Benchmark · Meta Conversions API
We compared $1M in ad spend with and without CAPI
We split 59 brands running Meta ads on BooleanMaths into two groups: 24 with Meta's Conversions API (CAPI) turned on and 35 without. Each group spent about $1M (~₹9 Cr). Brands without CAPI looked better in Ads Manager. Their real ROAS was worse.
1.82x
true ROAS with CAPI, vs 1.59x without
1.42x
orders Meta reports per real order with CAPI, vs 1.95x without
₹931
cost per real order with CAPI, vs ₹1,004 without
2x
click-to-order conversion rate, 0.84% vs 0.43% without
In short
With roughly the same spend, brands with CAPI earned more real revenue per rupee, paid less for each real order, and saw Meta over-report far less. Their clicks cost more, but those clicks converted at twice the rate, because Meta could finally see who was buying. Brands without CAPI saw a higher ROAS in Ads Manager (3.04x vs 2.78x) while their true ROAS was lower (1.59x vs 1.82x). Budget decisions made inside Ads Manager were being steered by orders that didn't happen.
The question
Does CAPI change real results, or just the dashboard?
Meta's browser pixel misses purchases. iOS privacy settings, ad blockers, slow mobile networks and third-party checkouts like GoKwik and Shopflo all break the trail between an ad click and an order. Meta then fills the gaps with modelled conversions, which is why Ads Manager often claims more orders than a brand actually made.
Most brands judge CAPI by what Ads Manager shows after they switch it on. That is the wrong yardstick, because Ads Manager is the thing being fixed. So we measured both groups against BooleanMaths true attribution instead, and compared what actually happened to real orders, real revenue and real costs.
The setup
Benchmark setup
59 D2C and e-commerce brands in India running Meta ads, all tracked in BooleanMaths. We grouped them by whether Meta's Conversions API was enabled and compared each group's total spend, true orders and true revenue against what Meta reported. Three brands without a Meta pixel or ad account were left out.
24 brands with CAPI
₹9.08 Cr in Meta spend, about $1M. Server-side conversion events sent to Meta through BooleanMaths, alongside the pixel.
35 brands without CAPI
₹9.73 Cr in Meta spend, about $1M. Meta sees only what the browser pixel captures.
True = BooleanMaths
Results - at a glance
Same spend. Different outcomes.
On every metric that matters, brands with CAPI came out ahead: 14% higher true ROAS, 7% cheaper real orders, 21% higher new-customer ROAS, and nearly twice the click-to-order conversion rate. Their average order value was 6% higher too, and Meta over-reported their orders by 42% instead of 95%.
Results - the ROAS trap
Looked better in Ads Manager. Did worse in reality.
Brands without CAPI saw a higher ROAS in Ads Manager, 3.04x against 2.78x for brands with CAPI. Measured against real orders, the order flipped: 1.59x without CAPI, 1.82x with it. Without CAPI, Meta's reported ROAS was 48% above reality. With CAPI, the gap narrowed to 35%. A brand scaling budgets off the Ads Manager number would have been scaling the weaker setup.
Meta-reported ROAS vs BooleanMaths true ROAS, by group. Same spend in both groups, about $1M each.
Results - deep dive
Where the difference comes from.
Three views of the same data: how much Meta over-reports, why pricier clicks still produce cheaper orders, and how individual brands spread out within each group.
Meta over-reported by nearly 2x without CAPI
Platform-reported orders ÷ BooleanMaths true orders
Without CAPI, Meta claimed 1.95 orders for every order that actually happened: 1.89 lakh reported against 96,870 real. With CAPI, it claimed 1.42. On revenue, the inflation was 91% without CAPI and 53% with it. When Meta can see real conversions, it has to model less, and the dashboard moves closer to the truth.
Pricier clicks. Cheaper orders.
Cost per click → click-to-order rate → cost per real order
Brands with CAPI paid 82% more per click (₹7.82 vs ₹4.30) and 38% more per thousand impressions (₹108 vs ₹78). That looks worse until you follow the click through: their clicks converted into real orders at 0.84%, nearly twice the 0.43% without CAPI. The net result was a cheaper real order, ₹931 against ₹1,004. When Meta can see every purchase, it stops chasing cheap clicks and starts finding people who buy.
