LTT Outsourced CFO builds and runs the bookkeeping back office for other accounting firms. White-label, on your platform, invisible to your client. You keep the relationship and 40–70% of the fee. We keep the keystrokes.
Three files. Ninety days. Card rate. No long-term commitment.
Captions on by default. The player follows the language you pick above — English and Mandarin are separate films, not subtitles of each other.
Bookkeeping is the one service line in a Malaysian practice where the work is high-volume, low-judgement, hard to staff and increasingly automatable. It is therefore the first thing worth taking off a firm's fixed cost base.
A junior bookkeeper is hired, trained and supervised whether or not the files are there. When a client leaves, the fee goes and the salary stays. Turnover resets the training every eighteen months.
The staff who can do a clean bank reconciliation are the staff you would rather have on chargeable audit or tax work, where the realisation rate is materially better.
Preparing the accounting records of your own audit client raises a self-review threat under the MIA By-Laws, and is prohibited outright for public interest entities. Routing that work to an unrelated firm removes the threat at source — and you still earn 40% of the fee.
SMEs increasingly expect cloud books, monthly management figures and e-Invoice readiness. Building that capability in-house is a software project, not a hire.
A referral from a professional firm is pre-qualified, arrives with records in some kind of order, and stays for years. Compared with digital advertising, content production and sales conversion, it is by a wide margin the cheapest way for us to acquire a client — and unlike an ad budget, it is only ever paid out of revenue that actually arrived. That is why the rates below are as generous as they are. We would rather buy distribution from firms that already have the relationship than from Meta and Google.
We are not an audit firm and we do not compete with you for audit, tax or company secretarial work. We do not sign, issue or file anything in your name. We do not approach your clients. Under the model we prefer, your client will never know we exist.
The simplest way to place us: we work as your firm’s junior staff, at the desk. And one thing we are not selling is the software. There is no licence and no subscription; the tooling is not a product and you cannot buy it. It exists to get the work done and to keep getting better at it, and the only way it gets better is by doing more of it — which is the whole of our interest here. That is the whole of it — so the engagement covers in-house bookkeeping only. We do not act as payment maker and we do not run payroll. At T1 and T2 we take no client contact and do not liaise with anyone outside your firm; selling, collecting and fielding the day-to-day questions stay where they belong, which is with you. Site visits happen only at T2, and only where the file needs them. Client liaison is a T3 add-on you choose, and even then it is done in your firm’s name. One distinction is worth drawing, because these two get confused: everything we produce goes out under your name and on your domain, with no LTT marking anywhere on it — your client never needs to know we exist. That is not the same as us speaking to them. We execute after your firm has spoken to the client, never instead of it. Within that boundary the offer is speed: for a partner firm we commit to ten days from complete documents, and for a firm we have worked with for a while — a file that suits us, records that arrive complete — five days is the norm. Not by working late, but because there is a standard process and internal QC on every file before it leaves. (Our own direct clients run on a six-week cycle. That six weeks is not the work — it is the talking around the work: telling a client what to prepare, waiting while they prepare half of it, waiting again because they are busy, settling what they expect, and the free extras that attach themselves to a direct engagement. Add paper to handle, an offline system to copy out of, and meetings to schedule. The rest is the documents themselves: the same simple page printed twenty times over, unrelated papers mixed into the box, and the queue — waiting for someone here to scan paper into files before the work can start at all, only to discover at the end of the scan that something is missing. Without tooling you then list it in a spreadsheet and sort it before the gaps even surface. With ours that stage is a few hours. Working firm-to-firm removes most of the rest, which is exactly why we would rather not take the client-facing side: giving it up is what buys the speed, and the records reaching us have usually already passed through a professional’s hands.) What stretches a file is document quality rather than volume: private spending mixed into the business account, items whose purpose nobody can state without asking, faded cash bills, foreign currency, accounts that need reconciling, a system that fights us, or a box of hardcopy. Both figures are committed where our tooling can carry the file, so they will not fit every client. When we cannot take one, that answer is fast too: normally three days, never more than seven, so you still have time to make other arrangements. That speed of answer has a reason behind it. This work is priced on transaction volume, and without tooling simply arriving at the number is most of a day: counting lines across bank statements, counting bills when a single scan may hold several, blank pages sitting in the batch, double-sided sheets and other page faults to resolve before anything can be counted at all. Our own tools take that stage — BillCleaver separates the bills inside one image, BinderReaper resolves the page and duplex problems — so counting the transactions takes about as long as producing the finished figures, on templates and a standard chart of accounts we already hold. The estimate costs us close to what the job costs. That is why a quote comes back in days rather than weeks. What separates 40% from 60% is not client contact but whether hardcopy, site visits and scanning are involved, and how much of the settling we carry rather than your own reviewer. Client contact is a separate matter: it is the twenty-point step to T3. Files are given to us by your firm; whether a file is feasible at the price you quoted and the scope you agreed is still ours to confirm.
