The hardest part of fixing a struggling Paid Ads account isn't fixing the ads.
It's telling the owner that the ads aren't the problem.
We audit hundreds of accounts a year. The pattern that shows up most often isn't a broken campaign structure, or a bad hook, or a misconfigured pixel. It's an account running ads for a brand that has no clear position in its market, and no amount of creative iteration or budget reallocation can fix that.
You can A/B test hooks for six months. You can rebuild your campaign structure four times. You can hire a new media buyer, switch to Advantage+ Shopping, layer on a new attribution stack. None of it works, because the underlying problem is upstream of all of it.
Meta isn't broken. The ads aren't broken. The brand has no reason to exist in the customer's mind that's different from the competitor running the same offer in the same feed, and Meta's algorithm can't fix that for you. No algorithm can.
This post is about the work that has to happen before any of the tactical Meta stuff matters. It's a four-phase framework, Discover, Define, Develop, Deliver, that takes you from where the brand is today to where it needs to be in order for paid media to actually do its job.
It's long. It's slow. It's the kind of work most founders avoid because it feels like it's not "doing anything." But it's the single highest-leverage activity in performance marketing, because everything downstream is multiplied or divided by it.
Let's start.
Before We Begin: Two Honest Questions
Before the framework, two questions worth sitting with. Not rhetorically. Actually answer them.
1. What is this brand trying to achieve?
Not your ROAS target. Not your monthly revenue goal. Not "more sales."
What is the brand for? Why does it exist? Why are you getting out of bed in the morning to run this business specifically, rather than any other one?
If you asked Nike what they're trying to achieve, the answer wouldn't be "sell as many shoes as possible." It might be a goal in there somewhere, but the brand has clearly transcended it. Same with Apple. Same with any brand you can name that has actually built positioning.
If your answer to this question is "make more money," that's fine, but recognise that you're operating one level below where the brands that win in your category are operating. They have a reason. You're running a transaction.
2. What do people actually think when they see your brand?
This is the brutal one.
Open your website. Open your Instagram. Open your Facebook ads library. Look at your last ten ads.
If a stranger saw this for the first time, what would they think, at a glance, in three seconds of feed-scroll attention?
- Impressed? Bored? Confused?
- Disrupting the market, or a flash in the pan?
- A real brand, or a gimmick?
Then, separately: what would your best customer say about your brand under a microscope? Not the surface scroll, someone who's actually engaged with it, used the product, gone deep on the offer. What would they say if they were being honest?
The gap between those two answers, the at-a-glance perception and the deep-dive reality, is the gap your brand positioning needs to close.
If those two views are aligned and strong, you have a brand. If they're misaligned, or both weak, you have a logo and a Shopify store. That's not the same thing.
Hold both answers in your head as you read the framework. Every step of what follows is in service of closing those gaps.
The Framework: Four Phases
The framework runs four phases in order. Skipping any phase breaks the whole thing.
Phase 1: Discover. Research. Find out who your customers actually are, what they actually think, and how they actually experience your brand.
Phase 2: Define. Synthesis. Take all that research and turn it into a clear statement of who you are, what you stand for, and where you sit in the market.
Phase 3: Develop. Decision. Choose the direction the brand needs to move in, and turn that direction into concrete creative concepts.
Phase 4: Deliver. Execution. Roll the new positioning out across every touchpoint, website, social, ads, copy, design, message, goals.
Each phase has its own outputs. Each phase feeds the next. By the end, you have a brand that has earned the right to be in market, and a Paid Ads strategy that has something genuinely worth saying.
Phase 1: Discover
The first phase is research, and it's where most brand work fails, because most brands don't do it.
They skip to the design. They skip to the slogan. They skip to the rebrand on the homepage. None of that is positioning. That's decoration. Positioning is built from research, or it's not built at all.
The Discover phase has one job: produce enough genuine knowledge about your customers, your products, and your market that everything downstream is based on evidence instead of opinion.
There are three research streams to run. Take them in order.
