meta ads strategy 2026: the india playbook for leads and sales

puneetraj chavanfounder, rapchik media17 oct 2026 · 11 min read
a performance marketer with arms crossed in front of a curved wall of phones showing many different ads

meta's andromeda system now picks who sees your ad mostly from what the ad says and shows. your creative is your targeting, and the signal you send back decides who meta finds next.

in this article11 sections
  1. 01what did the meta andromeda update actually change?
  2. 02should you use advantage+ campaigns and broad targeting?
  3. 03your creative is your audience: what "diversity" really means
  4. 04facebook lead generation ads: optimise for buyers, not form fills
  5. 05when do click to whatsapp ads beat forms?
  6. 06facebook ads for real estate in india: what changes
  7. 07how much do facebook ads cost in india in 2026?
  8. 08meta ads strategy 2026: a worked budget for a ₹3 lakh month
  9. 09facebook ads not working? check these five things first
  10. 10meta ads in 2026: follow-up questions we hear most
  11. 11who wrote this, how and why
  • andromeda is meta's ad retrieval system, the first step that narrows millions of ads to a few thousand for each person. meta reported +6% recall and +8% ads quality on selected segments when it launched in december 2024 (meta engineering).
  • meta said its advantage+ products passed a $75 billion annual revenue run rate in q2 2026, as reported in nasdaq's earnings call summary.
  • meta credited its newer ranking models with 8.3% more ad clicks and 15.7% more conversions on facebook in the same quarter. these are meta's own figures, as reported by ppc land.
  • optimising for conversion leads with your crm data works with instant forms, so meta can learn from sales, not form fills (meta for developers).
  • for marut drones, we turned ₹25 lakh into ₹8.7 crore with state-specific creative, not narrow audiences.

what did the meta andromeda update actually change?

the andromeda update replaced meta's old rule-based ad retrieval with one large machine-learning model. it reads each ad and each person together, then picks the few thousand ads worth ranking. in practice, the ad's content now tells meta who it is for. your audience settings are suggestions. your creative does the targeting.

meta's engineering team describes andromeda as the system that selects candidates from tens of millions of ads before the ranking models choose the winner (meta engineering, dec 2024). it replaced separate stages and hand-written rules. it was built for the far larger number of ad variations that generative tools create.

ranking moved the same way. on its q2 2026 earnings call, meta introduced the meta generative recommender. according to vectorshift's call summary, meta described it as weighing ad content and user preferences together, rather than scoring each ad on its own.

so here is the shift in one line. you used to tell meta who to show an ad to. now the ad tells meta. take an illustrative example. a video of a farmer speaking marathi next to a spray drone finds farmers in maharashtra. a slick 3d walkthrough with a ₹3 crore price tag finds a different person. same targeting box, different audience.

should you use advantage+ campaigns and broad targeting?

yes, for most indian lead and sales accounts in 2026. advantage+ campaigns and broad targeting give andromeda the room it needs to match each creative to its people. keep tight targeting only for hard limits: the cities you serve, legal age, or a language you can't support. everything else, let the creative decide.

a simple 2026 structure for a lead business:

layerwhat to runwhy
campaigns1 prospecting, 1 retargeting (optional)fewer campaigns pool data so learning finishes faster
ad sets1 per hard limit (for example, per state or per project)split only where the offer or the law truly differs
targetinglocation and age floor, advantage+ audience onthe creative carries the audience signal
adsseveral genuinely different concepts per ad setdifferent concepts reach different people. variants of one idea don't

you will also see the claim that "one ad set with 25 creatives beats five ad sets with five each". that comes from practitioners. meta has not published it, so treat it as a hypothesis to test in your own account.

meta keeps pushing in this direction. at advertising week new york in october 2026 it made advantage+ creative's static-to-video conversion generally available and expanded ads creative studio (meta newsroom, es; social media today). use these tools for more versions, but of different ideas.

