Meta Ad Fatigue in India: How Often You Actually Need to Refresh Creatives

Every founder we onboard asks some version of the same question: "my ROAS was great for three weeks and now it's falling apart — did the algorithm change, or did my ad just get tired?" Almost always it's the second one. Creative fatigue is the single most common cause of unexplained performance decay we see across the Meta accounts we manage, and it is also the easiest one to fix — if you catch it early enough.

The problem is that most teams catch it late. They wait for CPA to blow past target, then scramble to brief a new creative, shoot it, get it approved, and launch — by which point they've usually burned 5-10 days of budget at a degraded return. This post is about the actual signals that predict fatigue before CPA moves, and the refresh cadence that works for different product categories in the Indian market specifically, not the generic "refresh every two weeks" advice you'll find everywhere else.

What creative fatigue actually is (and isn't)

Fatigue is not "your ad is bad." It's a supply-and-demand problem inside Meta's auction. A finite pool of people in your target audience sees your ad repeatedly. Each additional exposure produces diminishing engagement — click-through rate drops, and Meta's ranking algorithm quietly de-prioritises the ad in the auction because its own predicted engagement score is falling. To hold your position in the auction, Meta charges you more per impression. That's the mechanic behind almost every "CPM suddenly went up for no reason" complaint we hear.

Three things get confused with fatigue that aren't fatigue at all:

• Audience saturation — your audience itself is too small and you've simply run out of new people to show the ad to, independent of how many times any one person has seen it.
• Seasonal demand shift — CPA rising because competing advertisers entered the auction (very common in the run-up to Diwali, Big Billion Days, or IPL season), not because your creative is stale.
• Landing page decay — the ad is still performing but a broken checkout flow or slow page load is killing conversion rate downstream.

Diagnosing correctly matters because the fix is different in each case. Swapping creative when the real issue is audience saturation just burns a new asset without solving anything — you need to widen the audience or move to Advantage+ instead. This is one of the reasons we built account-level diagnosis into AdsSarthi's unified campaign dashboard rather than a single "fatigue score" that conflates all four causes into one number.

The three leading indicators, in the order they move

Across the accounts we manage, the same sequence plays out almost every time a creative is genuinely fatiguing, and the signals appear well before CPA does. If you're only watching CPA, you're watching the last domino fall.

1. Frequency crosses 2.8-3.2

Frequency — average number of times a unique user has seen your ad in the selected window — is the earliest and most reliable predictor we track. In our data, the median Indian D2C account starts seeing engagement decay once 7-day frequency crosses roughly 2.8, and the decay becomes steep past 3.5. This threshold is lower than a lot of US-focused guidance assumes, and we believe it's because Indian audiences on Meta skew younger and more mobile-first, spending more total time in-feed, which compresses the exposure window. A campaign targeting a broad 25-45 metro audience can sustain frequency near 4 before real damage; a narrow retargeting audience of past purchasers fatigues far earlier, often by frequency 2.2-2.5.

2. CTR drops 15-20% from its first-week baseline

Click-through rate is the second signal, and it typically starts sliding 3-5 days after frequency crosses the threshold above. We don't look at raw CTR in isolation — we compare it against the same creative's own first-week average, because absolute CTR benchmarks vary wildly by category (a ₹399 fashion impulse-buy ad and a ₹45,000 furniture ad will never have comparable CTR). A 15-20% relative drop from baseline is the point where we flag a creative as "fatiguing" rather than "fatigued."

3. CPM rises while relevance/quality ranking falls

This is the signal most advertisers notice first because it hits the wallet, but it's actually the last one to move — by the time CPM is visibly climbing, frequency and CTR degradation have usually already been underway for 5-9 days. If you wait for CPM to confirm fatigue, you've already lost most of that window.

Category-wise refresh cadence: there is no single right answer

The single biggest myth in creative fatigue advice is that there's one universal number — "refresh every 14 days" or "every ₹50,000 spent." Spend velocity, audience size and purchase cycle length all change how fast a creative burns through its useful audience. Here's what the cadence actually looks like across the categories we work with most:

• Fashion & low-ticket D2C (₹300-₹1,500 AOV): fastest fatigue cycle. High spend velocity against comparatively narrow lookalike audiences means creatives often need refreshing every 7-10 days, sometimes faster during a sale period when daily budget is pushed 2-3x.
• Beauty & personal care: similar to fashion but slightly slower — 10-14 days is typical, extending to 3 weeks for hero/brand-story creative that isn't carrying the full conversion load.
• Home & furniture (higher AOV, longer consideration): 18-25 days. Smaller in-market audience at any given time and a longer purchase cycle mean the same creative can serve profitably for close to a month before frequency becomes a problem.
• Real estate & high-ticket lead gen: longest sustainable window — often 25-35 days — because audiences are broad (city-wide) and a single lead can be worth lakhs, so the marginal cost of slightly elevated CPM barely dents ROI.
• Marketplace-adjacent retargeting (cart abandoners, past purchasers): shortest window of all — often under 7 days — because the audience pool is small by definition and gets re-exposed constantly.

