Nathan Perdriau

$1.2B+ generated for DTC brands through paid ads.
Free Audits for Brands Doing $5M+ p/yr
@

[Nathan Perdriau @PerdriauNathan]

8x ROAS on Google, 2.5x on Meta. At face value, the answer looks obvious.

It isn't.

Someone sees your Meta ad, doesn't click, searches your brand three days later, and buys.

Channel ROAS tells you who's closest to the point of demand. It does not tell you who created it.

[Nathan Perdriau @PerdriauNathan]

Most brands judge in-cart upsells on AOV lift. That is exactly how you end up scaling your lowest margin SKUs. The metric you optimise to determines the outcome you get.

[Nathan Perdriau @PerdriauNathan]

Advantage Plus doesn't learn. It skips the learning phase, using historical conversion data to hit your highest intent buyers first.

That's why it starts strong then dips.

ABOs were the opposite. Learn first, improve over time.

Expect the dip. It isn't the campaign failing.

[Nathan Perdriau @PerdriauNathan]

Every ad has a fixed spend capacity. You can burn it fast or stretch it over six months, but you can't increase it. So the question isn't how to scale an ad. It's how many ads you need.

[Nathan Perdriau @PerdriauNathan]

@AlessandroLavis stopwatch

[Nathan Perdriau @PerdriauNathan]

Most brands track creative hit rate and ignore what their average winner actually spends. Two accounts can both sit at 4%. Only one of them scales.

[Nathan Perdriau @PerdriauNathan]

Audited a teeth whitening account running broad creative with no structure. Every ad in one pile.

The mistake: testing ads without concept-level segmentation means your data is noise.

The fix: build a matrix. Persona, angle, offer, ad type. Each ad set reflects one concept.

[Nathan Perdriau @PerdriauNathan]

Retargeting is not reminding people to buy. It is removing the reason they didn't. If your site converts at 5%, each of the other 95% had an objection, and each objection needs its own asset.

[Nathan Perdriau @PerdriauNathan]

Most Meta accounts are structured backwards.

There are only 2 levels that matter:

Campaign level = business logic
Ad set level = learning logic

Most scaling problems aren’t media problems.

They’re structure problems.

Fix the structure first.

[Nathan Perdriau @PerdriauNathan]

The top performing AI-generated ads across several categories right now are showing products doing things the products cannot actually do.

And it is working, for now.

The cost is double. Refunds on the front end. Dead trust on the back end.

Show what it does. Full stop.

[Nathan Perdriau @PerdriauNathan]

Segmentation feels like control. It is fragmentation. Most accounts have too many campaigns, not too few, and the 50 conversions a week rule people structure around was never real.

[Nathan Perdriau @PerdriauNathan]

Native statics are quietly outperforming polished creative in the supplements space right now.

The image stops the scroll. The copy does the selling.

The creative is not the hook. The story is the hook. The image just earns the first second.

[Nathan Perdriau @PerdriauNathan]

Most fashion brands judge paid media on first purchase gross profit against CAC. Returns make that ratio a fiction, and usually by half.

[Nathan Perdriau @PerdriauNathan]

Copying creative is a volume problem, not a moral one. Ten clones out of 9,000 ads a month is noise. Ten clones out of 20 ads a month is your brand.

[Nathan Perdriau @PerdriauNathan]

Ignore it for LTV measurement. Judge repeat behaviour by cohort, not by a ratio that improves when you acquire less. Full stop.

[Nathan Perdriau @PerdriauNathan]

Shopify's default LTV metric is not a repeat rate. It is returning customer orders divided by total orders. Which means it looks better when your acquisition gets worse.

[Nathan Perdriau @PerdriauNathan]

There is one use for it. Slicing new versus existing during a sale period to see where the promotion actually resonated. That is it.