SEA & India Growth Benchmarks: What's Real (2026)

We set out to build the benchmark hub we wanted to exist for Southeast Asia and India — CAC, CPC, conversion rate, retention, by market. We couldn't. Most of the numbers ranking on page one for these queries do not survive verification, and the region's most-cited report presents no acquisition data at all — its publishers have never claimed it does.

So this page is the audit instead. What's real, what the methodology actually says, and where the data simply does not exist — including the finding that matters most: Indonesia, the region's largest digital market, has no performance benchmark data with a disclosed methodology. None.

If you want the short answer, skip ahead: the availability matrix two sections down is the usable map. The rejections immediately below explain why it has so many holes.

Key Takeaways

• Google/Temasek/Bain's e-Conomy SEA — routinely cited for regional CAC — presents no acquisition-cost data anywhere in the materials all three publishers put out. Its publishers don't claim it does.
• Indonesia is 37% of SEA platform GMV and has no verified CAC, CPC, conversion rate, or retention benchmark. Every page we found claiming otherwise was unsourced — one discloses chatbots as a validating source.
• One "Indonesia Digital Marketing Benchmark 2026" report discloses ChatGPT, Gemini and Perplexity as data sources it cross-validated against. That is not a methodology.
• dentsu forecasts global 2026 ad spend growth at +5.0%. WPP Media says +8.9%. Same year, published three weeks apart, neither discloses a methodology.

What We Threw Away

Leading with the rejections, because the rejections are the finding. Every source below ranks for the benchmark queries this page targets. None of them survived.

The AI-laundered one. A widely-ranking "Indonesia Digital Marketing Benchmark 2026" states in its own methodology that it "integrates research from ChatGPT, Gemini, and Perplexity" alongside a client survey, and that its data points were "cross-validated" against those chatbots. Language models are not independent sources; cross-validating one against another validates nothing. Some of its own figures are self-marked "LOW confidence — directional only" and then tabulated as benchmarks anyway. It's the cleanest specimen of the genre we found.

The one whose report doesn't load. A "State of Indian D2C 2026" report ranks well while hosted on a CMS subdomain that returns a DNS failure. Its headline Meta CAC figures are contradicted by the same company's own blog. Its disclosed sources are "Reddit threads, founder communities, and industry reports."

The relabelled one. A "Digital Marketing Benchmarks Singapore 2026" page aggregates — by its own admission — from WordStream and others, without links. WordStream's benchmark report is explicitly built on 13,474 US-based campaigns. US data, wearing a Singapore label.

The one that contradicts verified data. A "Meta Ads CPC by Country 2026" page lists India at $0.20. Our own figure puts India's median at $0.11 — an 80% divergence. Ours is a 13-month median spanning July 2025 to July 2026, from Superads' aggregated ad-account data, with a range of $0.025 to $0.21. The page we're rejecting discloses this, in full: "projections for 2026 based on industry benchmarks and late 2025 data."

The one that's six years stale. The page ranking #1–2 for "CAC benchmark Southeast Asia" sources its figures from AppsFlyer 2020 and Liftoff 2019. Its $129 fashion CAC traces to a Shopify blog post about global data. Its $173.75 gaming figure cites a PDF that now returns 403 — so we cannot confirm the number exists in the source at all. The article itself concedes its data "may very well be outdated, and not geographically relevant."

The one that invented a regional multiplier. A "CAC Benchmarks 2026" page asserts Southeast Asia "runs 40–60% lower" than North America, with no citation, on a page that openly states its figures are "primarily North American." That uncited multiplier is now propagating verbatim across other sites.

Which Benchmarks Actually Exist

Here's the honest map. This is what a marketer in this region actually has to work with.

Metric	Singapore	India	Indonesia
Market size / GMV	Verified	Verified	Verified
Audience reach	Verified	Verified	Verified (best-documented)
Ad spend totals	Partial	Verified	Stale forecast only
Meta CPC	Verified	Verified	Does not exist
Google Ads CPC	Does not exist	Does not exist	Does not exist
Conversion rate	Does not exist	Vendor proxy only	Does not exist
CAC by industry	Does not exist	Does not exist	Does not exist
App revenue per install	Regional only	Regional only	Regional only
Retention	Regional only	Vendor proxy only	Does not exist

"Verified" = named source, disclosed date, methodology stated. "Vendor proxy" = real data from a vendor benchmarking its own customer base. Compiled 2026-07-12.