The practical implication: if you're running a fashion brand and a furniture brand out of the same agency dashboard with a blanket "refresh every 2 weeks" rule, you're refreshing the furniture ads too early (wasting production budget) and the fashion ads too late (bleeding CPA for a week before the swap lands).

What "refresh" should actually mean

Refreshing doesn't have to mean a full reshoot. In order of effort and typical effectiveness, based on what we've seen recover performance across accounts:

1. Hook swap — change only the first 2-3 seconds of a video or the headline of a static, keeping the core asset. Cheapest, fastest, and surprisingly effective against early-stage frequency fatigue because Meta's ranking model weighs early engagement heavily.
2. Format rotation — take the same message and re-cut it as Reels vertical vs. feed square vs. carousel. This resets exposure counting per placement and works well when the underlying offer is still strong.
3. Language/vernacular variant — swapping a Hindi or regional-language version in for an English creative targeting the same audience segment effectively creates a "new" ad from Meta's relevance-scoring perspective, while reusing the same visual asset. This is one of the fastest low-cost refresh levers available, and it's why AdsSarthi's 13-language vernacular creative generator gets used as much for fatigue-fighting as for new-market expansion.
4. New concept — a genuinely new creative idea, offer angle, or shoot. Reserve this for when hook swaps and format rotation have stopped moving the needle, because it's the most expensive and slowest to produce.

Building a refresh calendar that doesn't rely on you remembering

The honest reason most Indian D2C teams under-refresh isn't ignorance — it's bandwidth. Someone has to notice the frequency creeping up, brief the creative, get client sign-off, and swap it in, all while also running five other campaigns. A refresh cadence only works if it's a system, not a memory.

A few structural things that actually move this from "we know we should" to "it happens":

• Set a frequency alert threshold per audience type (broad vs. retargeting) rather than one number account-wide.
• Pre-brief your next 2-3 creative variants before you launch the current one, so the swap is a same-day action, not a 5-day production scramble.
• Route the "this creative is fatiguing" signal to wherever your team actually checks daily — for most of the founders and agency owners we work with, that's WhatsApp, not a dashboard tab nobody opens. Our morning digest surfaces fatiguing creatives at 8 AM IST with a one-line reply-YES-to-approve refresh recommendation, exactly because "log into the dashboard" is where most fatigue-monitoring intentions die.
• Tag creative variants by hook/format/language at upload time so you can tell, three weeks later, whether it was genuinely a new concept or just a colour swap — this matters when you're deciding what to try next after a refresh underperforms.

If you're managing this across multiple accounts or clients, the manual version of this — checking frequency across dozens of ad sets daily — doesn't scale past a handful of accounts. That's the gap we built the fatigue-detection layer inside AdsSarthi to close: it watches frequency, relative CTR decay and CPM drift continuously and flags creatives before CPA moves, not after. If you want to see where your live accounts stand against these thresholds, our free 60-minute AI audit runs this exact check against your real Meta data and sends the findings to WhatsApp with no commitment required.

When refreshing too often is also a mistake

It's worth saying plainly: over-refreshing is a real cost too, and we see it almost as often as under-refreshing, particularly among teams that read one "kill it every 7 days" rule and applied it everywhere. Every new creative restarts Meta's learning phase for that ad, during which the algorithm is still figuring out who to show it to and typically delivers a worse CPA for the first 3-7 days regardless of creative quality. If you're swapping creative before the previous one has even exited learning phase, you're paying the learning-phase tax repeatedly without ever letting a winning creative run long enough to compound.

The category cadences above are starting points, not hard rules — the actual trigger should always be the frequency/CTR signal for that specific ad set, not a calendar date. A creative that's still holding CTR at day 20 in a broad audience shouldn't be killed just because "it's been three weeks." Plans on AdsSarthi from Starter through Growth all include the fatigue-signal tracking needed to make that call on data rather than a fixed schedule.

A simple weekly check that catches most fatigue early

You don't need a sophisticated system to get most of the benefit here. A five-minute weekly review, done consistently, catches the majority of fatigue cases before they become CPA emergencies:

1. Sort active ad sets by 7-day frequency, descending.
2. Flag anything above 2.8 for broad audiences or above 2.2 for retargeting audiences.
3. For flagged ad sets, compare this week's CTR to the creative's own first-week CTR.
4. If CTR is down 15%+ and you don't already have a variant queued, brief one this week — not after CPA breaks.

The teams that build this into a weekly habit — whether manually or through an automated flag — consistently run flatter CPA curves than teams that only react once ROAS visibly drops. Fatigue is one of the few problems in performance marketing that's almost entirely preventable once you know what to watch for.