How to Build a Messaging Framework That Ships Ads

You've got the brief. The product team wants one thing, the paid social lead wants another, and the creative queue is waiting on hooks that should've been written yesterday. In a lot of teams, the problem isn't a lack of ideas, it's that the message was never pinned down before the first script got drafted. That's where a messaging framework earns its keep, because it gives the team a shared message system before anyone starts polishing a concept into dead-on-arrival creative.

Why Most DTC Ads Sound the Same

The Monday morning ritual is familiar. A fresh brief arrives, three people weigh in on the hook, and by noon the ad is somehow about everything and nothing at once. That's how teams end up with scripts that are technically fine and strategically flat, because the message never got narrowed to one point of view before the edit started.

A lot of brands treat messaging like a brand-book exercise, something to tidy up once the positioning deck exists. Paid social doesn't care about tidy. It cares whether the opening claim matches the audience's problem fast enough to earn a second look, and whether the proof is strong enough to justify the ask.

The difference shows up in the archive. A team that keeps one promise consistent across product, sales, and paid social ends up with ads that compound, because every new asset is another attempt at the same argument and the results stack into a read. A team without that consistency ends up with a library of one-off ads and no way to say which idea was worth repeating. A framework earns that consistency only when it is operational, which means it changes what gets briefed and killed, not when it sits in a deck.

The useful lens for deciding which message to open with is Eugene Schwartz's five levels of awareness. A message that works for a highly aware prospect won't carry the same weight with someone who's still problem-aware. If the team doesn't know which awareness state it's writing for, the hook becomes a guessing game.

Practical rule: if the first line could be swapped into five different ads without changing the meaning, it's too generic to test cleanly.

Why do so many DTC ads end up sounding interchangeable?

Because most teams start with format before message. They ask whether the ad should be founder-led, creator-led, voiceover-led, static, or UGC-style before they decide what claim the ad is actually making. Once execution decisions come first, the message gets forced to fill whatever container is already in motion.

Interchangeable ads also happen when teams confuse product description with strategic messaging. Listing ingredients, features, or benefits is not the same as making a focused argument. A messaging framework forces a choice. It asks, what is the one problem this ad is naming, what is the one promise it is making, and what is the one proof element that makes that promise believable?

What usually breaks before the script gets written?

The breakdown usually happens at the handoff between research and creative. Teams may have customer reviews, product claims, founder opinions, and competitor examples, but they have not converted any of that material into a shared message structure. So the writer receives inputs, not direction.

That creates a predictable pattern. One ad leans on aspiration, another leans on education, another leans on urgency, and none of them are traceable back to a stable messaging decision. After a few weeks, the team has a pile of creative, but no clean read on which underlying message is actually working.

What does a strong ad message do that a generic one does not?

A strong message creates exclusion. It makes some lines feel usable and other lines feel off-strategy. That may sound restrictive, but it is exactly what helps a team move faster, because the writer is not staring at a blank page deciding between ten possible stories.

A generic message does the opposite. It gives everyone room to interpret. That may feel collaborative in the moment, but it usually produces copy that tries to please every stakeholder and lands with no one in particular. Precision is what makes testing readable later.

What a Messaging Framework Is

A messaging framework is not a brand poem. It is a fixed taxonomy for tagging and clustering ad messaging into pain points, value propositions, proof elements, and calls to action, so creative can be compared across time instead of judged by vibe. The point is normalization. Once every ad is coded against the same message dimensions, you can spot concentration, overused angles, and actual positioning shifts instead of just creative variation. That is the difference between "we made a new ad" and "we learned something about what the market responds to."

Those four dimensions are the tag set you code a finished ad against. The document that generates them has five blocks. The positioning statement locks the one promise every ad has to defend. The target audience block defines who the message is for, what they care about, and what state of awareness they are in. The pillars become message clusters, usually a short set of them, and each one should be stable enough to generate multiple angles without collapsing into the same claim. The proof per pillar is the evidence stack, which keeps the team from building ads on unsupported language. The CTA library gives the team multiple close lines to test without rewriting the whole concept.

Inside a paid social workflow, that structure does real work. It tells a strategist which angle to brief, tells a writer which claim to open with, and tells a media buyer how to label the result so the next round is readable. A working framework is closer to a dataset than a deck, because the goal is to code ideas into Meta and TikTok ad angles, then prune them by hook rate and CPA until the strategy improves week by week. That is why this overview of creative strategy is a clean companion if you are building the framework from scratch.