your creative is your audience: what "diversity" really means

creative diversity means different people, problems, formats and hooks, not ten colours of one banner. ads that look and say the same thing tend to reach the same people, so near-copies compete with each other. build each new ad around a different buyer, reason to buy, or objection, and you widen the audience without touching targeting.

when we brief a set of concepts, each one has to change at least one of these:

  • who is on screen. a first-time buyer and an investor are different audiences.
  • the problem. "labour shortage at spraying time" and "spray drone subsidy" pull different farmers.
  • the format. talking head, demo, testimonial, static offer, carousel.
  • the language. telugu, hindi and english creative reach different people, even with identical targeting.

this is what worked for marut drones. across 13 states we ran state-specific creative and forms. ₹25 lakh of spend produced 23,420 leads and ₹8.7 crore in revenue, a 36x roas (read the marut drones case study). the targeting stayed broad. the creative did the sorting.

for how we build concepts that are truly different, see our ad creative strategy guide.

facebook lead generation ads: optimise for buyers, not form fills

facebook lead generation ads in 2026 should optimise for qualified leads, not raw form fills. if you only report "lead submitted", meta finds more people who submit forms, including the ones who never answer the phone. send back which leads qualified or bought, through conversion leads and the conversions api, and meta looks for more of those.

this is the second lever andromeda leaves you. creative tells meta who might care. your feedback tells meta who actually bought. meta's crm integration for the conversions api supports instant form leads. you can mark each lead's stage, such as "qualified" or "site visit done", and optimise towards it (meta for developers).

speed matters as much as the signal. in hbr's study of 2,241 us firms, those that contacted a lead within an hour were nearly 7x more likely to qualify it than those that waited longer (hbr, 2011). a slow call centre makes good leads look like junk. then you feed meta the wrong signal.

worked example: why the cheaper lead can cost you more

two campaigns, same ₹3 lakh budget. the numbers are an illustrative example, to show the method.

metriccampaign a: optimised for form fillscampaign b: optimised for qualified leads
cost per lead₹300₹600
leads1,000500
share that qualify10%35%
qualified leads100175
cost per qualified lead₹3,000about ₹1,714

campaign a looks better in ads manager. campaign b gives your sales team 75% more people worth calling, at about 43% lower cost each. if you judge on cost per lead, you switch off the better campaign.

when do click to whatsapp ads beat forms?

click to whatsapp ads beat forms when your buyer wants to ask a question before sharing details, and when you can reply within minutes. they open a conversation, not a record, so qualification happens in chat. meta reports one indian brand cut cost per lead 12% by optimising these ads for leads. they fail when nobody answers.

that figure comes from meta's own nutryfit case study. nutryfit ran click to whatsapp ads optimised for leads and got a 12% lower cost per lead than ads optimised for conversations, in a meta a/b test (whatsapp business). it is one meta-reported test, not a benchmark.

the catch is staffing. a chat that waits two hours is worse than a form, because the buyer expected a reply. our whatsapp business api guide for india covers setup and fast first replies.

facebook ads for real estate in india: what changes

facebook ads for real estate in india follow the same rules, with two additions. the quality signal matters more, because a site visit is worth far more than a form fill. and the law is stricter: under rera, every ad must show the project's registration number and the regulator's website.

section 11(2) of the rera act requires the registration number and the authority's website on every advertisement. state regulators enforce it, for example in andhra pradesh rera's advertisement guidelines and a haryana rera order that quotes the section. closer to home, tg-rera fined gnr infra ₹2.14 crore in 2026 for promoting an unregistered project (deccan chronicle). rules vary by state. this is general information, not legal advice.

for lodha, digital did the selling: 45 flats sold, attributed 100% to digital, at 3.8x the industry average conversion rate (see our work).

how much do facebook ads cost in india in 2026?

facebook ads cost in india depends on industry, city and season, and there is no official meta rate card. vendor-reported real estate cost per lead ranges from about ₹150 to ₹2,000. october to december is usually the most expensive window, because festive retail bids up the auction. budget on cost per qualified lead, not cpl.

