Playbook
The playbook I still use
I've taught these since SMX Advanced in 2013, because they still work. Product is king.
Product, Content, Amplification
Product, Content, Amplification. You can do really well on two. All three is how you build a business that lasts.
Product is king. I don't care how good your videos are. I don't care how good the media is. A great video cannot save a product people return.
Content is the website, the blog, customer reviews, and the product demos. That work grabs attention, builds the brand, and teaches why the product is different.
Then you put ad money behind it.
Related: What performance branding means.
Performance branding
Anybody doing performance by itself, or branding by itself, is doing it wrong. The right ads do both at once, and the sales pay for the brand.
Most brand virality is paid for.
Proof: Purple and Groove Life. What performance branding means.
How a video ad is built
- Open with a hook (a question, a pattern interrupt or a joke).
- Agitate the problem.
- Show how your product fixes it in a way the others don't.
- Compare it to what they already use.
- Entertain them enough that they don't notice they watched four minutes.
We reverse-engineered it off the Harmon Brothers work, then built it in-house. The Groove version of the hook was "Have you seen this ring?"
You don't even have to do a big production to start.
Diversify or die
Diversity is stability. I've called it Stacked Ads. One channel is a single point of failure. At Purple a team of six ran 15 or 16 channels on purpose. Stop being one-trick ad ponies.
I taught a version of this at SMX Advanced in 2013. Jay Leishman and I co-hosted Stack the Odds around the same idea.
When one channel has a bad month, the others carry the business.
Return by funnel stage
These are the return targets I taught at Resonate in 2019.
- Prospecting, about 1 to 2x.
- Brand search, about 2 to 4x.
- Remarketing, about 3 to 5x.
- Everything combined, 5x or better.
If one stage is way off its range, that's where I look first.
Put 10% on tests that can fail
At least 10% of the media budget tests something that could completely fail. If every dollar has to win this week, you will never find the next hook.
You could be 100% right. Let's test it.
Attribution is directional
Attribution is a four-letter word. I check platform data, brand-search lift, a geo or holdout test and blended MER against each other.
A directional answer beats false precision.
Proven keywords get most of the budget
Proven exact-match terms get most of the budget. A smaller campaign goes hunting for queries you've never seen, and the winners graduate into the core. Phrase match is dead in my book, as a positive match type. Google retired modified broad in 2021, but the split still works.
Each campaign gets its own budget, so the experiments never steal money from the winners.
The 2014 post, comments included: Phrase match is dead.
The crowd is the point
Crowdfunding validates the product, funds production, and builds the tribe before the DTC site has to carry the launch. Kickstarter, then Indiegogo InDemand, then import the orders so reviews exist on day one.
The backers matter more than the money. They're your first reviews, your first feedback and your first fans.
Proof: Purple Pillow, Groove Life, Ozlo. Crowdfunding · Ozlo.
Ads that feed organic rank
Paid traffic drives sales velocity. Sales velocity drives organic rank. Judge the ads on total units. CABN keeps each campaign on one job. Conquest, Awareness, Brand, Non-Brand. I built that structure at Pattern.
A sponsored-only ROAS will lie to you.
The ad picks who sees it
On Meta's current AI stack, creative and creator identity do a lot of the targeting. The system stopped rewarding lazy sameness. Same angle, same format, different creator, and you can reach a pocket the last ad never found.
You're the marketer. Own the angle.
I wrote the longer version for Triple Whale on February 18, 2026. It's Not Andromeda: Inside Meta's AI Ad Stack And Why Nothing is Working. The 2014 version of "don't look like everyone else" is Don't look like every other ad.
Feeling first, logic second
Everything we do is feeling first, logic second. By the time somebody opens ten tabs to compare, they've often already decided they want it. That decision was emotional.
Every one of those numbers is a human on the other side of the device. The moment you forget that, you lose the core of what marketing is.
I talked about this on the Unofficial Shopify Podcast in 2025. Listen to the episode →