Paid acquisition in Southeast Asia: what the platform benchmarks don't tell you

Meta CPMs in Indonesia often run 60 to 70 percent below Singapore rates. Every regional growth plan treats this as an opportunity. Many founders who have run the campaigns will tell you it's a trap — not because the traffic is bad, but because cheaper reach is not the same thing as cheaper customer acquisition, and the unit economics of SEA markets require a completely different model.

The CPM misreading

Let's be precise about what a low CPM means and what it doesn't. Cost per thousand impressions is set by advertiser auction pressure. When CPMs in Indonesia are, say, USD 1.50 and CPMs in Singapore USD 4.50, it means advertisers are bidding less for Indonesian inventory. They are bidding less because Indonesian consumers, on average, generate less revenue per customer acquired than Singaporean consumers.

The CPM number is not a mistake in the platform's pricing system. It is the market's accurate assessment of what each impression is worth to the advertiser buying it. When you see a 70% lower CPM and think "cheaper growth," you are seeing an output metric and mistaking it for an input opportunity. The platform knows something your spreadsheet doesn't: the customer on the other side of that impression generates less revenue over their lifetime.

Indonesia's per-capita GDP is roughly USD 5,000. Singapore's is roughly USD 85,000. That is not a 70% gap — it is a gap of about 17x. The CPM discount is not proportional to the economic gap, which means Indonesian impressions are arguably still overpriced relative to the LTV of a median Indonesian consumer for most product categories.

The three compounding problems

Low CPM in SEA markets comes with three structural problems that compound on each other and are invisible in the headline metrics.

Lower intent quality at equivalent awareness stages. An Indonesian user clicking a performance ad for a subscription SaaS tool is, on average, earlier in their consideration journey than an equivalent click from Singapore or Australia. The conversion rate from click to free trial, and from free trial to paid, is structurally lower — not because Indonesians are less decisive, but because the market for that category is less mature and the user's familiarity with the purchase pattern is lower. A campaign that converts paid clicks to free trials at a healthy rate in Singapore will often convert at a fraction of that rate in Indonesia.

Faster creative fatigue. Indonesia is one of Meta's largest user bases in the world, and a large share of its internet-connected population is on Meta's platforms. That sounds like a targeting advantage. It is also a creative fatigue problem: your target audience is smaller in absolute terms relative to the volume of ads being served. Campaigns that sustain effective frequency in Australia for two months or more before significant fatigue can show fatigue indicators in Indonesia within a few weeks. If you are not built for rapid creative iteration, your CPA will climb faster than your optimisation cycles can compensate.

ARPU compression and willingness-to-pay misalignment. For SaaS products priced for Western markets, the conversion rate from trial to paid in Indonesia is not just a payment intent problem — it is a price point problem. A USD 29/month tool is a meaningful slice of monthly income for a median Indonesian professional. In Singapore it is well under 1%. Products that have not localised their pricing for SEA markets are attempting to sell at a price point that is structurally inaccessible to a significant fraction of their addressable users, regardless of how good the product is.

When the math works and when it doesn't

Paid acquisition in SEA markets works well under specific conditions. The most favourable is a high-volume, low-ARPU consumer model — a gaming app, a food delivery platform, a short-video product — where the product-market fit is inherently adapted to the income level. A mobile game that monetises through small in-app purchases (USD 0.99 to USD 4.99) at volume can generate positive unit economics from Indonesian users at current CPM and conversion rates. The math closes.

It works for B2C subscription products that have localised their pricing. Spotify prices Premium in Indonesia at a fraction of its US price. This is not charity — it is the price point at which the willingness-to-pay curve intersects the subscription value for the Indonesian market. If you are running a consumer subscription and you haven't done this analysis, you are running your acquisition spend against a price point that will suppress conversion and inflate your effective CAC.

It works for products with network effects that make each user intrinsically valuable independent of their direct revenue contribution. A marketplace adds supply-side value when a new seller joins even if their transaction value is low. A communication tool adds value to existing teams when a new workspace member joins even if that member is on a free tier. In these cases, the LTV calculation needs to include the indirect value of network expansion, not just the direct ARPU of that user.

Where paid acquisition in SEA reliably does not work: SaaS priced above USD 20/month targeting individual users with no localisation, without deep product-market fit validation in the specific SEA market, relying on conversion funnels built for Western user behaviour patterns. The economics look reasonable at the CPM level and break down across every subsequent stage in the funnel.