Customer Avatars
A customer avatar is a representative person who stands in for a group of your buyers. Most marketing courses teach this badly, they have you write down a name, an age, an income bracket, and a few interests. That's not an avatar. That's a stock photo with a caption.
A real avatar is built from actual research with actual people and answers actual buying questions.
Let's take a concrete example. If you ran an agency-focused paid community, your avatars might look something like this:
Kevin, 32. Owns an agency. Five clients, growing. Wants to improve his own ad knowledge so he can run his accounts better without hiring another media buyer. Cares about price because he's reinvesting profit back into the business. Entrepreneurial. Reads Hormozi. Listens to podcasts on the drive in.
Jen, 41. Chief Marketing Officer at a 50-person company. Manages a team of four marketers. Doesn't run ads herself but needs to understand what her team is doing. Cares about being able to brief her team well and make strategic calls. Has a learning and development budget. Cares about staying ahead of the curve.
Those are two completely different people. They have different problems, different budgets, different language, different buying triggers. An ad written for Kevin would fail with Jen. An ad written for Jen would feel patronising to Kevin.
This is why one-size-fits-all creative doesn't work in 2026, and why Meta's algorithm now rewards creative diversity. Different avatars need different ads.
How To Research Avatars
Most agencies hand you avatars built from reviews and demographic guesswork. That's not enough. You need to go deeper.
Some methods worth running:
- Surveys to existing customers. Open-ended, focused on motivation. What problem were you trying to solve when you found us? What nearly stopped you from buying? What did you almost buy instead?
- Reddit threads in relevant subreddits. People talk honestly to strangers there in a way they won't in your customer support inbox.
- Discord communities in your niche. Lurk first. Then ask questions. Most are happy to help.
- One-on-one interviews with engaged customers or community members. Twenty minutes, recorded, transcribed.
- NPS surveys to your existing customer base. The numerical score matters less than the qualitative feedback you get from the open-text field.
- Post-purchase surveys, every ecommerce brand should be running one always-on.
The crucial framing: you're not selling them anything during this research. You're not pitching. You're trying to understand. The moment the interview becomes a sales call, the research becomes useless.
The goal of all of this is to surface the voice of the customer. You should end this phase with pages and pages of notes, phrases your customers actually use, problems they actually have, objections they actually raise, language that actually resonates with them.
What To Map For Each Avatar
Once you have enough research, build each avatar out across these dimensions:
1. Demographics and identity. The basics, age, location, job, income level. But also the deeper layer: what tribe do they see themselves as belonging to? What's their self-image?
2. Their relationship with your product or service. What's their current experience? Are they aware of you? Have they bought before? What was that experience like? How does your product fit (or not) into their life right now?
3. Where they discover you. What channels are they on? Is it a Facebook awareness ad? A conversion ad? Word of mouth? TikTok? An organic reel? A billboard? Direct mail? Different avatars find brands through dramatically different routes, and the discovery channel often dictates the creative format.
4. Where they convert. And what kind of ad makes them convert. A Kevin-type avatar might convert on a high-information, value-stacking video ad. A Jen-type might convert on a credibility-led ad with case studies and testimonials. Same product, different conversion triggers.
5. The problem that triggers their search. What's happening in their life right now that makes them go looking for a solution like yours, today? Kevin's problem might be "my ads aren't profitable and I can't afford a media buyer." Jen's might be "my team is making decisions I can't evaluate." Different problems, different solutions, different ads.
6. What factors are important to them. Price? Quality? Sustainability? Social proof? Community? Speed? Convenience? Status? Different avatars weight these completely differently. Identifying the dominant factor for each avatar is critical, it dictates which angle the creative leads with.
7. Their deal-breakers. What single piece of information would immediately rule you out for this avatar? If you're a roofer and your site says you're booked until next month, that's a deal-breaker for a customer who needs the work done now. Knowing your deal-breakers tells you what to not put on your site, in your ads, in your sales process.
8. Their "last push" triggers. What would close the deal? For a price-sensitive avatar, a discount might be the lever. For a community-driven avatar, it might be a testimonial video from existing community members. Same final decision, different mechanisms.
