Creative brief library

skills/creative-brief-library/SKILL.mdView on GitHub

Goal: a set of angles the company can keep producing ads from, each written up once as a brief that any writer, designer or tool can execute, and a routine that tests a few at a time, reads the numbers, keeps the winners and drops the rest. The library grows every month; the ads stop being made from scratch.

Before you start

This skill works on its own. If the person has their own marketing brain folder, read the positioning, messaging, ideal customer, brand voice and any ads notes. If they do not have one, ask: who the buyer is in a sentence, the problem in the buyer’s words, the two or three reasons the company wins, the offer and the first step, and what proof can be used publicly. Ask for the ad account results so far if any, and for any ads, posts or emails that got a response.

Tools this skill uses when the person has them. None are required to write the briefs.

Steps

1. Write the angle list. An angle is one reason a buyer in the segment would stop and read. Draw from six sources and aim for eight to twelve angles:

   • The problem in their words (one angle per major pain).
   • The cost of the workaround (time, money, risk, in numbers).
   • The outcome after (a before-and-after a customer actually had).
   • The objection answered (the top two objections, turned into the headline).
   • The comparison (versus the tool, vendor or habit it replaces; versus doing nothing).
   • The trigger (the moment that makes them look: a new hire, a lost bid, a deadline, a rule change). Rank them by how specific they are to the segment and how much proof backs them.

2. Write one brief per angle, in the same template every time:

   • Angle name and the one-line idea.
   • Who it is for (segment and role) and where they are in the decision (never heard of the problem, know the problem, comparing options, ready).
   • The promise: one sentence, in the buyer’s words.
   • Proof available: the number, quote or logo that backs it, with approval status.
   • Three headline options, each under eight words.
   • Primary text or description, 40 to 90 words, one idea, one call to action.
   • Visual direction: what the image or video shows (the product, the before and after, a real person, a number); what it must not show (stock handshakes, generic AI imagery, text-heavy slides).
   • Formats: which of static, carousel, short video, document ad, search text ad fit this angle.
   • Landing page: the existing page it goes to, or the note that it needs its own.
   • The objection it will trigger and the answer, for the page or the follow-up.

3. Map angles to platforms and stages. Search ads take the problem and comparison angles with high intent keywords. LinkedIn takes the problem, outcome and trigger angles for the role, with document and video formats. Meta takes outcome and problem angles with video and carousel for retargeting and for segments that are on it. Do not run every angle everywhere.

4. Set the first test. Three angles, two formats each, one platform, two to four weeks, a budget the company can lose without worrying. Define success before launch: cost per qualified lead against the target, not cost per click. Name the one variable being tested in each pair.

5. Produce the creative from the briefs. Hand the briefs to whoever makes the ads, or make them: copy from the brief as written, visuals per the direction, the brand kit applied. Two versions per brief at most for the first test. Every ad gets a UTM that names the angle.

6. Read the test. At the end of the window: leads and qualified leads by angle (from the CRM, not the ad platform), cost per qualified lead, and what the sales team heard from those leads. An angle that gets clicks and no qualified leads is a losing angle with a good headline.

7. Keep, cut, scale. Keep the angle with the best cost per qualified lead and give it more budget and two new executions. Cut the worst. Replace it with the next angle from the library. Record the result on the brief: what was tested, when, the number, the decision. The library is also the memory.

8. Repeat monthly. Each month: one new angle in, one out, two new executions of the winner, one note per brief. Audit the account for spend on search terms or audiences that never produce qualified leads and turn it off.

Rules

• Angles come from customers, not from brainstorming. Every angle names its source.
• One idea per ad. One promise per landing page.
• Qualified leads are the measure. Clicks and impressions do not pay anyone.
• Test one variable at a time, three angles at a time, with a window and a number decided before launch.
• Real proof only, approved before it runs. No invented numbers, no unapproved logos.
• The brief is the asset. Ads expire; the brief and its results stay in the library.
• A person approves every ad before it runs and reads the results before anything scales.

Output

Return the ranked angle list with sources, the brief for each angle in the template, the platform and stage map, the first test plan with success numbers, the UTM naming for angles, the read-the-test template, and the monthly routine.

If the person has their own marketing brain folder, save to it:

• 4-playbooks/ as ads.md: the angle list, the brief template, the test routine and the results log
• 3-strategy/messaging.md: the headlines that won, dated, appended
• 5-current/priorities.md: the current test and its end date
• 5-current/decisions-log.md: what was kept, cut and scaled, with dates

If they do not have one, return the library as one file the team keeps with the ad account. Do not set up a brain unless they ask.

Next: a platform-specific ads skill, if the person has one, runs the account; this library feeds it. homepage-rewrite supplies landing pages. call-mining refreshes the angles each quarter.