All figures current as of September 2026. Quarterly indicators are dated inline and will age; the argument they support is not built to depend on any one of them.
I. Three numbers that cannot all be true
In February 2026, IMARC Group published an estimate that the Philippine social commerce market reached USD 28.4 billion in 2025. Later the same year, the same firm published an estimate that the Philippine total e-commerce market reached USD 28.0 billion in 2025.
Social commerce is a subset of e-commerce. IMARC's two numbers say that subset is 101% of the whole.
In January 2026, Ken Research valued Philippine social commerce at USD 2.3 billion — a twelfth of IMARC's figure. In August 2026, Mordor Intelligence put total Philippine e-commerce at USD 17.65 billion for 2025. ECDB put it at USD 24.77 billion. ResearchAndMarkets put social commerce alone at USD 28.77 billion for 2026.
Across six firms publishing within a twelve-month window, the estimates of Philippine online commerce span more than an order of magnitude, and two of them are internally impossible.
None of these firms publishes its methodology. All of them sell the report the number is designed to make you buy.
The most charitable reading available is that some of this spread is definitional rather than incompetent — one firm may count informal Facebook and live-selling transactions inside social commerce while another counts only formal platform GMV inside e-commerce, and two figures scoped that differently could both be defensible. I cannot rule that out. Neither can you, and that is the point: without disclosed methodology, a reader has no way to distinguish a scoping difference from an error, and no basis for choosing between a USD 2.3 billion estimate and a USD 28.8 billion one for the same thing in the same year. A number you cannot interrogate is not evidence, whatever produced it.
This is not a story about sloppy vendors. It is the visible edge of a structural condition: the Philippines is a market where transactional data is tracked reasonably well and attitudinal data is barely tracked at all, and the gap between them has been filled by parties with a commercial interest in what you conclude. Most strategic decisions made about the Filipino consumer — pricing, channel, segment, forecast — are being made on that fill.
A narrowing, so the title is not read as more than it claims. Some organisations know a great deal about this market. PSA and BSP run large, disclosed, methodologically serious instruments. Worldpanel, NIQ and Kantar hold continuous panel and survey data, much of it behind subscriptions costing more than most Philippine companies will spend on research in a decade. The claim is not that the knowledge does not exist. It is that the freely available picture — the one that circulates, gets cited, and reaches the desk where the decision is made — is badly wrong, and the people using it have no way to tell. That is a smaller claim than the title. It is still the one that governs most commercial decisions in this country.
If you run commercial strategy in this market, that is your operating environment. This essay is about what survives scrutiny, what does not, and what the scarcity itself implies for how you should be building.
II. The zombie statistic
Search for Filipino consumer psychology and you will hit this within three clicks: 91% of Filipinos trust word-of-mouth recommendations — the highest in Southeast Asia.
It is cited constantly, usually without a date, usually in the present tense, usually as the foundation of an argument about why referral and community marketing work here.
It comes from Nielsen's Global Trust in Advertising Survey, released 28 September 2015.
The methodology was roughly 30,000 online respondents across 60 countries, measuring 19 advertising formats. In the same wave, 80% of Filipino respondents reported trusting editorial content, 75% trusting online consumer opinions, 59% trusting social network advertising, 56% trusting search advertising. The 2013 wave had put the word-of-mouth figure at 89%.
It is eleven years old. It predates TikTok's arrival in the Philippines, GCash at national scale, Shopee's market dominance, and the entire live-selling economy. Its online-panel method in 2015 over-represented connected urban respondents far more severely than the same method would today, in a country that has since added tens of millions of internet users.
I could not find a single post-2020 Philippine-specific replication of that measurement.
That absence is the finding. An eleven-year-old number continues to anchor commercial reasoning about Filipino buying psychology because nothing has replaced it — not because anyone verified it still holds.
To be precise about what this does and does not prove: stale is not the same as false. It is entirely possible that Filipino word-of-mouth trust remains among the highest in the region — the underlying cultural claim is plausible and may well be stable across a decade. What has been demonstrated is a sourcing failure, not a factual error. The problem is that a decision made on this number is being made on faith rather than evidence, and the people citing it do not know which.
III. Everyone selling you the live-selling story sells live selling
The current consensus is that live commerce is how the Philippines shops. The country is described as one of the most active live-selling markets in the world, shoppertainment as the default mode, the suki relationship — the repeat-buyer bond with a specific seller — as the cultural substrate that makes it inevitable.
