How Many Creatives to Test on Meta, Google and YouTube in India

Short answer: Split production roughly 50–60% Meta, 20–25% YouTube, 15–20% Google, 10–15% Amazon and quick commerce — and reuse the idea across platforms, never the file.

Each platform tests a different thing#

The mistake is treating "creative testing" as one activity you do in three places.

Meta tests concepts in an auction — different angles competing for the same attention. Google Performance Max tests combinations of text, images, video and feed assets, assembled by the system. YouTube tests narratives — whether the story holds long enough to matter.

Those are three different questions, so they need three different counts, three different timelines, and three different definitions of what a "variant" even is. Running the same monthly number across all three is how brands end up over-producing for Google and under-producing for Meta.

How many creatives across all platforms?#

That is total digital spend across every channel. For the Meta-only number, run testing budget ÷ blended cost per creative on your Meta budget alone.

A creative means a meaningfully different idea — new problem, promise, proof point, use case, creator, hook or offer. Changing the background colour is not a new creative.

Deriving the number from budget#

Monthly test capacity = testing budget ÷ minimum spend per creative

Monthly Meta budget:     ₹6,00,000
Testing allocation 20%:  ₹1,20,000
Test budget per creative: ₹5,000  (5× a ₹1,000 target CAC)
₹1,20,000 ÷ ₹5,000 = 24 theoretical tests

Twenty-four is the mathematical ceiling, not the plan. A reliable version: launch 12–15 new concepts, promote 5–8 winners, keep the rest of the budget for retesting and follow-up variations.

The divisor matters more than the ceiling. Set it from the test budget of 5–10× target CPA — ₹5,000–₹10,000 at a ₹1,000 target CAC — not from what you can afford to slice.

Meta#

Test one variable at a time: problem angle, product demonstration, testimonial or UGC, founder story, offer or urgency, Hindi/English/Hinglish execution, and format.

3–5 or 8–15 creatives per ad set?#

Both, depending on daily spend — the per-ad-set count by daily budget is the table to work from. What decides which end of the range you sit at:

Use 3–5 when the brand is new, spend is limited, CAC is high, conversion data is thin, or you need a winning angle quickly.

Use 8–15 when the campaign can distribute impressions across the portfolio, the creatives are genuinely different, you sell to several use cases, purchase volume is stable, or you are managing fatigue at scale.

The governing rule:

Daily budget per creative ≥ 1 × target CAC

Target CAC ₹800, five creatives:
5 × ₹800 = ₹4,000/day

At ₹1,000 a day, fifteen creatives will dilute delivery into nothing. Meta will hand most impressions to one or two ads and the rest stay statistically untested.

A useful six-creative batch: problem–solution UGC, founder explanation, customer testimonial, product demonstration, offer-led video, and a static or carousel built on proof, ingredients or comparison. Far better than six captions on one video.

Advantage+ Shopping#

Maintain 10–20 active creative variations for a mid-sized Indian account, 15–30 for high spend. Meta's own diversification guidance recommends at least 20 diversified ads — with the emphasis on diversified.

A sensible portfolio mix:

• 40–50% proven winners
• 30–40% fresh variations of winning angles
• 10–20% experimental concepts

Do not use Advantage+ as your primary testing environment. Test manually first, then transfer winners and strong variations in. And never swap the whole portfolio at once — you lose the ability to tell whether a change came from creative, delivery learning, seasonality or the offer.

Separate ad sets or one broad campaign?#

One broad campaign with consolidated ad sets for most purchase campaigns. Meta needs conversion data and budget concentration to optimise.

Scaling campaign: one broad prospecting campaign, one broad ad set where possible, 5–15 diversified creatives, Advantage+ placements, existing winners plus controlled variations.

Testing campaign: separate, one broad ad set or a few defined cells, 3–8 creatives per batch, 20–30% of Meta budget, winners promoted after validation.

Split into separate ad sets only for a real business difference: different language markets, different regions, prospecting versus retargeting, different products or price points, or a controlled experiment.

₹10,000/day across 10 ad sets = ₹1,000 each
— before auction variation, learning allocation and audience overlap

₹10,000/day × 20% = ₹2,000 testing budget
→ three to five test creatives, not ten fragmented ad sets

Google and YouTube#

YouTube Shorts cadence#

Start at 3 Shorts a week for the first 30 days. Move to 5–7 only when quality, consistency and variation all hold.