What is a messaging framework in practical terms?

In practical terms, it is the document that decides what the ad is allowed to mean before anyone starts writing lines. It is not there to make the brand sound polished in a workshop. It is there to reduce drift, so that strategy survives contact with production.

The simplest test is this: if two different writers use the framework, do they produce different executions that still defend the same claim? If yes, the framework is doing its job. If each writer wanders into a different promise, then the framework is probably too vague to be operational.

How is a messaging framework different from positioning?

Positioning is the higher-order market argument. It explains where the brand sits, who it matters to, and why its promise is distinct. A messaging framework takes that higher-order argument and breaks it into fields a creative team can actually use. If the team has already written a creative strategy statement, the framework is the layer underneath it: the statement settles the argument, the framework turns it into a tag set every brief inherits.

Put differently, positioning tells you the territory. A messaging framework tells you what to say inside that territory, what proof belongs there, and how to repeat those decisions consistently enough that performance data becomes useful.

Why should paid social teams care about taxonomy at all?

Because without taxonomy, every result gets trapped inside the ad that ran. You can tell that one script did well and another did not, but you cannot tell whether the difference came from the message, the creator, the cut, the offer, or simple timing. A taxonomy gives the team a way to label the message layer separately from the creative layer.

That matters when you review performance over time. You stop saying, "the founder video worked," and start saying, "decision-relief messaging with educational proof kept outperforming feature-led messaging." That kind of learning compounds.

When is a framework actually usable, not just well written?

A usable framework creates outputs on demand. A strategist can brief from it. A writer can draft from it. A designer or editor can infer what the proof needs to show. A media buyer can name and sort tests with minimal debate.

A well-written but unusable framework usually sounds impressive while leaving too much room for interpretation. It uses broad language like "premium quality" or "results-driven experience" without tying those phrases to a real customer problem, a defendable proof point, or a repeatable angle.

The Components That Make a Framework Usable

A framework only helps if a creative lead can grab it on deadline and turn it into ad angles without a workshop first. That's why each field needs a job. If a field doesn't change what gets written, tested, or killed, it's probably just filler in a prettier document.

Which parts of the framework are doing the real work?

Every field matters, but not every field carries equal weight. The heaviest lifting usually happens in the positioning statement, the audience block, and the pillar definitions. Those are the fields that prevent drift.

If those fields are weak, the rest of the framework becomes decoration. Proof gets attached to fuzzy claims, CTAs get written in a vacuum, and creative choices start standing in for strategy. The framework may still look complete on paper, but it will not guide production in a consistent way.

How do positioning and audience shape the whole framework?

The positioning statement sets the boundary. It tells the team what the brand is allowed to say, and just as important, what it should stop trying to say. The audience block should name the segment, the dominant pain, and the awareness state, because the hook is always written against a specific buyer condition, not a generic market.

A framework gets sharp when the audience block includes a real problem and a real state of mind, not just a persona label.

The pillars are the message clusters the team will keep seeing in briefs, ad copy, and review meetings. Each pillar should be specific enough that an ad can be tagged against it, because the tag is what lets you compare your backlog, your competitors' patterns, and your own historical winners without arguing about wording. If a pillar is too broad, it swallows everything and teaches nothing.

What makes a positioning statement strong enough to brief from?

A strong positioning statement makes one promise in one sentence. It does not try to summarize the whole brand. It simply names the audience, the problem space, and the outcome the brand wants to own.

It should also create productive tension. If the statement is doing its job, some otherwise appealing ideas will no longer fit. That is a sign of clarity, not loss. A positioning statement that accommodates everything protects nothing.

What should go inside the audience block?

The audience block should answer who the ad is for, what frustrating condition they are in, what they have likely tried already, and what they need to believe before they act. That last part matters because awareness shapes the opening claim.

A weak audience block tends to stop at demographics. A strong one adds buying context and mental state. It tells the team not just who the person is, but what kind of inner conversation the hook needs to interrupt.

How specific should message pillars be?

Specific enough that two people tagging the same ad would usually choose the same pillar. If a pillar can mean five different things, it is too soft. If it is so narrow that it only produces one ad, it is too tight.