Trace the sources. The confident version comes from live-commerce production agencies, logistics firms selling to live sellers, and market-research houses selling reports to both.
The measured version is thinner and more interesting.
Momentum Works and Tabcut, tracking TikTok Shop globally for the first half of 2026, found GMV above USD 50.3 billion, up 92% year on year. The channel split: slightly more than half through the Shop tab, roughly 40% through short-form video, and approximately 8% through live-streaming. A separate read of the same platform from Tabcut and Dashboardly puts it at 36% shop tab, 58% video, 14% live. The trackers disagree by nearly double on the live share. Both put live in the minority, and neither is Philippines-specific.
Global figures, not Philippines-specific. Shares shown as each tracker reports them; the second tracker's three shares sum to 108%.
On the Philippine side, Rakuten Insight surveyed 8,221 respondents aged 16+ across APAC in May 2024. Among Filipino respondents, 26% had purchased via livestream three to six times in the preceding twelve months. Eighteen percent neither watched live commerce nor bought through it.
There is no published Philippines-specific split of live versus video versus catalog GMV. It does not exist. Anyone presenting one is modeling, not measuring.
What is measurable is more precise and more useful than the hype. Seller-panel data from Cube.asia for FY2025 shows beauty and personal care at 28% of TikTok Shop GMV in the Philippines — the platform's largest category and the highest beauty concentration of any Philippine platform. Over the same window, beauty's share of Shopee's Philippine GMV declined from 17% in FY2023 to 14% in FY2025.
28% is the highest beauty concentration of any Philippine platform; on Shopee the category lost 3 points in two years.
That is a real, category-specific migration with a legible mechanism: beauty requires demonstration to convert, live selling supplies demonstration, and the category moved to where the demonstration is. It is a much narrower claim than "Filipinos shop by livestream," and unlike that claim, it is supported.
The platform rebalancing underneath is sharper still. Lazada Philippines GMV fell to USD 3 billion in FY2025, down 34% year on year, its share of combined platform GMV dropping from 28% to 16%. In Q4 2025 — the peak seasonal quarter, when every platform should benefit — Lazada generated roughly 10% of combined platform GMV. TikTok Shop grew 53% over the same period.
IV. Why the gap exists
Three structural reasons, none of which are anyone's fault, all of which are durable.
Attitudinal research has no sustainable buyer. Transactional data gets collected because someone monetizes it — platforms, payment processors, purchase panels. Attitudinal data requires someone to fund fieldwork with no direct transaction attached. In the Philippines, the sustained attempt is EON Group's Philippine Trust Index, running biennially since 2012. Its 2021 wave sampled 800 respondents. Its 2017 wave sampled 1,200. Those are serious efforts and they are also too small to disaggregate below the national level — which means there is effectively no published attitudinal data distinguishing a consumer in Davao from one in Makati.
At 800–1,200 respondents, the attitudinal index cannot be disaggregated below the national level.
Regional averaging erases the country. Most "Southeast Asia" consumer research reports a regional figure, and the Philippines is a minority of that sample. A market with 98 million internet users, near-total smartphone penetration, and a cash-heavy payment culture gets averaged against Singapore. The resulting number describes neither.
The measurement boundary is drawn where the trackers can see. Every Philippine e-commerce market-share chart covers Shopee, Lazada, and TikTok Shop. Facebook Marketplace and Facebook Group commerce appear in none of them, because Meta publishes no Philippine Marketplace volume and no panel captures it. Whatever share of real Philippine consumer transaction runs through Marketplace — and the anecdotal consensus is that it is large — is structurally invisible to every chart you have seen.
Tracked by seller panels and platform disclosures.
Meta publishes no Philippine Marketplace volume; no panel captures it. Anecdotal consensus: large.
V. The thin layer that holds
Strip out everything without disclosed methodology and something real remains. It is smaller than the literature suggests and considerably more reliable.
The Philippine Statistics Authority runs the Family Income and Expenditure Survey and the national accounts. The Bangko Sentral ng Pilipinas runs the quarterly Consumer Expectations Survey and the Consumer Finance and Inclusion Survey. Continuous purchase panels — Worldpanel by Numerator, Kantar — observe what households actually buy rather than what they say. Payment processors report their own volumes with defined scope.
Four things that layer says, which most commentary gets wrong.
1. The consumer is reallocating, not retreating
The BSP Consumer Expectations Survey current-quarter Confidence Index, from a sample of about 5,000 households per round:
| Quarter | Index |
|---|---|
| Q3 2025 | −9.8 |
| Q4 2025 | −22.2 |
| Q1 2026 | −15.8 |
| Q2 2026 | −42.0 |
A diffusion index: it counts whether pessimists outnumber optimists, not how much anyone spends.
