Hi Reader -

That’s Hemingway. 1926.

Accounts are the same way. Some crash overnight, some bleed out over weeks while you sit there thinking everything’s fine. Same cause either way: staring at the wrong numbers while the whole thing quietly falls apart.

We've come close to the brink on 3 accounts. Each time, the warning signs were there for weeks. Each time, we nearly missed them.

​The overnight crash. Account cruising at $10k/day with one ad doing all the heavy lifting. Dashboard looked beautiful. Then it slipped - and when one ad carries everything, “slipped” turns into “on fire” real fast.

​The slow bleed. Subscription app where cost per trial was rock solid while ROAS quietly bled out. Dashboard said everything was fine. Dashboard was lying. Took statistical analysis to figure out what was actually happening. Yes, stats. We’re those people now.

​The silent killer. Great ad, great metrics, a “winner” by every measure anyone bothers to check. Except downstream retention was garbage. A magnet for users who’d never pay. A win that was slowly poisoning the product.

Different symptoms, same disease: concentration.

And the annoying part is that concentration feels smart. Feels like focus. It’s actually a bomb with a timer you can’t see.

After these 3 near-misses seared into my mind the dangers of concentration in different flavors, I decided we'd make creative diversity a cornerstone of everything we do.

Not because Meta said 'Andromeda' - but creative diversity as insurance against things blowing up. And we track this in our key dashboards(even when, especially when Meta doesnt surface these by themselves) - that is how critical diversity is to everything we do. Here are just a few things we track.

• Spend freshness. What percentage is going to ads from this month versus three months ago? If everything scaling is old, your pipeline is dead and you don’t know it yet.
• Format mix. This is how we caught the slow bleed. Statics quietly took over, algorithm found worse users, cost per trial looked fine but the humans were garbage.
• Message concentration. If one angle is doing 70% of the work, that’s not strategy. It’s a bet with a clock on it.
• Velocity. No pipeline means concentration creeps back in. And then you’re right back where you started.
• Placement distribution. Do we have a sufficiently diverse distribution across placements?

...and a LOT LOT more(below is a small subset of the metrics we track in our internal dashboards).

This is what we're looking at every week while most people are refreshing their ROAS and tweaking bid caps. Different view, different game.

This isn't optional for us. This is the work that actually moves the business forward. Not the sexy stuff: the stuff that keeps accounts alive when things get weird. And things always get weird eventually.

None of this is in ads manager. Unlikely it ever will be: Meta's pushing for automation and concentration, not diversification(notice how Meta reps talk about account consolidation in one breath and creative diversity in the next one? LOL).

Anyway, if you're spending over $50k/mo and want to scale in a way that diversifies your risks and creative, hit reply and say "DASHBOARD". (Or don't. Ads manager will still be there tomorrow).

Later,

Shamanth

PS: From the socials:

​𝗡𝗼𝘁𝗲𝗲(AI notetaker) 𝗶𝘀 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗻𝗴 ~𝟴𝟬𝗞 𝗱𝗼𝘄𝗻𝗹𝗼𝗮𝗱𝘀 𝗮𝗻𝗱 ~$𝟯𝟬𝟬𝗞 𝗶𝗻 𝗺𝗼𝗻𝘁𝗵𝗹𝘆 𝗿𝗲𝘃𝗲𝗻𝘂𝗲. I ran semantic analysis on 50 of their video ads.​

​2013, early advertiser on Facebook install ads.​

​I scaled an app to $1.5M/month in ad spend.​

​Everything we know about ChatGPT ads:​

​"You need to fight creative fatigue" - LinkedIn Bro​