Good pillars create a middle ground. They are stable buckets of meaning that can generate multiple hooks, scripts, and proof combinations without losing coherence. That gives the team room to explore while still learning at the message level.

Why do proof and CTA usually get underbuilt?

The proof field keeps the team honest. A pillar without evidence is just a preference, and a lot of teams waste time polishing those preferences into scripts nobody can defend. The rule is simple, no proof, no pillar.

The CTA block matters more than teams think. Close lines are not afterthoughts, they're part of the message test, because a soft CTA can hide a strong angle and a pushy CTA can ruin an otherwise good hook. Give the team a short library of closers, then tag which close line pairs best with which pillar.

If the framework can't survive being turned into a tag set, it's not operational enough for paid social.

What counts as proof in a messaging framework?

Proof is any support that makes the claim easier to believe. That could be a product demonstration, a clear ingredient explanation, customer language, process transparency, side-by-side comparison, founder expertise, or a before-and-after narrative that shows the outcome in concrete terms.

What matters is not whether the proof sounds impressive in isolation. What matters is whether it directly supports the claim being made. Good proof reduces skepticism at the exact point where the audience would otherwise hesitate.

How should teams think about CTA options?

Teams often default to one CTA and reuse it everywhere, but different messages create different levels of readiness. An educational angle may need a softer close. A high-intent angle may be able to ask more directly. The framework should capture those pairings on purpose.

That does not mean writing a long list of clever end lines. It means creating a small library of practical asks that match the buyer's state of mind. The best CTA is usually the one that feels like the natural next step after the proof you just showed.

A Reusable Messaging Framework Template

A real framework should be fillable in one sitting, not debated for a week. Use the structure below as the working doc your creative team can pull from. It expands the five blocks into ten rows, because each pillar carries its own proof and the audience block gets easier to write when the problem and the promise have rows of their own.

The point of the template is speed with discipline. A strategist should be able to fill it out, hand it to a writer, and know the writer won't start freelancing the strategy. That is also where a disciplined ad copy process matters, because the framework only works when the message language stays tight from brief to script.

How do you fill out the template without overthinking it?

Start with the fields that constrain the rest. Write the positioning statement, the audience, the problem, and the promise first. Once those are stable, the pillars usually become easier to define because you know what kind of claim they need to support.

Then pressure-test every line by asking whether it would help a writer make a decision. If the answer is no, rewrite it in plainer language. Good framework entries tend to sound less impressive than bad ones because they are built to be used, not admired.

What separates a good entry from a weak one?

A good entry gives the next person a direction they can act on without needing interpretation. A weak entry sounds acceptable in a meeting but falls apart when someone tries to turn it into a hook, proof sequence, or creative brief.

The easiest way to tell the difference is to ask whether the field narrows choices. If it leaves the writer with the same number of options they had before, it is not doing enough work.

How much detail should live in the notes field?

Enough to protect the core idea, not so much that the notes become a second brief. The notes field is useful for constraints that matter to execution, especially when a certain kind of proof must appear early or a recurring misconception needs to be corrected quickly.

It becomes unhelpful when it starts dictating every shot, line, and edit choice. That level of detail belongs in the creative brief. The framework should stay one step upstream from that.

A Worked Example for a DTC Skincare Brand

A skincare brand in a crowded category doesn't need more adjectives. It needs a message system that can survive skepticism, because the audience has already seen every promise in the aisle and on the feed. Assume the brand's current problem is a tired hook library and a climbing CPA on Meta, which usually means the team has run out of distinct message angles before it ran out of edits.

The framework might look like this. The positioning statement says the brand helps simplify routine selection for busy shoppers who want clearer skin without a long ritual. The audience block names a problem-aware buyer who's tired of trial and error. The pillars could be ingredient transparency, routine simplicity, and confidence in choice, with proof attached to each one from customer feedback, product education, and usage guidance.

What does a full framework look like when every field is written out?

Below is the same framework written the way a strategist might actually enter it into the doc. Each field includes the literal sentence that goes into the framework, followed by a hook that could come directly from that sentence.

Positioning statement

• Framework entry: "For shoppers who are tired of guessing their way through skincare, this brand makes routine decisions simpler with clear education and straightforward products."
• Ad hook from that entry: "Skincare gets easier when you stop guessing what belongs in your routine."