Q2 2026, fielded 6–18 April across 5,503 households, is the weakest reading since Q4 2020. The historical mean from 2007 to 2025 is about −16.
The obvious inference — the consumer is pulling back — is wrong, and the same statistical system says so.
PSA household final consumption expenditure for Q2 2026, at constant 2018 prices, year on year: transport −7.5%, alcohol and tobacco −1.9%, recreation and culture −0.8%, restaurants and hotels −0.2%. Over the same quarter: education +13.0%, miscellaneous goods and services +5.8%, food +2.4%. Total household consumption grew 2.8%, down from 5.2% a year earlier. Food alone is 36.3% of the total.
BSP's own Q2 2026 read adds that households were less likely to save, plausibly because prices were pushing a larger share of income into current expenditure, and that the spending outlook for essentials in the following quarter was more upbeat, even as headline confidence hit its floor.
The Confidence Index is a diffusion measure — it counts whether pessimists outnumber optimists, not how much anyone spends. Read alongside the expenditure data, the picture is not a consumer who stopped buying. It is a consumer who moved money out of mobility and leisure and into food and education, hard, within a single quarter.
If your plan cuts forecast uniformly on a weak confidence print, you are wrong in half your categories and you will not find out which half until you have already reallocated against it.
2. Incomes are rising and being outrun
PSA's 2025 FIES preliminary results, released 26 August 2026, put average annual family income at ₱411,350 — about ₱34,300 a month — up 16.5% from ₱353,230 in 2023. Average annual family expenditure was ₱321,850, up 24.7% from ₱258,050.
Expenditure grew half again as fast as income, nationally, over two years.
Follow that through to what PSA itself calls average family savings — income less expenditure, the framing the agency uses in its own regional releases:
| 2023 | 2025 | Change | |
|---|---|---|---|
| Average annual family income | ₱353,230 | ₱411,350 | +16.5% |
| Average annual family expenditure | ₱258,050 | ₱321,850 | +24.7% |
| Implied average annual savings | ₱95,180 | ₱89,500 | −6.0% |
Nominal pesos. 2025 is preliminary and the first biennial wave; income minus expenditure is a residual, not a deposit.
The average Filipino family's implied annual savings fell by about ₱5,700 in nominal terms between 2023 and 2025 — during a period when its measured income rose 16.5%. In real terms the decline is materially worse. This figure does not appear in the press coverage of the release, which led on the income gain.
Three caveats, because this is the most attackable number in the essay and it should be attacked properly.
It is not literally savings. Income minus expenditure is a residual. It absorbs debt service, remittances sent outward, business reinvestment and measurement error. PSA uses the savings label loosely in its own regional releases and I have followed that usage, but the quantity is a gap, not a deposit.
The comparison crosses a design change. 2025 is the first biennial FIES; the survey ran triennially from 1985 until this cycle. The 2025 figures are also explicitly preliminary. Both facts weaken a two-point comparison.
There is a benign reading, and it is not weak. Headline inflation was 1.7% in 2025. Against that, a 24.7% nominal rise in expenditure is a large real increase in consumption — which is what happens when households feel better off and release deferred demand, not only when they are squeezed. The implied savings rate fell from 26.9% to 21.8%, which remains high by international standards. An economist could read this table as recovery rather than pressure and would not be being unreasonable.
What tips it toward the pressure reading is not this table alone but its company. Discretionary categories contracting in the same period (§V.3). Borrowing intentions falling while essential-goods spending expectations rise (§V.4). FMCG growth at a record low with hard discounters growing 77% (§V.5). Households less likely to save, per BSP's own Q2 2026 commentary. Any one of those admits a benign reading. Five of them pointing the same direction is harder to dismiss — and that, rather than the savings figure in isolation, is the claim I would defend.
The regional detail sharpens it. CALABARZON posted the highest expenditure growth of any region at 34.6%; CAR the lowest at 7.6%. In Central Visayas, income rose 16% to ₱379,220 while expenditure rose 28.4% to ₱279,900. Cebu City household expenditure rose 37.3%.
The dispersion is severe. NCR averaged ₱574,370 against BARMM's ₱246,050 — a 2.3x spread. Makati, the highest-earning highly urbanized city at ₱796,990, is 3.2x BARMM and roughly double the national average. The four cities behind Makati are all in Metro Manila.
