creative strategy, as we actually do it.

Your agency is still optimising audiences.

Most agencies have a creative team and a media team. The creative team makes the ads. The media team runs them. Sometimes those teams are in different rooms, sometimes at different agencies. Either way, the structure is the same: someone makes the ad, someone else figures out who to show it to.

That structure made sense for fifteen years. The auction was loose, audiences were targetable, third-party data was gold. A good media team could carry mediocre creative with smart lookalikes and interest stacks. Targeting was the lever. Creative was the wrapper.

Then privacy laws came in. The iOS update. The death of the cookie. Andromeda in 2024. The targeting layer collapsed in slow motion, then all at once. The lookalike audiences you spent three years building stopped working. Interest targeting started feeling like a coin flip.

Your agency is still optimising audiences. The algorithm doesn't need them to.

What replaced the old layer is creative signal. Meta's system now reads the ad itself — the imagery, the hook, the message, the format — and decides who to show it to. The creative is the targeting now. Different messages attract different people. A static about period cramps reaches a different person than a video about convenience. Same product, different message, different audience. That's the targeting.

Which means the conversation about how to grow a digital brand has moved. The question used to be who are we showing this to? The question now is what are we actually saying?

This page is about how we answer that question.

What we've been doing since 2014

Weird started as a brand and content studio in Mumbai. Brand identity, websites, films, ads. We picked up performance marketing early because the ads we were making needed to run somewhere, and running them ourselves taught us how the platform worked. The two practices ran in parallel for twelve years, which is unusual. Most agencies pick a side.

What that combination produces is specific. The creative knows what scales because it was made by people who've watched it spend. The media is read by people who wrote the ads. The two halves have been teaching each other for a long time, and what we ended up with is a creative strategy process built for how Meta's auction actually works now.

The process has six phases that run as a continuous loop. Research feeds strategy. Strategy produces briefs. Briefs become creative. Creative goes live. Performance feeds back into research. Winners get compounded, not replaced.

There's nothing magical about it. The advantage isn't a trade secret. It's that we actually run it, every week, for every account, without exception.

Phase 1: Research as a weekly habit

The first thing most agencies get wrong is treating research as an onboarding exercise. A kickoff workshop, a brand brief, a customer summary deck. The deck goes into a Drive folder. The ads start getting made on instinct.

Research isn't something you do once. It's the weekly operating rhythm of the whole agency. The brands producing consistent winners don't have better ideas. They have better inputs.

A week looks like this: Monday is account data — top three ads vs bottom three, by spend. The gap between them is the next brief direction. Tuesday is comments — twenty minutes on viral posts in the client's niche, logging new objections and transformation stories. Wednesday is reviews — 500+ recent reviews fed to Claude to extract customer language, emotional state, failed solutions, desired outcomes. Thursday is competitor audits via Meta Ad Library, looking at ads running 90+ days. Friday is brief finalisation for the following week.

Every brief has to cite a source. A review, a comment, a competitor signal, an ad account data point. If the source can't be cited, the brief isn't ready.

Winning angles are found through research, not invented through brainstorming. Briefs that can't be sourced are guesses.

Phase 2: The strategy map

Once research is feeding the system, the next phase is structural. Most agencies make creative the way you might cook a meal — taste a few times, adjust, see what works. The problem is the agency ends up running 80-90% of its ads at bottom-of-funnel, against the same persona, in the same format, without anyone noticing. The creative looks diverse because there are fifteen different statics. The strategy is identical.

The strategy map fixes that by treating creative direction as a math problem before it becomes a creative problem.

Product × Persona × Angle × Awareness Level = distinct creative directions

The FBU breakdown — Features, Benefits, USPs — produces the raw material. Features are facts. Benefits answer "so what?" USPs are what no one else can claim. Most agencies stop at features. Features create information, not desire.

From the FBU, we derive personas. Not demographics. Why someone buys, not who they are. A persona isn't "women 25-45 interested in skincare." That's a media buy. A persona is the specific buying motivation — the failed solutions they've already tried, the emotional state they were in when they searched, the lifestyle gap the product fills.

For each persona, we map awareness levels: unaware, problem-aware, solution-aware, product-aware. The same person at each stage needs a different conversation. The unaware person doesn't want a product. They want to be told there's a problem. The problem-aware person doesn't want a solution. They want to know what kind of solution works. The product-aware person doesn't need to be sold; they need to be reminded.

For each combination, we pick a format. Static, UGC, brand video, podcast clip, interview. The format isn't a style preference. It's a targeting decision. A static reaches a different person than a 60-second interview, regardless of what either says.

The math, fully worked: one product × three personas × three pain points × five awareness levels = forty-five distinct creative directions, before format variation. In practice we focus on two or three personas and their strongest angles, which gives roughly fifteen concepts per round. With three hook variations per concept, that's forty-five ads with genuine diversity — diversity that's structural, not aesthetic, built into the math before anyone writes a single line of copy.

Phase 3: The brief

A good brief is specific enough that two different designers, working independently from the same brief, would produce recognisably similar work.

Most agency briefs aren't this. They're vibes documents. "Make it feel premium. Show the product hero. Maybe a UGC angle." Then the designer makes their best guess, and either the strategy gets through to the ad or it doesn't, depending on luck.

Our briefs name five things, every time:

The persona, by buying motivation. Not the demographic. The reason this specific person is in the market.

The angle, with a cited source. Where this hook came from — a review, a comment, an ad account signal. The source is in the brief so the designer can read the original.