Read the Indonesia column. It is the largest market in the region by platform GMV and it has essentially nothing on the cost and conversion side. That absence is the single most useful fact on this page — because every article we opened that fills the Indonesia column with numbers got them from somewhere it declines to name.

The Market-Size Numbers Are Real

Where the region is well-served is size. Two sources carry it.

Google/Temasek/Bain, e-Conomy SEA 2025 (published 11 November 2025, the 10th edition) puts the Southeast Asian digital economy past $300B GMV and $135B revenue, both growing 15% year over year — 7.4x GMV growth since 2016. Sector splits for 2025: ecommerce $185B GMV, travel $51B, media $34B, food delivery $23B, transport $11.5B. Singapore's GMV is $29B, up 7%.

Momentum Works, Ecommerce in Southeast Asia 2026 (14 April 2026, 4th edition) puts SEA platform ecommerce at $157.6B GMV in 2025, up 22.8%, with 98.8% held by Shopee, Lazada and TikTok Shop — Shopee alone at 53%. Content commerce reached $49.7B, or 32% of platform GMV, up from 20% in 2024. Indonesia is 37% of regional GMV.

The caveat that applies to both: neither discloses a methodology. e-Conomy SEA names its inputs — Temasek insights, Bain analysis, Google Trends, expert interviews — but states no sample size. Momentum Works publishes GMV estimates with no stated estimation method at all. These are the best size numbers available and they are still, strictly, unaudited. Cite them for scale; don't cite them for precision.

And here's what nobody does with the e-Conomy data. Everyone quotes the GMV headline. Almost nobody divides it by the revenue line — which is where the operator-relevant story is. Travel converts $24B of revenue from $51B GMV: it keeps 47%. Ecommerce keeps $41B of $185B: 22%. Food delivery keeps $2.4B of $23B: 10%. The sector with the largest GMV is a middling monetiser, and the one with the loudest growth narrative keeps a dime on the dollar. If you're choosing a category to build in, that ratio matters more than the GMV headline it's buried under. One caveat: a ratio is more robust than either figure alone, because errors common to both the GMV and revenue lines partly cancel. It still inherits e-Conomy's unaudited status. Use it to rank sectors, not to model one.

The gap you should know about before citing e-Conomy at all: across all three publishers' summary pages, contents listings and press materials, there is no CAC, ad spend, or marketing cost data — and none of them claim there is. We should be precise about our own method here: the full PDF exceeds our fetch limit, so we are reporting what the publishers themselves present, not a page-by-page read of the report. That's enough to make the point. If someone cites e-Conomy SEA for an acquisition-cost figure, they are citing it for something its own publishers don't say it contains.

Indonesia: The Largest Market With the Least Data

Indonesia is 37% of SEA platform GMV, roughly $99B in digital economy GMV, with ecommerce at $71B — about 72% of its digital economy. It has 230 million internet users, 80.5% of the population (DataReportal, Digital 2026: Indonesia, 5 November 2025).

That DataReportal figure deserves a note: it is the best-documented source in this entire compilation. It names its inputs — ITU, GSMA Intelligence, national statistics offices, platform planning tools, Ookla — dates its data to October 2025, and explicitly cautions that advertising reach is not the same as monthly active users. That's what a disclosed methodology looks like, and it's rare enough here to be worth pointing at.

Its ad-reach figures: TikTok 180M adults, YouTube 151M, Facebook 121M, Instagram 108M, LinkedIn 37M.

And then it stops. We could not find an Indonesia-specific CAC, CPC, conversion rate, AOV, or retention benchmark with a disclosed methodology. Not one. The country-level Meta CPC data we trust elsewhere excluded Indonesia because currency couldn't be confirmed. The only Indonesian ad-spend figure we found is an 18-month-old forecast reported secondhand. Of the pages that appear to carry Indonesian cost data, every one we opened either cites no source or — in one documented case — lists chatbots among the sources it validated against.

What to do instead: treat Indonesia as a market you must measure yourself. Run a deliberately small test budget, let your own CPC and conversion data accumulate for a quarter, and benchmark against your own history rather than a number from the internet. That's slower than copying a table. It's also the only defensible option, and it's what we do for clients here.

India: The Best-Documented of the Three

India is the one market where a marketer has real numbers to work with.

