Why a merchant superapp, and why now
AnKedai's design rules trace to a 50-million-ledger acquisition machine, proven surplus economics at grocery chains, the unserved 1.1-million-merchant reality of Malaysia, and a compliance deadline that does the selling for us.
What the record actually says
One acquisition precedent, one surplus-economics pair, one market census, one deadline and one pricing truth.
A free digital ledger is the proven wedge into micro-merchants
Khatabook digitised the notebook of tens of millions of Indian micro-businesses by giving the core ledger away, then monetising adjacent value — payments, supplies, credit referral. Sequoia backed it because the ledger is the stickyest artifact a small business owns. AnKedai is the same machine, tuned for Malaysia: Bahasa-first, utang-aware, QR-native.
Khatabook · India's digital buku lima sen
Surprise surplus rescue works at grocery scale
Flashfood pushed CA$58M of near-date grocery through Loblaw stores in one year; Too Good To Go crossed 100 million users selling 'surprise bags'. The model is proven at chain scale — but nobody has priced it for the six-bungkus hawker. AnKedai's AnLebih surface is that gap: the same markdown curves, sized for a stall.
Flashfood × Loblaw (2024) · Too Good To Go 100M+ users
Malaysia's micro-economy is enormous, paper-run and unserved
Some 1.1 million MSMEs — the overwhelming majority micro — account for roughly 37–38% of GDP and nearly half of employment. The POS software industry serves the restaurant with floor space; the kedai runcit, the pasar-malam pitch and the weekend-moving trader run on a notebook. That unserved tier is the entire addressable market of AnKedai.
SME Corp / MIDIF statistics
Compliance is arriving whether the kedai is ready or not
Malaysia's phased e-invoice rollout reaches all remaining taxpayers by December 2026 — a deadline that lands hardest on exactly the businesses with no bookkeeping system. For AnKedai this is not a threat to mitigate but the demand signal to ride: ledger first, e-invoice as an export button on entries that already exist.
LHDN e-invoice mandate · phased rollout from Dec 2024
Merchants do pay — when the price is honest and flat
Malaysian digital-payment acceptance services have priced daily-coffee-level subscriptions that merchants accept when value is visible. The counter-example is commission and ad-auction models that tax the poorest margins in the economy. AnKedai's tiers are published, flat, and cancel-anytime — the pricing doctrine this number supports.
Senang.io pricing research (recorded in expansion batch 1)
The hidden tax on the merchant is admin, not rent
Counting stock, chasing utang, reconciling QR settlements and dreading month-end absorb hours a day that a photo-and-voice ledger gives back. Acquisition for a free tool runs on returned time, not promised revenue — the reason the AI surfaces lead with 'post nothing by hand'.
Micro-merchant time-and-motion synthesis, expansion research
So AnKedai is built this way
The ledger is free forever at the core
Acquisition happens at the notebook, not the checkout. Paid surfaces (queue, surplus logistics, e-invoice volume, staff) attach to a ledger the merchant already cannot lose. Never gate the bookkeeping itself.
Institution-pays or merchant-pays; members never pay
AnKedai's revenue never comes from the poorest margin-holder for basic dignity features — and council/market operators (AnPasar) pay for roster and lottery rails. The payer always names the surface they buy.
Records, never custody
Settlement rides licensed acquirer rails through DuitNow. Financing appears as referral to licensed partners using the merchant's own history. AnKedai holds books, not money — the blueprint refusal that keeps the posture clean.
Compliance as a by-product
The e-invoice trail exports from entries that already exist because bookkeeping is effortless — not from a parallel compliance module the merchant must feed. If the paperwork needs its own input, the ledger failed.
Waste math before growth math
AnLebih ships only when the surplus story is true for the stall: tonight's unsold, discounted, cleared — measured in recovered ringgit on the merchant's own ledger, not in mission statements.
Aggregate signals opt in, cohorts only
Price and demand telemetry that leaves the ledger (e.g. AnHarga index feeds) passes floor tests: thousands-deep cohorts, differential noise, explicit merchant opt-in, revocable. Individual books are not a raw material.
The Malaysian numbers
The tier AnKedai exists for, counted. Micro businesses are the overwhelming majority — and the software industry's least served.
Buku lima sen — fifty-sen paper ledgers — remain the default bookkeeping instrument for credit and cash across the kedai economy.
Two years from now, 'I keep a notebook' stops being an option at the exact moment it becomes an export from the notebook app.
No commissions on staples, no hidden take-rates, no visibility sold from behind the counter. The anti-Nextdoor clause of the pricing page.
One kedai evening, honestly described
At 20:40, six bungkus remain. Kak Maimun photographs them into AnLebih at the learned half-price; the students’ i-Saves pass makes the 21:00 window sweeter still; the sale posts to the same ledger as the day’s four hundred QR payments, and the e-invoice drafts itself from entries that were never typed. Nothing was commissioned, nothing was advertised, nothing about her customers left the shop. In the morning the restock note already knows Fridays move more selayang.
The supply side of the savings network — and of KPDN’s price index.
Ready to build something intelligent?
Let's engineer your next product — from idea to launch.