9. Post-purchase behaviour. What happens after they buy? Do they review? Do they refer? Do they fill out your survey? Do they come back? Do they complain? Understanding this loop tells you whether you have a transactional business or a relationship business.
By the end of this phase, you should have:
- Two to four fully built customer avatars, each with all nine dimensions above filled in
- Pages of voice-of-customer research in your customers' own words
- For each avatar, a single-line summary of what they want
For our agency-community example, the summaries might be:
Kevin wants a cost-effective community to learn ads from scratch.
Jen wants advanced content to keep her team ahead of the curve.
That single-line summary is more useful than any twenty-slide deck. It's something you can hold in your head while you write copy. It's something you can show to a designer. It's something you can use to filter every creative decision through.
If you can't summarise each avatar in one sentence, the research isn't done.
Phase 2: Define
You have the research. Now comes the synthesis.
The Define phase has a different job from Discover. Discover was about them, who your customers are, what they want, how they think. Define is about you, who you are, what you stand for, what you're offering. Then it puts the two side by side and looks at the fit.
Start with the foundation: your product or service itself. Strip away the marketing for a moment and answer some plain questions:
- Features: What does it actually do? What's in the box? What's included in the service?
- Benefits: What does the customer get out of it? Not "high-quality leather", "shoes that still look new after three years." Features describe the product; benefits describe the customer's life with the product.
- Guarantees: What are you willing to promise? A callback guarantee for the boiler installer. A 30-day return for the ecommerce brand. A pause-anytime policy for the subscription. These aren't features. They're commitments, and they signal something about how confident you are in your own product.
That foundation is the rational layer. Now comes the squishier work, the part most brands either skip entirely or treat as decoration. It's neither. It's the part that determines whether anyone remembers you.
- Values. What does the brand actually stand for? Not the values-page values that read like an HR document. The genuine, lived ones, the ones that show up in how you treat customers when no one's watching. Values are revealed by behaviour, not declarations.
- Tone and voice. How does the brand sound? Warm or formal? Playful or serious? Direct or diplomatic? Punchy or thoughtful? Tone is the thing that makes a piece of copy feel like you even before the customer reads the logo at the top.
- Vernacular. What's the lingo of the brand? What words do you use that your competitors don't? What words would never appear in your copy? The difference between "team members" and "ride-or-dies" is the difference between two completely different brands.
- Vibe and aesthetic. What's the visual identity? Cosy? Industrial? Polished? Raw? Minimal? Maximalist? The aesthetic is what your customer sees in the first 0.3 seconds of an ad, and it decides whether they keep watching.
The Two-Axis Positioning Maps
This is the part of the Define phase that most brand work is missing. Where most positioning conversations stop at adjectives ("we're warm and trustworthy"), this is where you make it visible, by plotting yourself and your competitors on actual axes.
There are three maps worth running.
Map One: Tone. One axis goes from human to mysterious. The other goes from says much to says little.
Plot some real brands. Average Joe-style relatable brand: high human, says much. A capsule-drop fashion brand that releases monthly: more mysterious, says little. A minimalist t-shirt brand like True Classic: human-ish but says little, they're selling t-shirts, they don't need to give a TED talk. Nike: human, but says a lot, they've moved beyond product into something closer to activism. Apple: knows exactly who they're talking to, but isn't trying to be your friend, they sit somewhere in the middle, leaning minimalist.
Now plot yourself. Then plot your three biggest competitors. What you're looking for is this: are you sitting on top of your competitors, or in your own space? If your competitors are clustered in one quadrant and you're in the same quadrant, you have no tonal differentiation. If you're in a different quadrant, or you could move to a different one, you have positioning leverage. You're not better. You're different. And different is far more valuable in a crowded feed than incrementally better.
Map Two: Vibe And Aesthetic. One axis: polished to raw. Other axis: cosy to industrial. Apple: highly polished, fully industrial. A hand-stitched dressmaker in Paris: more cosy, still polished. A skater jeans brand: raw, leaning cosy. A high-performance gaming PC brand: industrial, polished.