The four highly urbanized cities behind Makati are all in Metro Manila.
This is the number that should discipline any nationwide pricing decision. A price point calibrated to Metro Manila purchasing power is calibrated to roughly a third of the addressable country. And the composition data cuts the same way: in the 2021 FIES wave, SOCCSKSARGEN households spent about 55% of income on food against NCR's 37%. Discretionary headroom is not a national quantity.
Discretionary headroom is not a national quantity.
3. Seven in ten adults think borrowing is a bad idea
The BSP's 2025 Consumer Finance and Inclusion Survey — fielded February to July 2025, 8,784 completed interviews with adults 15 and over across all regions, from 10,836 eligible respondents at an 81% response rate — found that roughly seven in ten Filipino adults hold that taking on any type of loan is not a good idea.
The behavior matches the belief. Adults with outstanding loans fell to 25% in 2025, from 45% in 2021. Among those who did borrow, 58% paid on schedule and 19% paid ahead, while 34% reported frequent difficulty and 8% borrowed anew to service existing debt. Borrowing shifted from informal toward formal channels. The BSP's Q2 2026 expectations survey found households less inclined to borrow over the following year even while expecting higher spending on basics.
Borrower responses overlap, so they do not sum to 100%.
This is the most valuable datapoint available on Filipino buying psychology, for two reasons.
It is a directly measured attitudinal finding — rare in this market, on a large disclosed government sample.
And it is a standing challenge to the buy-now-pay-later growth narrative, which is promoted almost exclusively by parties selling buy-now-pay-later infrastructure, and which implicitly assumes this belief softens. Nobody has demonstrated that it does. A normative conviction held by seven in ten adults is not a friction to be engineered away over a product cycle.
If your commercial plan has a financing, instalment, or deferred-payment lever in it, calibrated on benchmarks from markets where consumer credit is normalized, that lever is weaker here than your model assumes.
4. Price beats value, and the discounters prove it
NielsenIQ reported in August 2026 that in the Philippines, low price remains the top shopping driver — against value for money across APAC generally. NIQ does not publicly disclose the sample or fielding behind that finding, which is a real limitation, so treat it as directional rather than settled.
The observed purchase data points the same way and is stronger evidence, because it records what households did rather than what they said.
Worldpanel by Numerator's 2026 FMCG Outlook forecasts Philippine fast-moving consumer goods spending growth of 3–4% for 2026, down from 5.2% in 2025, and describes it as among the slowest on record against a typical 5–6% range. Within that near-flat category, hard-discount formats — Dali, O!Save — are forecast to grow 77%, with e-commerce at 15% and traditional supermarkets and market stalls growing but slower.
Shaded extensions mark forecast ranges. Traditional supermarkets and market stalls: growing, but slower than e-commerce.
A category growing 3–4% in which the discount format grows 77% is not a market making nuanced value trade-offs. It is a market optimizing on price, and doing so fast enough to reshape channel structure inside a single year.
A necessary clarification, because the obvious objection is a good one. If Filipinos are price-sensitive, why is Metro Manila full of people buying ₱200 lattes?
Because price sensitivity is a statement about comparison behavior and elasticity, not about absolute spend. It does not mean buying cheap things. It means comparing continuously, switching readily, and holding tight control over transaction size — and those behaviors are category-specific within a single household. The same person can buy a premium coffee daily, comparison-shop three platforms for a phone charger, and switch detergent brands over ₱15. None of that is inconsistent. It is one budget with different elasticities across different lines of it.
The sharper problem is that the visible market and the actual market are not the same market, and almost all commentary is written from inside the visible one.
The Philippines is the world's largest importer of soluble instant coffee, accounting for roughly 28.85% of global instant coffee imports, per USDA forecasting for the 2025–26 marketing year. Within the domestic instant category, single-serve 3-in-1 and 2-in-1 sachets are the largest segment — explicitly because tingi retail lowers the upfront purchase price to a point every income level can reach.
So the Philippines is at once the most visible café-culture market in Southeast Asia and the largest instant-coffee buyer on earth. The second fact is by far the larger one and it appears in almost no consumer commentary, because specialty cafés cluster in NCR, Cebu and Davao, and NCR alone is roughly 14 million of 117 million people with average family incomes running well above the national figure.
This is the cleanest available illustration of the central problem. Anyone writing about Filipino consumers from Metro Manila — which is nearly everyone who writes about them — is sampling the most atypical 12% of the country and reporting it as the market. Not because they are careless, but because that is what is visible from where they stand, and no dataset corrects for it. It is also why the regional dispersion in §V.2 is not a footnote to the national averages. It is a warning that the national averages describe almost nobody.