Audience

• Framework entry: "Adults who want clearer skin but feel overwhelmed by too many product choices, too many ingredients, and too much conflicting advice."
• Ad hook from that entry: "If every skincare recommendation sends you in a different direction, start here."

Problem

• Framework entry: "They have bought products that sounded promising, but still do not feel confident about what to use, what to skip, or what actually fits their routine."
• Ad hook from that entry: "Buying more skincare is not the same as knowing what to use."

Promise

• Framework entry: "The brand helps make the next skincare step feel clear, manageable, and easier to stick with."
• Ad hook from that entry: "A better routine starts when the next step feels obvious."

Pillar 1

• Framework entry: "Ingredient transparency: explain what the core ingredients do in plain language and show why each product has a job in the routine."
• Ad hook from that entry: "Every product in this routine has one job. Here is what each one does."

Pillar 2

• Framework entry: "Routine simplicity: show how the products fit together so the customer can see exactly what to use and when."
• Ad hook from that entry: "Three products, in this order, and that is the whole routine."

Pillar 3

• Framework entry: "Confidence in choice: reduce decision stress by showing how the brand removes guesswork before the customer buys."
• Ad hook from that entry: "We narrowed the shelf down to one starting point so you do not have to."

Primary CTA

• Framework entry: "Shop the routine that makes skincare decisions simpler."
• Ad hook from that entry: "Shop the routine."

Secondary CTA

• Framework entry: "See how the routine works before you choose."
• Ad hook from that entry: "See how it works."

Notes

• Framework entry: "Open on confusion, show the simplifying proof early, and make the visual sequence feel like relief rather than overload."
• Ad hook from that entry: "Too many products, too many steps, too much guesswork. Here's the simpler path."

The value of writing the fields this way is that nothing stays abstract for long. A writer does not have to infer what "confidence" means or what "routine simplicity" should sound like. The framework already contains usable language that can be lifted, tightened, and tested.

How do you turn one field into an actual hook?

The easiest method is to strip each field down to its tension point. In the audience field, the tension might be overload. In the problem field, it might be uncertainty. In the promise field, it might be clarity. Once you identify that tension, the hook becomes a compressed version of it.

A good hook does not need to repeat the framework sentence word for word. It just needs to preserve the same strategic meaning. That keeps the ad connected to the framework while still allowing the copy to feel natural.

What does a strategist actually write in each block?

Strategists often stall because they think each field needs polished brand language. It does not. It needs working language that is clear enough for the next person to build from.

That is why short, plain sentences usually win. "They feel overwhelmed by too many products" is more useful than "They navigate a fragmented skincare landscape." One sounds ordinary, but it can become an ad. The other sounds formal, but it rarely becomes good creative.

How do the fields connect to one another?

The positioning statement should make the promise legible. The audience and problem fields explain why that promise matters. The pillars break the promise into recurring message paths. Proof makes each path believable, and the CTA gives each path an appropriate close.

If one field cannot be connected to the next, that is usually a sign something is off. For example, if the promise is about clarity but the pillars are all about luxury or status, the framework is pulling in two directions.

How can one framework produce multiple ad angles?

Angle 1, Ingredient transparency.

• Hook: "You do not need ten actives, you need to know what's in the bottle."
• Proof: a clean product breakdown that explains what each core ingredient does.
• CTA: "See how it works."

This angle works because it turns education into relief. The message is not just that the ingredients are good. The message is that the customer can stop guessing.

Angle 2, Routine simplicity.

• Hook: "If your shelf is full but your routine still feels messy, start here."
• Proof: a visual routine map that shows what to use and when.
• CTA: "Shop the routine."

This angle works because it names a familiar frustration and resolves it with visible order. The proof is doing more than showing products, it is making the routine feel manageable.

Angle 3, Confidence in choice.

• Hook: "The hardest part isn't skincare, it's picking the right thing in the first place."
• Proof: customer language that reinforces why the brand removed guesswork.
• CTA: "See how it works."

This angle leans into decision relief, which is often stronger than feature praise in crowded categories.

The important part is what's happening underneath the creative. The framework is not the ad, it's the substrate. The ad is just the selected cut of the message, shaped for one format and one audience problem.

How do you keep multiple angles from collapsing into the same claim?

By making sure each angle earns its own proof path. If two angles use the same claim, the same support, and the same close, they are not really separate angles. They are minor wording variations.