Where are you on this graph? Where are your competitors? Are you clustered or separated? The aesthetic map is particularly useful because it forces a decision your creative team has probably been making by default. Most brands' aesthetic emerges by accident. Plotting it consciously surfaces whether the accidental choice is actually the right one.
Map Three: Value. One axis: high quality to low quality. Other axis: expensive to budget. There are four quadrants. Three of them are viable. One is a trap.
- High quality, expensive. The premium quadrant. Apple, luxury brands, premium services. Works if you can sustain the operational cost of high quality and the brand work needed to justify the price.
- High quality, budget. The disruptor quadrant. Hard to sustain, running a high-quality operation on thin margins is operationally brutal. But brands that pull it off (Costco, IKEA at times) build cult loyalty.
- Low quality, budget. The mass-market quadrant. Works if you have the volume and the operational efficiency.
- Low quality, expensive. The trap. There's no story to tell here. You can't generate social proof because the product doesn't deliver. You can't sustain the price because customers churn. This quadrant is where most failing ecommerce brands accidentally end up.
Where are you? Where do you want to be? What needs to change to get you there?
The maps are doing the most important work in this phase. They turn positioning from a vibe into a decision. They surface the gap between current reality and future ambition. They make it possible to point at a specific shift, "we need to move from here to here", instead of waving vaguely at "becoming more premium."
Phase 3: Develop
Define ends with you knowing where you want the brand to move. Develop is where you decide how.
This is the most creative phase of the framework. It's also where most brand projects die, because the temptation here is to skip back to making ads instead of building the connective tissue that lets those ads work.
Themes
The first task in Develop is wrapping your direction into a theme, a short, memorable phrase that captures what you're trying to be.
A theme isn't a tagline. A tagline is what the customer sees. A theme is what the internal team uses to make every decision downstream. It's two or three words, maximum, that anchor the whole brand.
Some examples of what themes can sound like:
- "Sustainable old money." An ecommerce brand for understated luxury basics, made ethically. The theme tells the designer everything they need to know about the aesthetic. It tells the copywriter the tone. It tells the strategist the avatar.
- "Retro oversized." A streetwear brand pivoting toward heritage cuts and 90s-inspired silhouettes. Same product category as plain t-shirts, completely different brand.
- "Only linens." A premium apparel brand built entirely around a single material commitment.
You can probably tell, from just those three phrases, exactly what each brand looks like, sounds like, and sells. That's what a theme is for. It compresses a hundred decisions into three words.
The reason themes matter is that they pull the maps from Phase 2 into a single creative direction. If your theme is "sustainable old money," the maps come together immediately. On the value map, you're high quality, expensive. On the aesthetic map, you're polished, leaning cosy. On the tone map, you're more mysterious than human, and you say little, old money doesn't shout.
Now every creative decision is anchored. If a piece of copy reads too loud, you can name why, it's not on-theme. If a designer mocks up an aesthetic that's too industrial, you can name why. If an ad feels too transactional, you can name why. That's what positioning gives you: the ability to evaluate creative against something other than taste.
Mood Boards And Slogans
Once you have a theme, the work in Develop is to bring it to life concretely enough that the rest of the team can see it.
The exercise: build three mood boards and three slogans, each capturing a slightly different interpretation of the theme. Mood board A + slogan A. Mood board B + slogan B. Mood board C + slogan C.
This isn't about finding "the right one" out of three. It's about giving the team something to react to, debate, refine, and synthesise. You'll often end up combining elements from two or three options into the final direction.
Pinterest is your best friend for this. So is a quiet hour with the team in a room and a whiteboard.
By the end of Develop, you should have a clear brand theme (two or three words), three mood board + slogan combinations exploring different interpretations of that theme, a final synthesised direction with sign-off from the relevant stakeholders, and concrete creative reference points the rest of the team can build against.
Phase 4: Deliver
Deliver is the shortest phase to describe and the longest to execute.
Up to now, everything has been preparation. Research, synthesis, decision-making. None of it is visible to your customers yet. The brand on your website still looks like it did before you started. The ads in your account haven't changed.
