We audited 50+ Indian D2C ad accounts by April 2025.
All the brands had one thing in common:
“Performance dropped suddenly. The ad was a winner. ROAS was great and then it just died.”
We hear this a lot. But here’s what’s really going on:
→ You’re optimizing for ROAS.
→ Meta’s optimizing for recency.
↳ What does that actually mean?
→ You want to spend more on what’s converting.
→ Meta wants to spend more on what’s currently getting clicks.
→ If the same audience keeps clicking (even if they don’t buy), Meta keeps pushing it.
→ If you scale an ad with strong past performance but low recent activity, Meta deprioritizes it.
→ And just like that, your “winner” disappears from the feed.
That ad that crushed last week? It hasn’t stopped working and you just haven’t fed it fresh engagement.
If the hook hasn’t changed, if the offer hasn’t been reframed, if the story hasn’t evolved then Meta won’t prioritize it.
Even if it could still convert.
Meta optimizes at the creative level and not just campaigns.
It reads recency signals like clicks, comments, shares.
Not your internal ROAS dashboard.
So if you're scaling based on yesterday’s numbers, but Meta’s algorithm doesn’t see today’s intent, your scale stalls.
↳ Real-world signs this is happening:
→ ROAS was great last week.
→ CTR is still decent.
→ Impressions drop when you scale.
→ Meta starts favoring other ad sets.
→ CAC rises but your funnel didn’t change.
You’re not being penalized.
Meta just thinks it’s time to move on.
↳ How I Recommend Brands to Fix It:
1. Keep creative signals fresh
→ Add 1–2 new variations every week, even slight modular edits (new hook, swapped visuals, new opening line)
2. Build engagement intentionally
→ Warm up your “hero ad” with new audience pools, fresh TOFU is fresh recency
3. Track engagement velocity, not just ROAS
→ Shares, saves, CTR trend line is Meta’s real-time feedback loop
4. Stop chasing old winners.
Start building creative families.
→ A single winning concept split into TOFU, MOFU, BOFU variations sustains scale
Recap:
✅ Meta’s not loyal to yesterday’s winners.
✅ You need to feed the algo with recent intent, not just performance history.
✅ ROAS tells you what happened. Recency tells Meta what to prioritize.
✅ Real scale comes from systems, not surprises.
The ad didn’t break. The signal faded.
If you want Meta to bet on you and give it a reason this week, not last month.
Spending ₹10L–₹50L/month and stuck trying to revive your “winning” ad? Let's Chat.
This post was originally shared by on Linkedin.