AI ads · Scaling

AI Ads for Growing Brands: Getting Started at ₹10 Lakh+ a Month

At ₹10 lakh a month, small leaks become expensive. The brands that grow with AI ads at this level fix their structure, data and review rhythm before they add more budget.

Ajay Kumar • •7 min read
Illustration of a budget label on a scaled ad account

Quick answer: how should a brand spending ₹10 lakh+ a month start with AI ads?

Size each campaign so it can reach about 50 conversions a week, which usually means fewer, larger campaigns. Send purchases from browser and server with values, feed the AI several new creative angles every week, track cost per new customer separately, and set clear rules for who can change budgets and when.

Why ₹10 lakh a month changes the rules

Below about ₹2 lakh a month, most brands are still finding a product and an offer that sells. Above ₹10 lakh, the job changes: the offer works, and the risk is wasting a large budget through poor structure, weak data and too many hands on the account. For the overall operating plan, start with our complete guide to AI ads.

Small leaks become large

A tracking gap that hides 10% of purchases is an annoyance at ₹1 lakh a month. At ₹10 lakh it means the AI is optimising roughly ₹1 lakh of monthly spend on wrong information.

More people touch the account

Founders, an in-house marketer, an agency and a creative team may all make changes. Each edit can restart learning. At this level, process matters as much as settings.

Structure your campaigns from the learning-phase maths

Meta says an ad set needs about 50 results in the week after its last significant edit to leave learning. You can turn that into a simple budget rule.

Worked example: a supplements brand at ₹12 lakh a month

The inputs

Cost per purchase on Meta is about ₹700. The brand spends ₹9 lakh a month on Meta and ₹3 lakh on Google.

How many ad sets can learn

Each ad set needs about 50 purchases a week: 50 × ₹700 = ₹35,000 a week, or about ₹1.5 lakh a month.

Ad sets your budget supports = monthly budget ÷ (target conversions per week × cost per conversion × 4.3)

₹9,00,000 ÷ (50 × ₹700 × 4.3) = ₹9,00,000 ÷ ₹1,50,500 ≈ 6 ad sets.

If the brand runs 15 ad sets, most of them never leave learning. Merging into about six gives each one enough data to learn.

A typical budget split

The split below is a starting point for a store with a proven offer. Adjust it after four weeks of new-customer data.

CampaignShare of budgetMain job
Meta broad prospecting40 to 50%Find new buyers at scale
Meta sales, existing customers capped15 to 20%Clean read on new demand
Meta retargeting or catalogue10 to 15%Recover product viewers and carts
Google Search and Performance Max20 to 30%Capture people already searching

A typical structure at this level

  1. One broad prospecting campaign with Advantage+ audience, holding most of the new-customer budget.
  2. One sales campaign with existing customers excluded or capped, so you can see real new demand.
  3. One retargeting or catalogue campaign for people who viewed products or added to cart.
  4. On Google, Search for brand and top products, plus Performance Max once the product feed and tracking are clean.

Fix the data before adding budget

At this spend level, the data layer pays for itself fastest.

Server-side events

Meta recommends sending events from both the browser and your server with deduplication, so each purchase counts once and fewer are lost to browsers and ad blockers.

Enhanced conversions on Google

Enhanced conversions send hashed first-party data such as email with each conversion, so Google can match more of them to ad clicks.

Values, not just counts

Send the real order value, and if you can, a value closer to margin. A ₹3,000 order with 20% margin should not teach the AI the same lesson as a ₹3,000 order with 60% margin.

Monthly data check

Compare purchases in Meta and Google with completed orders in your store. A gap of more than 15 to 20% is worth investigating. Our guide on what AI needs from your pixel, CAPI and store has the full checklist.

Creative volume becomes the growth lever

Once structure and data are right, the AI's results depend mostly on what it has to show. At ₹10 lakh a month, ads tire faster because they reach the same people more often.

Plan angles weekly

Set a weekly target, for example three to five new ads covering at least two new angles: a new problem, a new proof point, a creator video, a comparison or a new offer.

Retire on evidence

Pause creatives when cost per purchase rises and frequency climbs for two weeks in a row, not on a single bad day.