5. The convenience premium is real and small, and it is losing to price
There is a widely held view — I held a version of it myself before running these numbers — that Filipino consumers over-index on convenience and lifestyle relative to comparable markets. The success of super-apps here is usually offered as the proof.
The strongest available test of that view is ride-hailing, and it fails.
Grab's Philippine dominance is genuine. Analyst estimates put its share at roughly 80–85% as of early 2026, down from an earlier position above 90%. These are estimates, not audited figures, and should be read as directional.
The direction is what matters. Grab's share is declining, and it is declining to inDrive — a challenger whose entire proposition is price. inDrive operates an auction-based pricing model at a commission rate around 10%, against an industry norm of 20–25%, and reporting on its 2025 Philippine expansion cites roughly 8x growth in rides and 7x in passengers. Grab has responded with vouchers and incentives, converting the Philippines from what analysts described as a straightforward profit pool into a market requiring defensive spend to hold position.
Share figures are analyst estimates, not audited results; read as directional.
A market that genuinely prized convenience above price would not do this. The incumbent owns convenience — deeper coverage, integrated payments, the full super-app stack. It is losing share anyway, to a competitor offering less convenience at a lower price.
The same pattern appears in FMCG, where hard discounters are forecast to grow 77% inside a category growing 3–4%, and in NIQ's finding that low price outranks value for money in the Philippines specifically.
So what is the convenience intuition picking up? Something real, but different from what it looks like.
The best-sourced evidence is SM Retail's own account of its 2026 customer behavior, given to BusinessWorld in September 2026: shoppers buying smaller packs, waiting for promotions, comparing prices more closely — while still reserving budget for small indulgences, a daily coffee or a meal out. This is a large retailer describing its own observed traffic, which makes it considerably better evidence than the survey commentary surrounding it, though it remains a self-report from an interested party.
Philippine dining figures circulating commercially suggest a high frequency of out-of-home eating at a low average ticket. I have not been able to verify any of them against a disclosed methodology and would not build on them, so the claim here rests on the SM Retail account and on the structural evidence elsewhere in this essay rather than on a per-meal number.
That pattern — continued purchase of small indulgences alongside active downsizing and promotion-waiting on staples — is not lifestyle-led consumption. It is indulgence preserved by keeping the ticket small, which is the same tightly-controlled transaction structure visible in the COD value gap, in pack-size downsizing, and in the credit aversion above, applied to treats rather than necessities.
The commercial implication inverts the obvious one. Convenience and experience are not price-insensitive categories here; they are categories where Filipinos have found a way to keep buying by keeping the ticket small. A premium convenience play priced on the assumption that Filipinos will pay for convenience is reading the frequency and missing the ticket size.
6. Influence and trust are not the same thing, and Filipinos keep them separate
A common working theory among Philippine operators — and an intuitive one — is that Filipino consumers trust third-party creators more than they trust brands, and that sponsorship disclosure does not much damage that trust.
The first half survives testing. The second half needs a correction that matters commercially.
On reach, the gap is genuine. Rakuten Insight found in 2023 that 86% of Filipino social media users follow at least one influencer, against 22.6% globally — a roughly fourfold gap. That figure reaches me through agency citation rather than Rakuten's own publication, so treat the exact number cautiously, but a gap of that magnitude is unlikely to be an artifact.
One widely circulated supporting statistic should be handled with tongs. The claim that over 90% of Filipino consumers prefer influencer accounts to brand accounts appears across Philippine marketing content, attributed variously and traceably to nobody's published methodology. It is the second zombie statistic in this essay, and it is younger than the first only because it has had less time to age.
Where the theory breaks is the equation of influence with trust. Kantar's Media Reactions 2025 Philippines work draws the distinction explicitly: traditional environments — television, out-of-home — continue to carry institutional authority, while digital platforms carry relevance and drive discovery and conversion. Kantar's own summary of the Philippine finding is that credibility earns belief while relevance earns action. Filipino consumers, on this reading, have not transferred trust from institutions to creators. They have learned to use each for a different job.
Apply the same commercial-interest test used elsewhere in this essay and it lands on Kantar too. Kantar sells traditional media measurement, and a finding that traditional media retains institutional authority runs in the direction of Kantar's revenue. I find the distinction persuasive because it is falsifiable and because it explains an otherwise awkward pattern — extremely high creator following alongside stubborn brand-trust research findings — but it deserves the same discount I applied to the live-selling agencies, and I nearly failed to apply it because the conclusion suited my argument.