Building a SEA-specific acquisition model

The correct approach is not to import your existing CAC model and apply a CPM discount factor. It is to build the acquisition model from the bottom of the funnel up, using market-specific data at each stage.

Start with ARPU by market. Not your global ARPU — your actual average revenue per paid user in Indonesia, or the Philippines, or Vietnam. If you don't have this data because you haven't launched there yet, use public benchmark data from comparable products and discount it by 40% for conservatism. This is your LTV ceiling. Work backwards: at a 3:1 LTV:CAC ratio, what is your maximum sustainable CAC? Now model the funnel — what click-through rate, trial conversion rate, and trial-to-paid rate do you need to hit that CAC at current CPM levels?

For many B2B SaaS products at Western price points, this calculation produces a required trial-to-paid conversion rate above what most products achieve even in their home markets — and SEA markets without localised pricing typically convert well below the home-market rate. The math does not close. Knowing this before you spend the budget is valuable.

For consumer products with localised pricing, the calculation often closes at lower conversion thresholds because the CPM discount is larger relative to the ARPU gap. A product where Indonesian ARPU is 35% of Singapore ARPU but Indonesian CPM is 25% of Singapore CPM has positive unit economics differential — you are getting a larger CPM discount than the ARPU discount warrants.

The creative and channel reality

Creative strategy for SEA paid acquisition is categorically different from creative strategy for Western markets. Three differences matter most.

First, language fragmentation. Indonesia has over 700 regional languages and a lingua franca (Bahasa Indonesia) that is most people's second language. Philippines campaigns need to distinguish between Tagalog, English, and code-switched creative. Thailand requires Thai-language creative to perform — English-language performance creative typically underperforms equivalent Thai creative on click-through and conversion, often by a wide margin. If your creative production pipeline is not built for localisation at this level of specificity, your CPM discount will be consumed by conversion rate underperformance.

Second, platform mix. The dominant platforms in SEA are not identical to Western markets. TikTok Shop has captured a significant share of product discovery and purchase intent in Indonesia and Thailand that Meta and Google don't fully replicate. LINE is the dominant communication platform in Thailand. Shopee's platform advertising reaches purchase-intent users in Indonesia, the Philippines, and Malaysia at conversion rates that can exceed Meta performance for commerce categories. A paid acquisition strategy that treats Meta as the primary channel in all SEA markets is leaving significant efficiency on the table.

Third, video-first creative is not optional. The SEA mobile user base is predominantly on low-to-mid tier devices consuming video content. Static creative typically underperforms video creative on CTR in most SEA markets. Short-form video (6-15 seconds) with on-screen text captions outperforms longer video because a significant fraction of mobile usage occurs without sound. If your creative operation produces primarily static assets, you are structurally handicapped in SEA paid channels.

• 01
  Model the full funnel before the first dollar of spend. Build market-specific conversion rates at each stage — click to trial, trial to paid, paid to 12-month retention — using SEA benchmarks, not global averages. A CPM that looks favourable will often produce a CAC that is identical to or worse than a higher-CPM Western market once the full funnel is modelled.
• 02
  Localise pricing before localising creative. Launching in Indonesia at USD 29/month and testing creative variants is a waste of budget. The price point issue will suppress conversion regardless of creative quality. Solve the pricing model first — the creative optimisation question only becomes meaningful at a price point the market can absorb.
• 03
  Plan for three-week creative refresh cycles. Creative fatigue in high-penetration SEA markets (Indonesia, Philippines) accelerates faster than most Western markets. Production pipelines that assume six to eight week creative cycles will show frequency-driven CPA increases at week three to four and struggle to recover without a significant creative volume buffer.
• 04
  Treat each SEA market as a separate market. Indonesia, Thailand, Vietnam, Philippines, and Malaysia are not one market. CPMs, platform preferences, language requirements, payment infrastructure, and willingness-to-pay profiles are materially different across these five countries. A campaign that performs in one will not automatically transfer to another.
• 05
  Instrument intent quality, not just volume. Track the conversion rates at each funnel stage by acquisition channel and market from day one. A CPM that is 70% lower but produces a trial rate that is 60% lower is not a 10% efficiency gain — after compounding through the rest of the funnel, it may be a net loss. Volume metrics without conversion quality data produce misleading unit economics.

Frequently asked questions.