A useful check is to ask what skepticism each angle is trying to resolve. Ingredient transparency may answer, "What is actually in this?" Routine simplicity may answer, "How do I use this without overcomplicating things?" Confidence in choice may answer, "How do I know this is the right place to start?" Different skepticism points create cleaner tests.

When should a team kill an angle instead of rewriting it?

Kill the angle when the underlying message is weak, not just the wording. If the pillar does not connect to a real problem, or the proof is thin, polishing the script usually wastes time.

Rewrite the angle when the message is sound but the expression is flat. In that case, the framework still holds. The execution just has not found its sharpest entry point yet.

Turning the Framework Into a Paid Social Testing Plan

A framework only becomes useful in paid social when it gets turned into a test queue. Start by prioritizing the pillars that are closest to the audience's dominant pain, because those are usually the fastest to diagnose. Don't launch every angle at once. Launch the clearest ones first, then let the market prune the weaker claims before you spend time on polish.

How do you turn the framework into a test queue?

Start by listing the candidate angles under each pillar, then rank them by clarity, not novelty. The first round should contain the messages that most directly reflect the audience problem and the brand promise. This makes the early learning easier to interpret.

It also helps to define the variable you are actually testing. If the goal is to test message, then the hook, proof sequence, and CTA should all stay anchored to the same pillar. That way, the team learns whether the message frame deserves more production energy.

Which angles should get tested first?

Usually, the best starting angles are the ones that sit closest to existing customer tension and have the cleanest proof. They may not be the most entertaining ideas in the room, but they are often the easiest to diagnose.

A common mistake is to lead with the most creative execution rather than the clearest strategic claim. That can create excitement internally while producing muddy results, because the team still does not know what message actually earned the response.

How should you judge outcomes at the message level?

A test should be named so the result can be read without opening the file and guessing. The point isn't aesthetics, it's making sure the verdict can be traced back to the exact message frame that ran.

Hook rate matters early because it tells you whether the opening claim is landing. The bands worth holding your message tests against on Meta come from Motion Creative Benchmarks 2026, which puts 25% at workable, 30 to 35% at strong and 40% or better at elite. What this means for a DTC brand: an angle sitting under the workable mark is a message problem before it is an edit problem, and hook rate and hold rate tell you which of the two you are looking at. If the opening hook doesn't match the audience's dominant problem, identity, or promised outcome, viewers leave before the proof or offer can do its job, so CTR and conversion testing become noisy rather than diagnostic.

Practical rule: test the message frame first, then the edit. If the angle is weak, better cuts won't save it.

What naming system makes message tests easier to read?

Use a naming system that surfaces the strategic choice first. A simple convention is enough, something like audience, pillar, hook type, format, then date. The key is not elegance, it is traceability: when the name carries the audience, the pillar and the hook type, the team can review outcomes without reopening every asset.

This becomes especially important over time. Clean names make it easier to compare different ads that are really defending the same message. Messy names make the archive useless.

How do you separate a weak message from a weak edit?

Start by asking whether the core claim is compelling even before you think about visuals. If the message sounds tired, unsupported, or disconnected from the customer problem, the edit is probably not the main issue.

If the message is strong on paper but the ad still underperforms, then the edit becomes the next suspect. Maybe the proof appears too late, the pacing blunts the hook, or the close weakens the payoff. A framework helps because it lets the team isolate those layers instead of guessing.

What should a first testing round actually look like?

It should be narrow enough that the results teach something simple and useful. That means choosing a small set of distinct message frames, making sure each one has clear proof, and avoiding unnecessary variation that clouds interpretation.

The first round is not about exhausting all possibilities. It is about identifying which parts of the framework deserve more creative investment. Once one pillar starts proving itself, the team can branch into more nuanced expressions of that same message.

Plugging the Framework Into the Weekly Loop

The cleanest way to use a framework is to treat it like the spine of the weekly creative loop. Research pulls language into the pillars. Plan chooses which pillars become angles. Produce turns those angles into scripts, visuals, designer notes, and creator shortlists. Learn tells you which message earned its keep and which one should be cut.

How does the framework fit into the weekly creative loop?

The framework should sit upstream of every recurring decision. It should shape what research gets captured, what angles make it into the plan, what briefs get written, and how results get interpreted afterward.