A five-Short week:

Seven weak videos lose to three strong, locally relevant ones. And five Shorts showing the same product shot with different captions are five pieces of content but one creative idea.

Paid YouTube: start with 3–5 distinct concepts, two opening hooks per winning concept, and different lengths — 10–15s and 20–30s. Vertical for Shorts placements, horizontal for standard.

3 concepts × 2 hooks = 6 video variants

Performance Max asset requirements#

One asset group:

Google's own recommendation is roughly 17 text assets, 10 image assets and three videos. These are asset variations within one group, not separate campaigns.

Create separate asset groups only for substantially different themes, products or audiences. "Summer skincare", "acne solution" and "sunscreen for outdoor workers" justify separate groups. Minor copy changes do not.

Example concepts for an Indian skincare brand: "Reduce visible acne marks" · "Dermatologist-tested daily routine" · "Lightweight formula for Indian weather" · "Before-and-after customer proof" · "Starter kit under ₹999."

YouTube test duration#

Run 7–14 days before a major optimisation decision. Use the full 14 on low-budget Indian campaigns, where delivery is uneven. Controlled Google video experiments need longer.

Review hook retention, view-through rate, cost per completed view, CTR, add-to-cart rate, purchases and assisted-conversion signals. A creative can post a strong view rate and weak sales — optimise toward the objective, not cheap views.

Splitting volume across platforms#

Split production volume, which is not the same as splitting media spend.

If Meta drives 70% of sales, shift toward 60:20:10:10. If Amazon or quick commerce carries a large revenue share, raise its production allocation.

Reuse the idea, not the file#

One product demonstration becomes:

• Meta Reel — 9:16, fast hook, captions, social-proof CTA
• YouTube Short — 9:16, stronger narrative and retention pacing
• PMax video — 9:16, 1:1 and 16:9 with clearer product branding
• Amazon video — product visible early, concise benefit, shopping-oriented end card
• Quick-commerce banner — product, price, pack size, delivery-led message

Keep the core proof point where it works. Adapt aspect ratio, the first three seconds, language and subtitles, CTA, price or offer, pack size, landing destination and platform restrictions.

FAQ#

How many creatives should be in one Meta ad set in India? Three to five below ₹3 lakh of monthly spend, 5–8 at ₹3–10 lakh, and 6–10 per ad set above ₹10 lakh. Keep daily budget per creative at 0.5–1× target CAC.

How many ads should run in an Advantage+ Shopping campaign? Ten to twenty active variations for a mid-sized Indian account, 15–30 for high spend. Meta recommends at least 20 diversified ads. Test manually first and promote winners in rather than testing inside Advantage+.

Should each creative get its own ad set? No. Use one broad campaign with consolidated ad sets. Ten ad sets on ₹10,000 a day gives each ₹1,000, which fragments delivery. Separate only for different languages, regions, products, price points, or prospecting versus retargeting.

How many assets does a Performance Max asset group need in India? Roughly 11–15 short headlines, 3–5 long headlines, 4–5 descriptions, 6–8 landscape and square images, 3–5 portrait images, 3–5 videos and 2–3 logos. Google's guidance is about 17 text, 10 image and 3 video assets.

How many YouTube Shorts should an Indian D2C brand post weekly? Three a week for the first 30 days, rising to 5–7 only if quality holds. Five Shorts of the same product shot with different captions count as one creative idea, not five.

How long should a YouTube creative test run? Seven to fourteen days before a major decision, using the full fourteen for low-budget Indian campaigns where delivery is uneven.

How should creative volume be split across platforms? By production, not spend: 50–60% Meta, 20–25% YouTube, 15–20% Google Search and PMax, 10–15% Amazon and quick commerce. Adjust toward whichever channel carries your revenue.

Can the same creative be reused on Meta, YouTube, Google and Amazon? Reuse the idea, not the asset. Keep the core proof point and adapt aspect ratio, opening three seconds, language, CTA, pricing and landing destination for each platform.

---

About the author#

Rohit Sharma is a performance marketer at Profitbox360, specializing in paid media, D2C growth, creative strategy, and conversion optimization. They write about practical marketing frameworks, campaign benchmarks, and strategies for scaling customer acquisition.