SEA unit economics Year 1 Year 4 divergence is the structural problem behind most of the regional startup failures of the past three years. The regional capital environment documented in the 2024 Bain and Temasek e-Conomy SEA report shows a market that has tightened meaningfully on capital availability while the operating cost base for scaling businesses has continued to grow, which compresses the window in which a business can discover whether its Year 4 economics actually work. The metrics that look strong at Year 1, including first-order contribution margin, early cohort lifetime value estimates, customer acquisition cost on performance channels, and gross merchandise value growth, are not the metrics that determine whether the business survives Year 4. By the time the business has aged into a mature cohort distribution, faced channel saturation on the paid acquisition stack, absorbed the compensation cost of a scaled team, and encountered the operational complexity that comes with cross-market expansion, the economics that looked clean at launch often look broken. The founders who do not model this divergence ahead of time typically raise against Year 1 economics and run out of capital trying to reach the Year 4 math.
The part most founders get wrong is the assumption that unit economics improve with scale. In SEA, unit economics usually get worse before they get better, and the period of degradation typically sits between Year 2 and Year 4. This is the trap. The business looks like it is scaling healthily because gross revenue and gross merchandise value are growing. The contribution margin is compressing because the channel mix is shifting, the customer mix is broadening beyond the initial high-intent cohort, and the operating cost base is growing faster than the revenue it is supposed to enable. By the time the numbers show the compression clearly, the capital runway is already planning for the Year 1 trajectory.
What Year 1 Math Actually Measures
Year 1 unit economics in SEA typically measure three things that do not persist. The first is the performance of the initial cohort, which is self-selected, high-intent and usually responding to launch-stage marketing efficiency that cannot be replicated at scale. The first thousand customers on almost any SEA consumer product come in at a customer acquisition cost that is structurally lower than the ten-thousandth customer, because the first cohort is pulled from the founders’ networks, early word-of-mouth channels, and the most efficient segment of paid acquisition before the targeting exhausts.
The second is the performance of the initial product configuration, which is deliberately narrow and deliberately aligned to a specific high-margin segment. The unit economics look clean because the product has been stripped to its highest-contribution features for a specific customer type. As the business scales, the product necessarily broadens to address adjacent segments, lower-margin use cases, and the retention dynamics of the long tail. The contribution margin compresses in a way that the Year 1 model did not anticipate.
The third is the performance of the initial operating structure, which is lean in a way that cannot be sustained. The Year 1 team is typically five to twenty people carrying functions that will require fifty to two hundred people by Year 4. The overhead is low because the functions have not been built yet. As the functions get built, the operating cost base grows on a step-function basis, often faster than the revenue that is supposed to justify them. This is the operating cost dynamic our analysis of why ERP deployments in SEA mid-market keep stalling examines from the enterprise tooling angle, and the pattern generalises across functional areas.
What Year 4 Math Actually Measures
Year 4 unit economics in SEA measure a different set of things. The first is the mature cohort distribution, which includes customers acquired at the efficient marginal CAC, customers acquired at the expensive marginal CAC during growth pushes, customers who churned, customers who reactivated, and the retention curve averaged across all of them. The true lifetime value is now measurable because enough time has passed to observe the tail of the retention curve. In SEA consumer categories, the Year 4 LTV typically comes in between 40 and 70 percent of the Year 1 estimate, not because the product got worse but because the initial LTV extrapolation was built on insufficient cohort maturity.
The second is the fully-loaded customer acquisition cost, which includes the paid channels, the brand spend, the content production, the partnerships, the sales team where applicable, and the share of overhead allocable to acquisition. Year 1 CAC typically captures only the performance marketing cost. Year 4 CAC captures the full cost of buying a customer, which is usually 1.5 to 3 times the Year 1 figure depending on category. In regulated categories or B2B categories, the gap can be wider.
The third is the operational cost of serving the customer, which includes the logistics, the customer service, the payment processing on a mature channel mix, the warranty or replacement reserve, and the allocation of the operations team supporting the customer base. Year 1 contribution margin is typically calculated on a simplified service cost that reflects the launch configuration. Year 4 contribution margin captures the real operational reality, which for most SEA consumer businesses is 10 to 20 percentage points lower than the Year 1 figure.
The Year 4 question is whether the business, operated at scale with the real cost base and the mature cohort behaviour, produces contribution margin that is sufficient to fund operating leverage over the next several years. For many SEA startups that raised against Year 1 math, the answer is no, and the capital runway ends before the answer changes. The venture-funded SEA technology landscape surveyed in the DealStreetAsia data on regional startup funding through 2024 registers the consequences, with write-downs, down-rounds and shutdowns clustering in the Year 3 to Year 5 cohort of companies whose Series A math has not matured into Year 4 math.
Modelling the Divergence Into the Year 1 Plan
The founders who navigate this successfully build two models at Series A. The first is the reported Year 1 unit economics, because that is what the business is currently demonstrating. The second is the projected Year 4 unit economics, with explicit stress tests on CAC, LTV and contribution margin. The investor conversation that makes sense is not whether the Year 1 numbers are real. It is whether the Year 4 numbers are defensible under realistic assumptions about cohort aging, channel saturation and operating cost scaling.
The pattern our analysis of why digital banks in SEA grow users but not revenue examines is an extreme example of this dynamic in the fintech category, where Year 1 user acquisition economics looked strong and Year 4 revenue economics have consistently disappointed the original thesis. The same dynamic, with different specifics, plays out in e-commerce, vertical SaaS, consumer health, and most platform categories across the region. The founders who model the Year 4 math honestly can structure the capital raise and the operational plan around the real trajectory. The founders who do not, raise on Year 1 math and discover the Year 4 math the hard way.
The Practical Takeaway
The practical question for operators is how to identify, inside the Year 1 numbers, the variables that will drag on the Year 4 numbers. The diagnostic is usually straightforward. Channel concentration is the first flag. If more than 60 percent of customers are coming through a single paid channel at Year 1, that channel’s CAC will inflate materially as targeting exhausts and competitors enter the auction. Cohort maturity is the second flag. If the LTV estimate is built on less than six months of observed retention, it is not yet an LTV estimate. It is a cohort projection. Operating leverage assumptions is the third flag. If the model assumes that gross margin improves with scale without specifying the mechanism, that assumption almost never holds. The founders who ask these three diagnostic questions in their own planning typically produce Year 4 numbers that resemble the Year 1 deck. The founders who do not, typically do not. The discipline that makes the difference is consistent with the operator analysis in the Boston Consulting Group Center for Digital Government and growth research on SEA digital business models, which documents the cohort-scaling dynamics that drive the Year 1 to Year 4 gap in actual regional data.