Ad market: dentsu and exchange4media's Digital Advertising Report 2026 puts India's ad industry at ₹1,21,339 crore for 2025, growing to ₹1,40,001 crore by 2027. Digital grew 19% in 2025 to ₹71,621 crore and is forecast to reach roughly 70% of total ad spend by 2027. WPP Media separately has India crossing ₹2 trillion in ad revenue, up 9.7%. Note both reports gate their methodology — and that the crore-level precision above is the trade press reporting them, not a figure we could audit.

D2C operations: Unicommerce's April 2026 D2C report is built on 6,000+ D2C brands and 410 million shipments between April 2025 and February 2026 — a genuine, disclosed sample. India D2C GMV grew 33% in FY2026. Category growth: Health & Pharma 48%, Beauty & Personal Care 41%, FMCG 32%, Fashion 21%, Home 19%. Tier 2/3 cities drove 66% of incremental order volume.

The operationally useful finding: cash-on-delivery returns run 58% during festive periods versus under 15% for prepaid, and return-to-origin rates swing from 39% in November to 21% by March. If you're modelling Indian D2C unit economics on a flat return rate, that seasonality will break your forecast. Standard vendor-bias caveat applies — Unicommerce is measuring its own customers, not a random sample.

For the acquisition-cost side, our CAC and CPC benchmarks for Southeast Asia and India covers India's Meta CPC in detail, including its extreme month-to-month volatility.

App Economics: The Regional Gap Doesn't Close

The clearest regional numbers in the whole compilation are app monetisation, from RevenueCat's State of Subscription Apps 2026 — 115,000+ apps, $16B in revenue, over a billion transactions, one of the few disclosed samples here.

Median revenue per install at Day 14: $0.38 in North America versus $0.08 in India/SEA — roughly a 5x gap. Western Europe sits at $0.25, APAC broadly at $0.28. By Day 60 the North American figure reaches $0.55 and India/SEA reaches $0.11 — still roughly 5x apart. The gap doesn't narrow with time. Year-one revenue per payer runs $32 in North America against $14 in India/SEA, and Day-35 download-to-paid conversion is 2.6% versus 1.4%.

We'll resist the temptation to say the gap widens, tempting as 4.75x-to-5.0x looks. These are two-significant-figure published values, and that apparent movement is smaller than the uncertainty in the inputs. Not closing is the finding. Anything more precise would be reading noise.

A correction worth flagging: earlier Coact posts cite the 2025 edition of this report ($0.06–$0.09 versus $0.39). The 2026 edition supersedes those figures with a larger sample. Both were accurate when published; this page uses the current edition.

The implication for anyone scaling here is unforgiving and we've written about it separately in how to scale app installs without losing retention: at a fifth of the revenue per install, there is almost no margin to subsidise users who churn. Retention discipline isn't an optimisation in this region — it's the precondition.

RevenueCat is benchmarking its own customer base. Directionally strong, not census data.

The Forecasters Don't Agree With Each Other

Here's a fact that should temper how confidently anyone cites a forecast.

For global ad spend growth in 2026, dentsu forecasts +5.0% (27 May 2026, 56 markets). WPP Media forecasts +8.9% (16 June 2026). Same year. Three weeks apart. For APAC specifically, dentsu says +5.9% and WPP Media says +6.7%.

Neither discloses a methodology in its public release, so the gap cannot be reconciled — you cannot tell whether they disagree about the economy, about what counts as ad spend, or about something else entirely. A 3.9-point spread on a $1T+ market is not a rounding difference; it's roughly $40B of disagreement.

The honest use: quote a forecast as one house's view, attributed and dated, never as "the market will grow X%."

How to Use Benchmarks Without Being Misled

Four rules, drawn from everything above:

1. Open the methodology before you quote the number. Not the article citing the report — the report. Most of what failed here failed at that step, and it takes two minutes.
2. Ask what the sample actually is. Vendor benchmarks — Triple Whale, RevenueCat, Unicommerce, Adjust — measure their own customers. That's directionally useful and it is not the market.
3. Treat a missing methodology as a missing number. Momentum Works and dentsu are credible outfits publishing figures with no stated method. Use them for scale, never for precision.
4. For Indonesia, measure rather than benchmark. There is no honest shortcut. Anything that looks like one was invented.

Our companion posts apply these to specific questions: what's a good ROAS by industry, Google Ads versus Meta Ads for SEA ecommerce, and why CAC is increasing.

Frequently Asked Questions

What is the average CAC in Southeast Asia?