And sponsorship tolerance is conditional, not unconditional. Research presented by UCN in March 2026 describes Filipino audiences moving toward lo-fi, unfiltered creator content and away from polished production, with brands responding by shifting from one-off sponsored posts toward sustained creator relationships. UCN reports that 85% of Philippine brands now run integrity checks on creators before signing, driven by audiences becoming adept at identifying scripted endorsement. Commercial interest applies — UCN operates in the creator economy — so read it as directional.
The synthesis: Filipino consumers are not indifferent to the fact that a creator is paid. They are indifferent to disclosure and highly sensitive to performance. A visibly sponsored recommendation that reads as genuinely used survives. A visibly sponsored recommendation that reads as scripted does not, and disclosure was never the variable.
This maps directly onto EON's own finding about business trust — that patronage frequently reflects constrained choice rather than belief. The same separation runs through both: the thing that gets the action and the thing that holds the belief are not the same thing, and the Philippine consumer keeps them apart more deliberately than most frameworks assume. A creator strategy that buys action and assumes it has bought credibility has bought one of the two.
7. Bundling: a real pattern with no serious measurement, and a contradiction inside it
The view that Filipino consumers respond unusually well to bundles is widely held and, as far as I can determine, entirely unmeasured. Philippine marketing content asserts strong response to bundle deals, buy-one-get-one offers and family packs; none of it discloses a methodology, and I found no study comparing Philippine bundle attach rates against any other market.
There is also a genuine tension worth resolving rather than ignoring.
The Philippines is the archetypal sachet economy. The tingi purchasing pattern — buying in the smallest available unit — is the most distinctive documented feature of Philippine retail, deeply embedded in the sari-sari store network, and is explicitly attributed to the daily-budgeting constraints of low-income households. Figures circulating on daily sachet consumption run into the hundreds of millions of units, though the provenance of those figures is a distributor blog and I would not publish the number.
Buying in the smallest possible unit and responding strongly to bundles are, on their face, opposite behaviors.
Inference, mine, offered as a hypothesis rather than a finding: the reconciling variable is not quantity but visible saving inside an affordable single ticket. Tingi minimizes the ticket when the saving is invisible. A bundle wins when it makes the per-unit saving legible while keeping the total within one affordable transaction. Both are the same behavior — tight control of transaction size with continuous price comparison — expressed under different offer structures.
If that is right, the practical consequence is that bundle size is capped by ticket affordability rather than by perceived value, and a bundle that improves unit economics while pushing the total past what a household will commit in one go will underperform a smaller bundle with a worse unit price. I have no data testing this, and it is the kind of claim a company could settle in a fortnight with its own transaction records. Which is, again, the argument of this essay.
8. The foreign entrant's discount playbook is not a misreading of the market. It is a misreading of the mechanism.
Nearly every international consumer business entering the Philippines runs a version of the same play: aggressive discounting, heavy promotional spend, price positioned visibly below perceived value. The country is read from outside as a pure price market, and the entry strategy follows.
The first thing to say is that the read is broadly correct. Everything in this essay supports it. NIQ finds low price outranking value for money in the Philippines specifically, against value for money across APAC. Hard discounters are forecast to grow 77% inside a category growing 3–4%. inDrive is taking share from the region's most entrenched super-app. A firm concluding that Filipinos buy on price is not confused.
The second thing to say is that the playbook fails anyway, and the reason is precise.
Temu entered Southeast Asia through the Philippines in August 2023 — chosen first in the region, reportedly for its cross-border fulfilment infrastructure, English-language operation, and the fastest e-commerce growth rate in the world at the time. The launch ran the standard playbook, advertising discounts up to 90%.
By 2025, the Philippines was contributing roughly 0.3% of Temu's global traffic — on the order of 1.3 million monthly visits against a global figure near 446 million, a volume one analysis compared to that of a single large independent seller. Southeast Asia became Temu's weakest region globally, in the market where its playbook had won hardest elsewhere.
Then the detail that explains it. Cube Asia's product-level price comparison found Temu's price competitiveness in the region mixed, with three in five sampled products priced higher than on Shopee, Lazada or TikTok Shop.
Temu advertised 90% off and lost the shelf price on most of the basket.
~1.3M monthly Philippine visits vs. ~446M global
Advertised: up to 90% off. Priced higher on 3 in 5 sampled products.