When teams skip that integration, the framework becomes a one-time exercise. It may still exist on paper, but it no longer controls the flow of work. The weekly loop is where the framework either becomes operational or gets forgotten.

How should research feed the framework?

Research should pull real customer language into the framework, not just confirm what the team already believes. Comments, reviews, and competitor patterns are useful because they show how people talk about the problem in the wild. Once that language is in the pillars, the plan stage becomes a selection exercise, not a brainstorming free-for-all.

What should happen during the planning stage?

Planning should be the moment where the team chooses which messages are worth turning into work this week. That means selecting pillars, identifying the proof each one needs, and deciding what kind of hook expression best matches the audience state.

A good planning stage is discriminating. It does not reward every idea equally. It uses the framework to filter out angles that sound catchy but cannot be defended.

What should production output from the framework?

Produce is where the framework starts paying off fast. Every chosen angle should output something usable, a script, a visual direction, a brief, or a creator shortlist. That is the job of a content team, not a notes app.

How do you make sure the team actually learns from results?

Learning is where many teams underinvest. Every tracked ad should be graded against the business outcome you care about, then fed back into the pillar system so the next week's plan reflects what the market did, not what the brainstorm preferred. If you skip that loop, the framework becomes a document instead of a learning system.

The critical habit is to record the lesson at the message level. Do not just note that an ad won or lost. Note what claim it defended, what proof it used, and whether the CTA matched the buyer state implied by the angle. That is what helps the next round improve.

What to Ship Monday

A usable framework has five blocks filled before the team calls it done: a positioning statement, an audience block, a set of pillars, proof attached to every pillar, and a CTA library. Each one has to be clear enough to brief from without extra interpretation. If any of them is missing, the document is still a draft.

Three habits block progress. Stop making ad-hoc messaging, stop running unstructured brainstorms that leave no tags behind, and stop treating measurement like an optional extra. The fastest teams don't write more slogans, they build a cleaner message dataset and use it to decide what gets tested next.

Ask one question before the next ad goes live, which pillar is this ad defending, and which proof element backs it up. If you can't answer that in one sentence, the ad isn't ready. That's usually the moment to pause, cut the weak angle, and go back to the framework instead of trying to save the script.

How Selzee Runs the Messaging Loop

Most teams already know the manual version of this. The hard part is holding it together while the account keeps moving, because the framework decays the week nobody has time to refill it.

Selzee is an AI content team for ad creative, with its own interface, and it runs the same four stages this article just described, under its own names. Research pulls your customer reviews, ad comments, account data, competitor ads and the organic feed into usable language, which is exactly the raw material the pillars need. Concepts is the plan stage: it turns that language into angles tied to a pillar and a proof element rather than to somebody's taste. Create is produce: it writes the brief, plans the test, and matches creators to the angle. Learning brings the result back as a verdict against the threshold you set, so the next week's plan reflects what the market did.

The distinction that matters here is that the framework stays live. It is not a document someone rewrites quarterly. Every round of briefs and verdicts is another pass at the same pillars, which is what turns a message system into a compounding one.

FAQ

What is the purpose of a messaging framework?

A messaging framework gives the team a shared system for deciding what an ad should say before production starts. Its main job is to make messaging more consistent, more testable, and easier to learn from over time.

How many message pillars should a framework have?

Three is the working default, and the worked example above uses three. Fewer than that and the framework cannot cover the real buying motivations. More than five and the tags stop being comparable, because each pillar collects too little data to read.

Who should own the messaging framework?

One person should be accountable for keeping it clean, but it should be informed by multiple functions. The best owner is usually the person closest to both strategic messaging and production decisions.

How do you know if a framework is too vague?

If different writers produce completely different promises from the same framework, it is too vague. Another signal is when the fields sound polished but do not make brief-writing or creative review any easier.

Is a messaging framework only useful for paid social?

No. It can also improve landing pages, email, and product marketing. Paid social simply makes the value more visible because the feedback loop is fast and the need for repeatable testing is constant.

What is the fastest way to improve an existing framework?

Replace abstract language with real customer tension and attach proof to every major claim. The fastest gains usually come from making the fields more specific, not from making the document longer.

---

If your team needs the message system to actually ship creative every week, Selzee runs the loop for you. It turns customer language, competitor signals, and ad account learnings into briefs, test plans, and creator matches, then brings back a verdict you can plan the next round from.