we agreed to lead with is at the bottom where nobody will reach it.
Notice that none of those mention taste, and all of them are solvable.
Two rules make this work. Give all your feedback in one pass rather than dripping it out over three days, because a trickle of small changes will quietly dismantle whatever coherence the ad had. And separate "this is wrong" from "this is not how I would have done it." Both are allowed. They just need different labels, so the person doing the work knows which notes are load-bearing and which are preferences they are welcome to push back on.
If you cannot tie a note to something in the brief, that is useful information too. It usually means the brief was missing a line, and now you know which one to add before the next round.
The Cheapest Twenty Minutes in the Whole Production
Everything downstream of the brief costs real money. The shoot costs money. The editor costs money. The revisions cost money and time, and the media budget behind a bad ad costs the most of all, because it buys you a confident, expensive answer to a question nobody bothered to ask properly.
The brief costs twenty minutes and a piece of paper, and it is the only part of the process that gets cheaper every time you do it, because next month you edit last month's rather than starting over.
We write these for clients every week, and the pattern is boringly consistent: the businesses whose ads work are not the ones with the biggest production budgets. They are the ones who can tell you, in a sentence, who the ad is for and what it is supposed to do. If you would rather hand the whole thing over and get back to running the business, that is what we do. And if you would rather write your own briefs from here, genuinely, good. That is the part worth owning.