Traffic share is a proxy for commercial performance; PDD Holdings does not disclose Philippine results.
The obvious objection: Temu struggled across the whole region. Its Malaysian operation saw little activity, a planned Thai launch slipped, and Vietnam suspended it outright within two months of entry on regulatory grounds. If Temu underperforms everywhere in Southeast Asia, the explanation may be Temu's cross-border model, logistics economics and regulatory exposure rather than anything about Filipino consumers — and a single case pulled from a regional pattern is exactly the reasoning error this essay complains about elsewhere.
That objection is fair against the outcome and not against the mechanism. The traffic share tells you Temu failed; it does not tell you why. The Cube Asia price comparison does, and it is a direct product-level measurement rather than an inference from a result: the discount was advertised and the shelf price still lost. That finding stands whatever else was wrong with Temu's regional operation.
The harder objection: Shopee. Shopee leads Philippine e-commerce in both traffic and GMV, and Shopee is the most aggressively promotional platform in the market — monthly free-shipping vouchers, stacked discounts, the 9.9 and 11.11 events. If discount were simply a losing strategy here, the biggest discounter should not be winning.
The resolution is in Sea Limited's own disclosures, and it is precisely the distinction this section is about.
Shopee's Q3 2025 GMV rose over 28% year on year to USD 32.2 billion — with cost of services up 38.8%, driven by shipping subsidies, and adjusted EBITDA margin at 0.6% of GMV. The promotions are real, expensive, and thin-margin. But on Sea's Q4 2024 earnings call, management framed the strategy explicitly: long-term success in e-commerce would hinge on structural cost advantages and operational excellence, citing SPX Express's cost leadership and a reduction in Shopee's logistics cost per order. In the Philippines specifically, Shopee has been consolidating fulfilment into its own delivery network and pushing same-day and sub-four-hour delivery through its Instant Mart programme.
Shopee is not running promotions instead of building a cost structure. It is running promotions while building one, and using them to buy time and volume until the structure carries the price. Temu ran the promotion without the structure, and the price comparison exposed it. Lazada ran the promotion with neither and lost a third of its Philippine GMV.
Whether Shopee's bet works is genuinely open — a 0.6% EBITDA margin is not a settled victory, and the subsidy is currently larger than the structural advantage it is buying. But the mechanism is the same either way: the promotion is a bridge to a price, not a substitute for one.
Note also that inDrive, which is taking share from Grab, is not running a promotion at all. It operates at roughly a 10% commission against a 20–25% norm. Its price advantage is structural from the start.
The distinction the entrant playbook misses is between discount and price.
A discount is a claim about price. Price is a property of a cost structure. In most markets an entrant can run the first for long enough to buy habit, and by the time buyers check, switching costs have accumulated.
That arbitrage does not work here. This is a market where consumers spend among the highest hours online in the world, hold a smartphone almost universally, buy in small frequent tickets with low switching cost, downsize pack sizes to track unit economics, and — per NIQ — treat low price as the primary decision variable rather than value for money.
Those conditions are consistent with a population running continuous price comparison. They do not demonstrate it, and I want to be exact about that, because this is the essay's closing argument and it would be indefensible to smuggle an unmeasured mechanism into it after spending several thousand words objecting to exactly that. Nobody has measured Filipino price-comparison behavior — not frequency, not number of sources checked, not the size of the price gap that triggers a switch. What exists is a set of conditions under which such behavior would be cheap and easy, and a set of outcomes consistent with it happening. That is inference, and it is the weakest link in this essay.
What is measured is narrower and survives on its own: Temu advertised a 90% discount and was priced higher than competitors on three in five sampled products, and it did not take the market. That is a direct product-level observation, not a claim about what was going on in anyone's head.
So the defensible version of the argument runs: a discount claim did not survive contact with the actual shelf price in the one case where someone measured both. The explanation I find most plausible — that Filipino buyers check, continuously, at basket sizes where leaving costs nothing — is a hypothesis that fits the conditions and the outcomes, and it is one any company could test against its own repeat-purchase and switching data inside a month.
So the entrant's real error is not thinking Filipinos are price-sensitive. It is assuming price sensitivity can be satisfied by a promotion — that a market this focused on price can be bought with a claim about price rather than beaten on the underlying number.
The winners in this market are not the ones running the deepest discounts. They are the ones who can hold a lower price after the promotional budget stops. Discount is a promise. Price is a structure. Filipinos check.