The only question that matters is who owns the client relationship. Everything else — who invoices, who collects, who chases documents, who gets what share — follows from that one answer.
Under Model A you make one introduction and keep 30% for the life of the engagement; the client is ours. Under Model B at T3, LTT does the same production work and the same client liaison, and you keep 20% — but the client stays yours, in your name, on your engagement letter. Those ten points are the stated price of ownership. And the twenty-point step from T2 to T3 is the stated price of the front office: document chasing, query resolution, sales conversion and the day-to-day relationship. Both are priced openly, so you choose with the numbers in front of you.
Under Model B the base is 40% to LTT. Only two things move it: who handles the documents and the site, and who handles the client. Three tiers, each a clean step up. You can put different clients on different tiers, and move a client up or down at any renewal.
| Tier | LTT delivers | Your firm handles | LTT's share | Your share |
|---|---|---|---|---|
| T1Cloud & Softcopy | The whole bookkeeping job on a purely cloud file: classification, coding to the standard chart, posting into Bukku Cloud or the LTT finance app, bank reconciliation from statement files, subledgers, schedules, monthly trial balance and management accounts in audit-ready format. No hardcopy, no site visits. | Everything up to softcopy: document collection and every follow-up with the client, landed in a Google Drive, OneDrive, Dropbox or equivalent shared folder your firm runs. All client contact, the final review and the sign-off. | 40%30% at volume | 60%70% at volume |
| T2Hardcopy & On-site | T1 plus the physical records: collection, scanning, indexing and records management, and site visits where the file genuinely needs them. LTT also carries the settling — chasing every open item through to a clean close — so your review is a final read rather than a working pass. Still no client contact: queries come back to your firm. | Sales, quotation, engagement letter, fee collection, all client contact, and the professional sign-off. | 40–60%assessed per file | 40–60% |
| T3Client Liaison | T2 plus the client: liaison, document requests direct to the client, query resolution, sales conversion, marketing and the day-to-day relationship — all in your firm's name. Your client still sees your firm. | Ownership of the client and the engagement letter, your name on the fee, and the professional sign-off. | 80%60% + 20% liaison | 20% |
Volume is the only thing that moves T1 from 40% to 30%: once your files together pay LTT more than RM5,000 a month at the 30% rate (about RM16,700 of fees), measured on a rolling three-month average and trued up quarterly, the ten-point cut applies. That volume is what lets us schedule capacity instead of holding it idle, which is what the cut actually pays for. T2 is assessed per file: hardcopy records management we can usually carry within 60%, while on-site work is costed on the visits the file needs. T3 adds a flat twenty points on top of T2 for the front office. There is no volume discount at T2 or T3, because at those tiers our cost is people-time, not spare capacity.
LTT's share will not fall below RM250 per entity per month at T1, regardless of the percentage. Small files still cost a fixed amount to run. Every rate is also tested against the agreed fee at scoping: if the share does not cover the work, we say so before we take the file, not after.
Opening balance migration, backlog catch-up, chart conversion and historic clean-up are one-off work, quoted per file — not taken from the monthly percentage.
A named LTT team is assigned to your firm, with a shared job tracker and a monthly capacity planning call. Rate reviewed by agreement at that point.
A share of a fee only works if the fee reflects the work. On every tier LTT's share must be at least 30% of what LTT would itself have quoted that client for the same scope. At the 40% base rate that puts the client's fee at roughly three-quarters of our standard price. Below it we say so at scoping and the file is re-priced or declined — not a comment on your pricing, which is yours, just the point below which we are the wrong supplier.
The Model B percentages assume your firm holds the engagement: you quote, you invoice, you collect, and you stay between LTT and the client. Where the firm's part is only to put LTT in touch with the client, and LTT then carries the relationship, the queries and the exposure, that is Model A — 30% recurring to you, and LTT prices and engages the client itself. We do not carry client-facing responsibility on a Model B percentage. Settled per file before work starts.
A firm's fully loaded cost for one junior bookkeeper — salary, EPF, SOCSO, EIS, bonus, leave, training, supervision and desk — is a fixed monthly number that exists whether or not the files do. At T1 we convert that fixed number into a percentage of fee that only exists when a fee exists. Run your own figures: take one junior's loaded cost, the number of files they carry, and the fees those files bill. Compare the contribution with the same fees less 40%. We are confident enough in that comparison to put it on a public page.
Be honest about one thing when you run it: at every tier your review time does not disappear. What disappears is the fixed cost, the recruitment, the leave cover and the turnover.
