Strategy & creative iteration.
Most brands jump straight to producing. That's the mistake. The work that determines whether a creative converts happens before any tool is opened. Personas, hooks, angles, and competitor research. By the end of this module, you'll have a brief ready for production.
Production is cheaper than ever. AI tools have collapsed the cost of generating an asset. What hasn't changed is that plenty of ads underperform because the strategy was wrong before any tool was opened. Production matters too, but no amount of polish saves a creative that targets the wrong angle or speaks to the wrong person.
This module covers the four steps you need to complete before producing anything:
Four steps before production.
Define personas
Spy on competitors
Generate concepts
Write the brief
Define the personas you're producing for
A persona isn't a demographic. "Women 25 to 45 interested in fitness" is useless for creative direction. A useful persona answers a different question: what's actually going on in this person's life that makes them want what I sell?
Most ecommerce brands have 3 to 5 distinct personas. Same product, different reasons to buy. The goal is to make sure every persona has at least one creative speaking directly to them. Even if a persona only represents 10% of your potential sales, you still want those 10%. A generic ad that tries to speak to everyone usually speaks to no one well enough to convert.
For each persona, you need four things written down. The trigger (what makes them notice your category), the pain (the friction in their current solution), the objection (why they hesitate to buy), and the desired outcome (what success looks like in their life). These become the raw material for hooks, scripts, and visuals.
Pull directly from customer reviews, support emails, and post-purchase surveys. The exact phrasing customers use when describing their problem is what should appear in your hooks. If you can't get this from your own data yet, scrape competitor reviews on Amazon or Trustpilot for the same product category.
Spy on what's already working
The Meta Ad Library is the most underused free tool in paid social. Every brand running ads on Meta has their library publicly visible. Search 5 to 10 direct competitors, filter to active ads, and look for ads that have been running 30+ days. If a brand keeps a creative live for over a month, it's almost certainly profitable. Save the winners into a swipe file.
A swipe file is just a place where you save inspiration. Could be a Google Doc with screenshots and links, a Notion page, or a folder on your phone. For each ad, save the link, a screenshot of the visual, the hook (first 3 seconds of copy or video), the format, and a note on what angle it's hitting.
Example entry: "Glossier — UGC talking head, hook: 'I gave up on serums until I tried this'. Pain-led problem statement, social proof framing. Active 47 days."
The goal isn't to copy. It's to recognise which angles, formats, and hooks already work in your category, then put your own spin on them. Spying is also the first step of the Creative Flywheel methodology, which goes deeper into how to systematise this monthly. Read the full Flywheel post if you haven't already.
Frame your hooks before generating concepts
A hook is the first 3 seconds of an ad. It's what determines whether someone keeps watching. Without a strong hook, the rest of the ad doesn't matter because no one sees it. Most underperforming ads aren't bad ads, they have weak hooks.
Six hook frameworks that consistently work for ecommerce:
Problem statement
If your skin still feels tight after washing, here's what's actually happening
Provocative question
Why do most face creams stop working after a month?
Specific transformation
From 14 step routine to 3 minutes a day
Pattern interrupt
Stop putting this on your face if you're over 30
Expert or insider angle
A dermatologist told me this is the only thing that actually works
Social proof opener
12,000 customers, one product. Here's why
For every persona, draft 3 to 5 hooks per framework. You won't use all of them. The point is to give yourself a wide range of starting angles before committing to production.
Generate concepts at scale with the HWC Ad Generator
Doing all of the above manually for every persona is exhausting. The HWC Ad Generator was built specifically to compress this work. It runs on the Claude API (Anthropic's most capable model) to generate personas, competitor analysis, hooks, and creative concepts. Feed it your product details and target market, and you get strategy outputs ready for the production phase.
Compresses days of strategy work into a single session. Personas, competitor analysis, hooks, and concept variations across formats.
The Generator runs through the same flow you've just learned, but at speed:
Personas
Generates 3 to 5 distinct personas based on your product. You can regenerate or refine them. The output covers trigger, pain, objection, and desired outcome for each.
Competitors
Analyses competitive ads in your category. Surfaces angle patterns and gaps you can exploit.
Strategy
Provides a written angle strategy across personas, with reasoning for each angle.
Hooks
Generates dozens of hooks across the six frameworks above, tagged by persona. You can regenerate hooks for any persona until you find ones that resonate.
Concept variations
Produces creative concept variations: full ad ideas including hook, body, format suggestion, and CTA. These are your starting points for production.
Treat the Generator's output as a first draft, not a final answer. Refine, regenerate, mix angles across personas, and sanity check against the Meta Ad Library before locking in your brief. Your judgement is still required, the Generator just gets you to that judgement faster.
Your output: a brief ready for production
Strategy isn't done until it's written down. The brief is the bridge between thinking and producing. Without it, production becomes guesswork. With it, every tool you use in Module 03 has clear instructions.
A complete brief contains the following:
What to capture before producing anything.
The single persona this batch is for. Trigger, pain, objection, desired outcome.
The specific creative angle. e.g. "highlighting the simplicity of a 3 minute routine for busy parents"
3 to 5 hooks tied to the angle. Which framework each hook falls under.
Video, image, carousel. If video: length and rough scene structure.
Lifestyle, demo, talking head, before/after, UGC style. Reference ads from your swipe file.
The core point the ad needs to deliver. One sentence summary.
What you want the viewer to do next. e.g. "Shop now", "Try free for 30 days"
How many variants of this concept. Different hooks, different visuals, different durations.
Paste your Generator output, your swipe file notes, and your persona research into your LLM. Ask it to draft the brief in this exact format. You'll get a starting point in 30 seconds. Edit and refine. We'll go deeper on this orchestration approach in Module 03.
What you have at the end of this module
By the end of the strategy phase, you should have:
3 to 5 personas with trigger, pain, objection, and desired outcome
A swipe file with competitor winners, hooks, and angles to draw from
A bank of hooks across the six frameworks, tied to each persona
One brief per concept ready to feed into the production pipeline
With those four things in place, you're ready to start producing. Module 03 covers the production pipeline itself: how to chain AI tools to turn each brief into finished ad creatives.