VI. What this means for how you are actually reasoning
Three consequences, in increasing order of importance.
Audit what your commercial assumptions rest on. Not whether your team has market data — whether the specific numbers underneath your segmentation, your price points, and your channel mix have a named methodology, a date, and a producer without a stake in the conclusion. My experience running that audit on public Philippine sources is that most load-bearing figures fail at least two of those three tests. The eleven-year-old trust statistic is the clean example, but it is not the exception.
Report ranges where sources disagree, and say why they disagree. Cash on delivery is the instructive case. Stated preference: about 42%. Share of e-commerce value: about 23%. Share of transactions including other cash methods: about 22%. Share of orders: 14–15%. These four numbers are all roughly correct and they answer four different questions — and the spread between value share and preference share is itself informative, because COD orders run systematically smaller and lower in value than prepaid ones. A strategy deck citing one of them without saying which question it answers has imported a hidden assumption into a decision.
The gap between preference and value share is itself informative: COD orders run smaller and lower in value than prepaid ones.
And the consequential one: in a market with no reliable external data, internal instrumentation is a competitive asset rather than an administrative cost.
This inverts the usual logic. In markets with deep syndicated research, your own funnel data is one input among many and its marginal value is modest — you can buy a good approximation of what you would otherwise have to measure. In the Philippines you cannot buy that. What you can measure inside your own business is, for many questions, the only reliable read available on the Philippine consumer.
A company that knows its own conversion rates by region, its own price elasticity by segment, its own channel attribution, and its own repeat behavior is not merely better organized than a competitor who does not. In this market it is operating with information that competitor cannot purchase at any price. The advantage is structural, and it persists precisely because the data environment does not improve.
VII. What I do not know
This is desk research on public sources. I do not operate a consumer business in the Philippines — no seller account, no live-selling operation, no retail P&L — and nothing here is triangulated against transaction data of my own. Where I have read behavior into the numbers, I have tried to mark it.
I am also writing from Metro Manila, which makes me a case of the problem described in §V.4. My own consumption is not a useful guide to this market and I have not used it as one: the households in the FIES data spend a third to a half of income on food, and NCR sits at the top of a distribution whose bottom is less than half its average. Any operator reasoning from what they personally buy, or from what they see other people buying within a few kilometres of a business district, is working from the least representative sample the country offers. That includes me, and the only defence available is to stay on the survey data and say so.
What I bring instead is a specific and narrow advantage: no stake in the answer. I do not sell live-selling production, payment infrastructure, market-research subscriptions, or supplements. A meaningful share of the sources cited above sell exactly one of those, and their findings run in the direction of their revenue. In an environment this thin, disinterest is a credential. It is also the only one I am claiming.
Several things here will age, and I would rather name them than have a reader discover them.
The quarterly confidence figures are perishable by construction. The −42 print may prove a shock spike rather than a level shift; the subsequent rounds will settle that and this essay will not.
The ride-hailing share figures are analyst estimates rather than reported results. Grab does not publish Philippine-specific share, and inDrive's growth multiples come from reporting rather than audited disclosure. If those estimates are materially wrong, the specific illustration weakens — though the FMCG discount data and the NIQ price finding carry the same conclusion independently.
The Temu figures carry a similar limitation. Traffic share is a proxy for commercial performance, not a measure of it, and PDD Holdings does not disclose Philippine results. The Cube Asia price comparison is the stronger half of that argument, because it tests the mechanism directly rather than inferring it from an outcome.
Several holes may close, and I hope they do. There is no measurement of Facebook Marketplace volume, no Philippines-specific split of live versus video versus catalog commerce, no post-2020 replication of the word-of-mouth measurement, no published attitudinal data below the national level, no study comparing Philippine bundle response against any other market, and no research at all on consumer-level price negotiation — the tawad and suki dynamics invoked constantly as the cultural explanation for how Filipinos buy, and measured by nobody.
That last one is the gap I would most like to see closed, and its persistence is the argument in miniature. The mechanism everyone agrees explains Filipino buying behavior has never been measured. It is repeated because it sounds right, and it sounds right because it is repeated.
The thesis does not depend on any single figure above, and that is deliberate. It depends on the shape of the evidence: a market where what people buy is tracked adequately, where why they buy is barely tracked at all, and where the difference has been filled by interested parties. Individual numbers in this essay will be superseded. That condition will not be — and until it is, the companies that measure themselves will know things about this market that the companies buying reports cannot.
Curious where your own revenue system is losing the most ground?
Run the Constraint Index →