From ₹106.66 to ₹32.29 CPL: Scaling Meta Lead Generation in Patna
77 → 1,540 leads across observed campaign cohorts, while reported CPL fell 69.7%.
₹106.66
2025 cohort CPL
₹32.29
2026 cohort CPL
77 → 1,540
recorded leads
69.7%
reported CPL reduction
Context
- Industry
- Real Estate
- Platform
- Meta Ads
- Market
- Patna + surrounding Bihar cities
- Campaign focus
- Residential real estate lead generation
How to read this comparison
The cohorts below are grouped by campaign start year and include only campaigns that recorded results.
This is an observed comparison, not a controlled before/after experiment. Spend levels, projects, creative and platform delivery all changed between the two cohorts, so the difference in cost per lead cannot be attributed to any single decision.
The cohort comparison
Cost per lead
2025 cohort
2026 cohort
Recorded leads
2025 cohort
2026 cohort
- 2025 cohort
- ₹8,212.69 spend · 77 leads · ₹106.66 CPL
- 2026 cohort
- ₹49,722.95 spend · 1,540 leads · ₹32.29 CPL
- Spend increased approximately 6×.
- Lead volume increased approximately 20×.
- Reported CPL fell 69.7% across the observed cohorts.
Note the wording. Reported CPL fell across the cohorts — the exports do not isolate a cause, and attributing the fall to a single decision would be a claim this data cannot carry.
What Meta was actually optimising for
These four result types are not interchangeable. A messaging conversation, a lead form submission, a quality-optimised lead and a native call are different events with different downstream value, so comparing their cost per result only means something once you have decided which event actually matters to the business.
That decision is a business question rather than a platform question — and it is the one worth settling before any of these numbers is used to move budget.
| Result type | Cost per result |
|---|---|
| Messaging conversations | ₹31.33 |
| Standard lead generation | ₹39.12 |
| Meta quality-lead optimisation | ₹50.22 |
| Native calls | ₹157.48 |
Not independently verifiedThe ₹50.22 figure is Meta’s cost per quality-optimised result, not a verified human-qualified lead cost.
Five campaigns that carried the volume
Five campaigns account for 1,170 of the recorded leads. The spread between the cheapest and the most expensive is close to 1.8×, which is the more useful number: it says the difference sits in the campaign, not in the platform.
| Campaign | Recorded leads | CPL |
|---|---|---|
| SCE Gayatri | 296 | ₹19.37 |
| Nutan Construction Project 4 | 341 | ₹21.87 |
| South City Centre 2 | 71 | ₹29.62 |
| Balaji Leads | 181 | ₹30.91 |
| Garden Leads | 281 | ₹34.84 |
| Combined — ₹30,678.43 spend | 1,170 | ₹26.22 blended |
Variant observations
Three campaigns ran alongside a variant. The exports record what each one spent and produced; they do not record why the variant was created.
Aqua City
Original
₹2,519 · 34 leads
₹74.09 CPL
Variant
₹7,180 · 190 leads
₹37.79 CPL
Urmila
Original
₹419 · 4 leads
₹104.66 CPL
Variant
₹733 · 12 leads
₹61.09 CPL
South City Centre
Original
₹2,103 · 71 leads
₹29.62 CPL
Variant
₹105 · 0 results
—
The exports show performance differences between campaign variants, but they do not establish the original decision rationale. Nothing here should be read as a claim that a particular creative or copy change caused a particular outcome.
The conservative blended number
1,617
Meta form submissions
₹40.77
conservative blended CPL
₹40.77 divides all lead-objective spend by Meta form submissions, and it includes ₹7,990.31 of lead campaigns that recorded no results.
A more flattering figure is available by excluding that spend. It is not used here. Spend that produced no recorded result is still spend, and a cost-per-lead number that quietly removes it is not a number worth planning a budget against.
Operational scale
What the account looked like in practice — the structure behind the numbers above.
- campaigns
48
campaigns
- ad sets
51
ad sets
- ads / creatives
83
ads / creatives
- Bihar cities
6
Bihar cities
- campaign objectives
3
campaign objectives
The exports record that Hindi and English creative variants existed. They do not record a language attribute for every creative, so 83 is a count of ads, not a count of bilingual creative pairs.
What this evidence does not establish
The available evidence covers platform-recorded activity. It does not confirm any of the following:
- Qualified lead rate
- Site visits
- Bookings
- Sales
- Revenue
- Lead-to-sale rate
- True cost per acquisition
- Offline attribution
Takeaway
Related
Service this evidences
Related case study
Cross-Channel Meta + Google AdsThe same real estate market, measured across two platforms.
Next case study
Auditing My Own Ad AccountsRead case study