Deliver is where the new positioning rolls out across every customer touchpoint, and the discipline here is alignment.
Alignment Is The Whole Game
The most important word in the Deliver phase is alignment.
A brand only works when every touchpoint says the same thing. If your website says "premium" and your Instagram says "discount", you have no brand, you have a confused customer.
If Apple's Instagram suddenly went 100% raw UGC tomorrow, you'd notice immediately. Off-brand. Doesn't fit. Something's wrong. That's what alignment looks like in practice, and the absence of it is what most underperforming brands suffer from without realising.
This is also why pretty agencies that focus only on Meta creative often produce flat results. They make a great-looking ad, but the moment the customer clicks through to the website, the alignment breaks. The ad promised one brand. The landing page delivers another. The trust evaporates. Conversion dies.
What Has To Change
The Deliver phase covers every customer touchpoint. In order of usual impact:
- Website. The hero section, the brand voice in the copy, the imagery, the typography, the colour palette, the offers being presented, the messaging hierarchy. The website is the single biggest signal of brand positioning you control.
- Social channels. Instagram grid, Facebook page header, the editorial tone of your captions, the kind of content you post. Your social presence has to match the brand you're building, not the brand you used to be.
- Ads. This is where most brands start, which is a mistake. The ads have to be the last thing to change, because the ads point to everywhere else. If you change the ads first and the landing page is still old-brand, you've broken the customer journey.
- Email. Your transactional emails, your newsletter, your post-purchase sequence. All of it has to sound like the new brand.
- Customer support. The way your team responds to enquiries. The tone of your help articles. The personality of your auto-responders.
- Product and packaging. For ecommerce specifically, the unboxing, the inserts, the product photography on the PDPs. For services, the proposal templates, the onboarding documents, the welcome materials.
- Internal docs. Brand guidelines, design systems, copy guidelines, tone of voice documents. The internal version is what protects the brand from drift over the next three years.
The New Message
Once the visuals and the tone have shifted, the deeper question is: what's the message now?
After all the research, after the avatars, the maps, the themes, the alignment audit, you should be able to answer four questions with precision:
- Who specifically are we talking to? Not "our customers", which avatar, in what state of mind, at what point in their journey.
- What are we saying to them? The core promise. The single line.
- How are we solving their problem? The mechanism that makes the promise believable.
- What's the call? What do we want them to do, right now, in this specific moment?
That answer is the input to every piece of Meta creative you make from this point on. It's also the test for every ad: does this creative actually deliver on those four answers, or has it drifted?
The final layer of Deliver is reconsidering the goals. Now that you have positioning, your ad goals might change. A brand that was running pure ROAS campaigns might shift toward awareness campaigns to build the new positioning in market. A brand that was chasing scale might pull back to maximise quality during the rebrand period. Your goals are downstream of your positioning, not separate from it.
Why This Is The Highest-Leverage Work In Performance Marketing
Step back and look at what this framework actually does.
Discover gives you a clear picture of who your customer is. Define gives you a clear picture of who you are. Develop gives you a concrete direction. Deliver rolls that direction out across every touchpoint.
The output is a brand that has a reason to exist in the customer's mind, a specific position, in a specific market, for a specific person, with a specific message.
That is what makes Paid Ads work.
Everything that goes wrong inside a Meta account that isn't a technical issue can usually be traced back to a missing answer somewhere in this framework. You can't write a hook for an avatar you've never researched. You can't produce on-brand creative without a defined aesthetic. You can't differentiate from competitors without a positioning map. You can't sustain performance without alignment across touchpoints.
This is the upstream work. The reason most agencies don't do it is that it's slow, expensive in time, and impossible to bill for in the same way as media buying. The reason it matters is that no media buyer in the world can compensate for its absence.
If your Paid Ads aren't performing the way you want, run the framework. Discover, Define, Develop, Deliver. Then go back to the account.
You'll find the problem you've been trying to solve at the campaign level wasn't a campaign-level problem at all.