The card rate is a standard rate for standard work. These are the boundaries. Anything outside them is not refused — it is quoted.
| Assumption | Standard rate applies when… | Quoted case by case when… |
|---|---|---|
| Platform | The books run on Bukku Cloud Accounting or the LTT standard finance app, and the client has no fixed preference of their own. | The client mandates SQL, AutoCount, Xero, QuickBooks, MYOB, Million, Odoo or similar. We can work in them; we lose the automation bridge, so the rate changes. |
| Record condition | Legible softcopy in a recognised format — PDF or image bank statements, supplier invoices, POS or sales summaries. | Handwritten or unreadable documents; consolidated multi-bill statements needing manual splitting; heavy mixing of personal and business spending; multi-currency; foreign records. |
| Reconciliation load | Ordinary monthly bookkeeping with a normal volume of banked transactions. | Heavy or forensic reconciliation, unbanked cash takings, intercompany reconstruction, backlog beyond one prior year, group consolidation. |
| Fee basis | Your normal in-house rate for work done remotely or at your own office, on records delivered to you or collected by you. | The fee already includes value-added services — those are stripped out of the base before the percentage is applied. |
| Excluded services | Tax computation and filing, audit fieldwork, company secretarial, advisory and CFO work, e-Invoice middleware build and statutory lodgement are not included at any tier. Each is separately scoped and separately priced if you want it. | |
Ten working days from complete records, defined in the engagement schedule so the clock only starts when the documents actually are complete. Queries raised within five working days of receipt, not at month end.
At every tier the final review, the adjustments and the professional sign-off are yours. LTT does not sign, issue, lodge or file anything in your name, and does not hold itself out as your firm.
One round of review adjustments is included in the tier rate. Systematic rework caused by source data materially different from what was scoped is re-quoted before it is done, never after.
Under Model B the client is yours. LTT will not approach, quote to or accept a direct engagement from your client during the term and for 24 months after. The same restriction binds you against LTT's clients.
A mutual NDA precedes any file. Data is held only in the systems named in the schedule, access is by named personnel, and everything is returned or destroyed on termination at your election.
LTT invoices the firm, not the client, and the share falls due whether or not that client has paid — collecting is the half the share pays you for. Tell us before the next month's work if a payment turns doubtful: most of a bad debt is the work done after it was already clear. Beyond 30 days overdue we may suspend work, but we never withhold records. A client who genuinely defaults is put to us and considered case by case. Under Model A the credit risk is ours.
Either side may terminate on 60 days' notice. LTT hands back every file in native, usable format — the client's books are the client's books. No data is held hostage to a fee dispute.
Nobody signs a framework agreement on the strength of a web page. Five stages, each with something to hand over at the end of it, and a real exit at every one.
Agree the principle and the model. Sign a mutual NDA. Pick the pilot files together — we work out which ones fit and where the problems are.
Three to five live files at card rate, no volume commitment, no framework agreement, no exclusivity. Measured against criteria agreed in advance — turnaround, review adjustments per file, your internal hours per file, and your realised margin against your in-house margin.
Sign the outsourcing framework agreement: rate card, service levels, volume commitment, data protection, non-solicitation, exit. The commitment period starts here, not before.
Migrate the existing book progressively. A named LTT team, a shared job tracker, your own onboarding pack under your brand, and a monthly capacity planning call.
Once the bookkeeping runs itself: audit-ready file packs that cut your audit hours, unaudited financial statements and audit-exemption compilations, company secretarial support tooling, and AI agents for client intake and CRM.
Every firm now says it uses AI. Very few can say what the machine does, where the human sits, and what stops a confident wrong answer from reaching the accounts.
AI does the keystrokes. Arithmetic proves the keystrokes. A person signs. Nothing in our pipeline is trusted because a model produced it — every number has to reconcile to something the client's own documents already say.
Every line carries a running balance, so every line carries its own proof. Where the chain breaks, the true amount is recovered from the balance movement; a statement is accepted only when the printed total debit and total credit agree to the cent.
Our naming standard puts each document's amount into its filename at intake, so every listing line is checked against the document it came from — two records of the same figure, produced by different routes.
Supplier statements, POS readings, gateway settlement reports and payroll summaries are third-party totals. We reconcile to those rather than to ourselves.
One published chart of accounts, one authority, administrator-only writes. What the pipeline learns on a client goes to a proposal queue, because most of it is client-specific and has no business becoming a practice standard.
The model that reads a figure is not the model that interprets the narrative, and neither decides whether a page was read correctly — arithmetic does. A page whose own totals fail is re-read automatically.
No output is issued on the strength of the machine. Where LTT works as your back office, that person is your reviewer, and the sign-off never leaves your practice.
Figures are from live engagements, anonymised — if we would name them to you, we would name you to the next firm. The full capability paper, including what the pipeline does not do, is available on request.
Ninety days, card rate, no commitment, no agreement to sign beyond an NDA. If the scorecard is not obviously better than doing it in-house, we shake hands and you have lost a quarter's worth of nothing.