No verified figure exists. We searched specifically for one and found only stale data, US figures relabelled as regional, or invented numbers — including a widely-ranking page whose sources are from 2019 and 2020. The best available dollar-denominated proxies are ad-platform CPC and app revenue-per-install, not full-funnel CAC. Any source quoting one precise SEA CAC number should be checked against its methodology before use.

Does the e-Conomy SEA report contain CAC data?

Not in anything we could access, and none of its publishers claim it does. We checked Bain's, Google's and Temasek's summary pages, contents listings and press materials for the 2025 edition: it covers GMV, revenue, and sector splits, with no acquisition cost, CPC, or ad cost data presented. The full PDF exceeds our fetch limit, so we're reporting what the publishers present rather than a page-by-page read. It is nonetheless routinely cited as a source for SEA CAC claims.

What is the average CPC in Indonesia?

There is no Indonesia-specific CPC benchmark with a disclosed methodology. The pages claiming one are either unsourced or, in at least one case, generated with help from AI chatbots that the report lists as validating sources. Indonesia is 37% of Southeast Asian platform GMV and remains the region's largest data gap.

How big is the Southeast Asian digital economy?

Past $300B in GMV and $135B in revenue for 2025, both growing about 15% year over year, per Google/Temasek/Bain's e-Conomy SEA 2025 published in November 2025. Ecommerce is the largest sector at $185B GMV. Note the report names its inputs but discloses no sample size.

How much revenue does an app earn per install in India and Southeast Asia?

About $0.08 by Day 14 and $0.11 by Day 60, against $0.38 and $0.55 in North America — roughly a fifth, per RevenueCat's State of Subscription Apps 2026, based on 115,000+ apps. The regional gap stays around 5x rather than closing over time. RevenueCat benchmarks its own customer base.

Why do ad spend forecasts for 2026 disagree so much?

dentsu forecasts +5.0% global growth for 2026 and WPP Media forecasts +8.9%, published three weeks apart. Neither discloses its methodology publicly, so the roughly 3.9-point gap cannot be reconciled from the outside. Treat any forecast as one house's attributed, dated view rather than a market fact.

Which Southeast Asian benchmark sources can I actually trust?

The ones that disclose a sample: DataReportal for audience reach, RevenueCat for app monetisation, and Unicommerce for Indian D2C operations — with the caveat that the latter two measure their own customer bases. e-Conomy SEA and Momentum Works are credible for market size but publish no methodology. Everything else in our search failed verification.

Conclusion

The useful version of a Southeast Asia benchmark hub in 2026 is mostly a map of what isn't there. Market size is well covered and unaudited. Audience reach is genuinely well documented. Cost, conversion, and retention data ranges from thin to entirely absent — and in Indonesia, the region's biggest market, it's absent. The pages that appear to fill that gap have filled it with recycled US data, six-year-old reports, or output from a chatbot.

That leaves an unglamorous conclusion: in this region, your own account history is not a fallback when benchmarks are unavailable. It is the benchmark. Everything on this page is context around it.

How this page was compiled. Every figure is attributed to a named source with a publication date, and its methodology status is stated: disclosed sample (DataReportal, RevenueCat, Unicommerce), named inputs but no sample size (e-Conomy SEA), or no methodology published (Momentum Works, dentsu, WPP Media). Vendor selection bias is flagged wherever it applies. Six ranking sources were rejected; each is characterised by report title and its failure mode documented rather than quietly omitted — a DNS failure, a 403 on a cited PDF, a self-disclosed AI methodology, and an 80% divergence from our own figure. We have deliberately not linked them: the reasons are reproducible from the titles, and we would rather not send traffic to them. Two claims we could not fully verify: Indonesia's country-level GMV split comes from trade-press reporting of e-Conomy SEA rather than the primary PDF, which exceeds our fetch limit; and the dentsu India figures come from trade-press coverage because the report itself is gated. Compiled 2026-07-12. Coact is a performance marketing agency working with ecommerce and app businesses across Singapore, India, and Indonesia.

Continue Learning

• Real CAC Benchmarks for Southeast Asia & India (2026) — the country-level CPC data behind this page
• What's a Good ROAS by Industry? — and how to build a target from your own margins
• Google Ads vs Meta Ads for Ecommerce in Southeast Asia
• Why Is My CAC Increasing?
• Performance marketing agency in India
• Talk to our team about measuring a market that has no benchmarks.