Who makes the creative

Most brands at this level combine an in-house designer, a few creators and AI tools for variations and resizing. Keep a human check on every claim, price and product detail before an ad goes live.

Keep a creative log

Record each ad's angle, format and result. After three months you will know which angles work for your brand, which makes every new brief better.

Measure new customers, not just ROAS

Automated campaigns are good at finding people likely to buy, including existing customers who would have bought anyway. At high spend, that can hide slowing growth behind a healthy ROAS. Our comparison of AI ads and manual campaigns shows how much this can matter.

Three numbers for the weekly review

Use the platforms' new-customer settings

Google reports that advertisers using New Customer Value mode in Performance Max improved their new customer ratio by 5% and cut new-customer acquisition cost by 7% on average, according to its internal studies. On Meta, use existing-customer budget caps or exclusions in prospecting campaigns.

Set rules for who changes what

Write a one-page operating rule and share it with everyone who can touch the account.

  1. One owner for budget changes, with changes made on fixed days.
  2. Limit budget moves to about 15 to 20% per campaign per week unless something is clearly broken.
  3. New creative goes into existing campaigns, not new ones, to avoid restarting learning.
  4. Every change goes in a log with the date, the person and the reason.
  5. A weekly review looks at cost per new customer and profit, then decides next week's one main change.

For a full strategy template, see our guide to building an AI advertising plan for a D2C brand.

What most scaling advice tends to get wrong

"Just double the budget on winners"

Large jumps restart learning and often push the AI into weaker audiences. Scale in steps and watch cost per new customer.

"More campaigns means more control"

At this spend, more campaigns usually means thinner data and more ad sets stuck in learning.

"Automation removes the need for a team"

It removes manual bidding work. It adds work on data, creative and review. Brands that scale well put people on those three jobs.

Disclosure

ScaleWin, our product, is an AI ad manager for Meta and Google Ads. The structure and rules above apply whichever tools you use.

Key takeaways

  • Work out how many campaigns your budget can really support from your cost per purchase and the 50-conversions-a-week guide.
  • At this spend, tracking gaps cost lakhs. Server-side events and enhanced conversions come first.
  • Creative volume becomes the main growth lever: plan new angles every week.
  • Track new customers separately, because automation also finds people who would have bought anyway.
  • Write down who can change budgets and how much, so the AI is not disturbed by daily edits.

Ai ads: frequently asked questions

How many campaigns should a brand spending 10 lakh a month run?

Work it out from your cost per purchase. Each ad set needs about 50 conversions a week to leave learning. Divide monthly budget by 50 times cost per conversion times 4.3 weeks. For many brands at this level the answer is about five to eight ad sets across Meta.

What should a high-spend brand fix first?

Tracking. Send purchases from browser and server with deduplication on Meta, set up enhanced conversions on Google, and pass real order values. At this spend, a 10% tracking gap means a large share of the budget is optimised on wrong data.

How often should we add new creatives?

Weekly. A practical target is three to five new ads covering at least two new angles each week, added to existing campaigns so learning is not restarted. Retire ads when cost per purchase and frequency rise together for two weeks.

Why track new customers separately?

Automated campaigns also reach existing customers who might buy anyway, which can make ROAS look healthy while growth slows. Cost per new customer and new customer share show whether ads are actually bringing in new buyers.

How fast should budgets be increased?

In steps of about 15 to 20% per campaign per week, made by one owner on fixed days. Larger jumps can restart learning and push delivery into weaker audiences, which raises cost per new customer.

Sources

Ajay Kumar
Founder, ScaleWin · 12+ years in performance marketing

Ajay is the founder of ScaleWin. He works hands-on with Meta Ads, Google Ads and SEO for Indian D2C and service brands, and built ScaleWin so that the daily work of a performance marketer - reading numbers, pausing losers, scaling winners - happens automatically.

Drafted with AI assistance. How these articles are written

Let ScaleWin read your ad data every day

ScaleWin checks your Meta and Google Ads daily, judges every ad on profit and tells you exactly what to change.

Start free, no card

Get marketing tips that grow your business

One practical email a week on ads, SEO and growth. No spam, no fluff. Unsubscribe anytime.