The goal of a creative strategist is to keep beating themselves.
There's no one person who just turns out great ads. You need someone who can keep looking at the data and asking: now what do I do next?
Week after week, new ideas over and over.
The best performance we've ever had from media buying alone is 100% ROAS with no new creatives. That's the extreme. From creative, we've seen 5x that on total impact. But if you don't have good media buying, it's really hard to get creative unlocks.
With paid ads you want the lighting and audio just good enough that it disappears. When it feels like Netflix-level production, a defense pops up in people's mind. They think: either I'm watching Masterclass, Netflix, or an ad. Then they see that it's sponsored.
About to run out of stock? Don't cut your ad budget.
Raise price 20-30% instead. It slows conversion velocity, stretches stock to your restock date, and lifts AOV.
Bonus: you learn your real price sensitivity. Most brands find they were underpriced all along.
About to exit your DTC? Be careful how you run sales.
Deep discounts juice top-line revenue, but buyers read the P&L line by line. A high discount-as-% raises alarm bells.
Switch to Gift with Purchase. Cost hits COGS, not your top line. Margins and revenue stay clean.
Europe has built some of the best DTC brands in the world. A few favorites: Loop Earplugs (0 to €190M), air up (€30M to €200M+ in four years on one product), Gymshark (past £600M, 12 straight years of growth), Huel (just sold to Danone for ~€1bn). Who am I missing?
Lift contribution margin by engineering bundles around your shipping weight tiers. Find where cost spikes (e.g. $3 to $5.50 past 0.99 lbs) and build bundles that land just under the line. Shipping is one of the few DTC costs you can cut without inverse CAC impact.
I've tested 20,000+ ads on Meta.
How many creative concepts to produce per month: one per $10k in monthly spend. Drop to one per $5k if you're under $50k/mo.
A concept is the core idea. You make multiple (3) ads off each one.
My whole agency thesis in one line:
Don't be a big agency that does okay work. Be a small agency that does really excellent work.
You can't do a good enough job with one or two hours a week on a brand. So we don't go wide. We go deep.
The top ads are overnight successes.
You watch one and think: they shot that in a few minutes. Cool.
No. That was a year of research and hours of work to get a clean 30-second clip. It just never feels forced. That is the whole skill.
Videos are more information-dense. Statics have cheaper CPMs.
Someone who's already aware and just wants the discount? A static gets them in and out fast.
Someone you still have to convince? Pay the higher CPM for a video that carries more information.
When I review creative, I'm always wary of saying what NOT to do. I'd rather say what to do.
An ad that spent $16 and drove 6 conversions isn't "never run this again."
It's "this is one to test again."
Marketing has no right answer.
There are a lot of frameworks, but it comes down to: here's an idea, can you test it, and do you have the understanding to evaluate it properly?
That is marketing. Knowing that you don't know, and being willing to test a lot of stuff.
The US is shipping world-class DTC brands.
20 of the biggest and best, from SKIMS ($5B) and ALO Yoga ($10B) to Liquid Death ($1.4B) and Ridge.
Bootstrapped, venture-backed, acquired. The DTC playbook is alive.
Who have I missed?
Ads don't always turn back on the same way they were paused.
The longer the gap, the worse it is. You keep the engagements, but you lose the algorithm's trajectory.
78% of shoppers say they'd rather see ads made by people. Marketers trust AI to buy media, not build the brand. Match this in your creative budget. Use AI to multiply outputs after the founder/UGC concept hits. Don't use it to invent the concept.
Amazon's Rufus drove $12B in incremental sales. Now they're selling the tech to other retailers. The DTC implication: Meta CPMs aren't your only acquisition problem. Brand search on Amazon AI is coming for brand search on Google. Build the answer, not just the ad.