that ₹150 to ₹2,000 range comes from a crm vendor, sell.do, and is vendor-reported (sell.do, 2026). the q4 cost rise is practitioner observation, not meta data. meta itself said its average price per ad rose 12% year on year in q2 2026. it linked this partly to better ad performance, as reported in vectorshift's call summary.

our rule: budget from the qualified leads sales can handle. then check the cost per qualified lead still leaves a profit.

meta ads strategy 2026: a worked budget for a ₹3 lakh month

a sound meta ads strategy for 2026 puts most of the budget in one broad prospecting ad set, keeps a fixed slice for new creative concepts, and only then retargets. size it so each ad set can reach about 50 optimisation events a week. below that, meta keeps learning and results swing.

meta's developer guidance says to structure ad sets for at least 50 events over 7 days (meta for developers). that decides how many ad sets you can afford. an illustrative example, for a lead business spending ₹3 lakh a month:

budget linesharemonthly spendwhat it does
broad prospecting, advantage+ audience on70%₹2,10,000one ad set, 6 to 10 different concepts
new concept testing15%₹45,0003 or 4 new concepts a fortnight
retargeting (video viewers, form openers)15%₹45,000answers objections from warm prospects

prospecting spends about ₹49,000 a week. at the ₹600 cost per lead from the worked example above, that is roughly 80 leads, past 50. optimise it for qualified leads at about ₹1,714 each and you get about 29, so it stays in learning. pick an earlier stage that clears 50, such as "contacted", and go deeper as volume grows.

split the same budget across three state ad sets and each gets about 27 leads a week. none exits learning.

facebook ads not working? check these five things first

when facebook ads are not working in 2026, the cause is rarely the audience. check creative diversity, the event you optimise for, how fast leads get a reply, whether your account is split too thin, and whether tracking reaches the crm. fix them in that order before touching targeting.

symptomlikely cause in the andromeda erafirst fix
spend stuck on one adother ads are near-copiesadd 3 or more new concepts with a different person or problem
cheap leads, no salesoptimising for form fillssend qualified stages back via conversion leads
good leads go coldslow first replyauto-reply on whatsapp, call within the hour
learning never endstoo many ad sets splitting budgetmerge ad sets without a hard reason to split
crm and ads manager disagreeno server-side or crm eventsset up the conversions api with lead ids

meta ads in 2026: follow-up questions we hear most

the questions that come up after a strategy call are practical: how many creatives a smaller account needs, whether to send people to instant forms or a landing page, and how long to wait before judging a new campaign. here are short answers, with the longer guides linked where we have them.

how many creatives does a smaller account need?

fewer than people think, but more different. our rule of thumb on ₹1 lakh a month is four to six live concepts, since each needs enough spend for meta to learn who it suits. ten near-copies of one idea reach fewer people than four real concepts.

instant forms or a landing page?

instant forms win on volume: the form opens inside facebook or instagram, pre-filled. a landing page wins when the buyer needs to read before deciding, such as a ₹2 crore flat, and it gives you tracking you control. test both on one offer and compare cost per qualified lead. our guide to a landing page for lead generation covers what that page needs.

how long before i judge a new campaign?

a full week, or about 50 optimisation events, whichever comes later. for cost per qualified lead, add the time sales takes to call the first batch.

who wrote this, how and why

puneetraj chavan wrote this. he is the founder of rapchik media and has spent 8+ years running performance marketing for 30+ brands across india and the us. sources were checked on 18 october 2026. platform-reported and vendor-reported figures were not independently audited. client results come from rapchik case studies.

the worked example uses illustrative numbers. we wrote this because most 2026 meta advice repeats tactics, and explaining the mechanism lets you judge new tactics yourself.

want this run for you? see our performance marketing services or talk to us.

want this on your brand?talk to puneetraj on whatsapp reply within 24h →

want this on your brand?

tell us your numbers. we'll tell you exactly where the money's leaking, free, no obligation.