The awareness level. So the designer knows whether the customer needs to be informed, persuaded, or reminded. Different awareness levels need different opening lines, different proof structures, different CTAs.

The format. Static, video, UGC, interview. Matched to the awareness level, not the strategist's preference.

Five hook variations, minimum. Not one hook with copy variations. Five genuinely different opening lines on the same concept. In practice, one of five hooks usually spends significantly while the other four spend almost nothing. Without all five, the winner never runs.

Briefs are named according to a convention — persona, angle, awareness level, format, variant. When an ad wins, we know exactly what worked. When an ad fails, we know exactly what we tested. The naming is the institutional memory.

Phase 4: Production

The handoff from strategy to production is where most agencies bleed quality. The strategist has the context, the designer doesn't, and the brief is too thin to transfer the thinking. The ad gets made on the designer's instinct, and the strategy that justified it never makes it into the work.

Our production phase runs against the brief, not against the strategist's mood that day. The brief is specific enough that the designer can execute without constant clarification. When clarification is needed, the brief gets updated, not the strategy.

The designers are the same designers who made the ads in 2023, 2024, 2025. Same people, same brands. They've read the briefs. They've seen what scaled and what didn't. There's an institutional knowledge inside the production team that no new hire could replicate in less than a year, which is why we don't replace people.

Phase 5: Launch and the daily reading

Ads go up on Friday. Every Friday. Regardless of what the previous week's data said.

This confuses most agencies, because most agencies are reactive. They upload more when things are going well and freeze when things are going badly. We upload on the same cadence either way, because the cadence is the system. The upload doesn't depend on the result. The result depends on the upload.

What changes is the reading. Every morning, the account gets read. Not metrics-dashboard-checked. Read. We're looking for three things:

Where Meta is placing spend. Spend allocation is the primary signal. When the algorithm pushes spend to a specific ad, that's a vote of confidence from a system that's already analysed billions of behaviours. An ad taking 20%+ of total spend is a top contributor. Under 10% is Meta still learning. The shape of spend across ads tells us which angle the market is responding to.

CPMr. Cost per mille reached. CPM × frequency. This is the early warning metric. ROAS is a lagging indicator — by the time ROAS drops, the damage happened two to six weeks ago. CPMr tells us what we're actually paying to reach a unique person. When CPMr rises for two consecutive weeks, CPA will follow. Frequency tells the same story from a different angle: around 1.1 per day means Meta is prospecting, above 2 per day means it's recirculating the same audience. The trigger isn't to increase budget on what's already running. It's to brief new creative against a different persona.

Creative diversity. Weekly, we check the demographics on the top five spending ads. If they all hit the same age, gender, and geo profile, the account has saturation on one segment — not diversity. The signal there is to brief a different persona, not more variants of the same one.

The reading is interpretive work. It can't be fully automated, though we've built an in-house intelligence system that pulls the data and posts the day's reading to Slack at 8am, so the team is briefed before the first coffee. The point of the reading isn't to make decisions for the strategist. It's to make sure the next decision happens with actual market data, not memory.

Phase 6: Compound the winners

When an ad wins, the work is beginning, not done.

Most agencies do the opposite. They get a winner, scale spend on it, ride it until it dies, then start the next test from scratch. The institutional knowledge from the winner doesn't compound. The agency restarts every six weeks.

We treat a winning ad as a signal. The signal can be extracted three ways:

Same hook, new body. Take the winning hook and place it on a previously winning body. Two proven elements combined into a new ad that reads as genuinely different to Meta's system. The signal compounds.

Same angle, new format. Turn a winning UGC into a static. Turn a winning static into a podcast clip. Same core message, different delivery. Different formats reach different people. The signal extends.

Five hook variations on the same body. Minimum five. Not three. Five genuinely different opening lines on the same proven concept. One of five usually spends significantly. Without all five, the winner never runs at scale.

Every winner gets logged in the angle leaderboard — a running document per client with persona, angle, awareness level, format, total spend, ROAS, CPMr at peak, status. Updated every Monday. Institutional memory that compounds across months instead of starting from scratch each round.

When CPMr rises for two weeks on an ad set, we don't increase budget on the fatigued creative. We brief new creative against a different persona or awareness level from the strategy map. The map guarantees we always have unused directions to deploy. The system never runs out of next moves.

How this becomes brand building

The reason this whole apparatus exists isn't just to make better ads. It's that the loop produces something most agencies can't: an ongoing, evidence-based answer to the question what is this brand actually about?

The creative testing isn't just optimisation. It's a conversation with the brand's market. Which messages compound. Which images stop people. Which positioning customers respond to in their own words. That information is the brand, more accurately than any positioning workshop could produce — because it's based on what people actually do, not what they say they'd do.

That information flows into everything we build downstream. The website that converts because it says what the ads taught us to say. The brand film whose script is built on signals from a thousand variants. The packaging and identity grounded in the positioning the ads are pressure-testing every week.

The agency's job isn't to make ads. The agency's job is to figure out what the brand should say, and then to say it a thousand different ways.

Most agencies build brands top-down — strategy deck, identity system, everything else made to match. We build them top-down too, because the foundation work matters. But we also build them bottom-up, with the ads acting as the brand's ongoing conversation with its own market. Both directions, simultaneously.

The principle

Ninety days of systematic research and iteration compounds more than three years of spray-and-pray. The brands that consistently produce winners didn't get lucky. They built the system first, then they scaled it.

This is the system.

If you want to see it applied to specific brands, the work is here. If you want to talk about whether it fits what you're building, book a call.