Not sure how accurate this insight into Meta’s ad algorithm is, but it’s still helpful to think through.

The 15-Second Framework:
• 0-3 sec: Hook
• 3-5 sec: Value prop
• 5-8 sec: Benefits
• 8-15 sec: Problem/Solution

I’d make that the 16 second framework and extend the Value Prop or Benefits if needed. Use ThruPlays to your advantage.

What Meta’s AI actually sees:
→ Setting = income level
→ Person = demographic
→ Language = education
→ Problem = life stage
→ Music = cultural signals
→ Colors = emotional state

Computer vision + sentiment analysis + trove of behavior data


The key to branding and messaging in a post-AI world? Personality

Average is easy to churn out at a superhuman clip. Stand out by saying things that only apply to your brand.

I’ve seen this advice a lot lately, this classic New Balance ad is a great example.

Say something only you can say.

A gray New Balance 990v5 sneaker is displayed with the caption, Worn by supermodels in London and dads in Ohio.

The Year of the Splinter 🐀 continues (more of a decade, maybe an era).

Relevant headlines:

The top 10% of Americans account for nearly half of consumer spending

highlighting the growing divide between high-income earners and middle/lower-income earners

Over half of consumers consider grocery prices a ‘major stress’

Nearly 9 in 10 US adults (88%) are stressed about grocery prices—including 53% who say food costs are a major source of stress

Gen Z holiday spending tanks in 2025

See also:

Holiday spending among US consumers is expected to decrease by 22.5% for Gen Z in 2025, while other generations show slight increases.

Plants need water to grow.

But they also need periods without water.
Watering everyday has the same end as never watering.

Algorithms want you to water your audience daily. Because it benefits them.

What is the right cadence for the results you want?


Beware your metrics: Google capped search results at 10 / page. No more 100 results on a page.

A lot of rankings tools are reporting issues and I’m noticing improved reported average ranks across Google Search Console accounts, likely from deep page data loss.

How are ChatGPT results changing?