Start from the debt service and build backwards
A lender does not read projections front to back; they go straight to whether projected cash flow covers the proposed payments, so build the model around that line from the start. The measure is debt service coverage: cash available for debt payments divided by the payments themselves, counting the new loan and everything the business already owes. A common benchmark is coverage of 1.25 times, meaning a dollar and a quarter of cash flow for every dollar of principal and interest, and a file that projects thinner coverage than the lender's threshold is declined arithmetic, however good the story.
Two adjustments separate the lender's coverage number from the one owners usually calculate. First, the cash available is measured after reasonable owner compensation, because the lender assumes you have to live; a projection that works only if you pay yourself nothing will be normalized until it does not. Second, all existing obligations count, including equipment loans, leases treated as debt, and the shareholder vehicle in the company's name. Compute coverage the way the credit department will, not the way the pitch deck wants to.
Build the payment line on honest financing assumptions. Use the rate the lender actually quotes for your profile, not a teaser, and for a floating-rate facility test coverage at a cushion above today's rate, because the lender's own stress test will. Include program and standby fees where they apply, and amortize the loan on the schedule being requested rather than an idealized one. If coverage only works at yesterday's rates, the model is telling you something about the ask.
Working backwards from coverage also disciplines the ask itself. If the cash flow the evidence supports covers a seven-year amortization but not a five-year one, the request should be structured accordingly before submission. The term should match the life of what is being financed: working capital on an operating line, equipment over its useful life, real property over a long amortization. Lenders read a mismatched structure as inexperience, and it is the cheapest thing in the whole package to fix.
This is also where the financing program landscape enters. Smaller asset purchases can fit the Canada Small Business Financing Program, where the federal government guarantees most of a loan used for equipment, leasehold improvements or real property, for businesses under the program's revenue ceiling. The projections for a guaranteed loan follow the same logic; the security conversation is what changes.
Lenders expect three linked statements, monthly for year one
The standard package is a projected income statement, cash flow statement and balance sheet, built monthly for the first twelve months and annually for two or three years after that, and the three must tie together arithmetically. Linked matters: net income flows to equity, collections and payments flow to cash, the new loan appears as principal on the balance sheet and payments in the cash flow. A set of statements that do not reconcile to each other tells the analyst the model was assembled by hand, and every number in it inherits the doubt.
The monthly grain exists because businesses fail in months, not years. An annual projection can show comfortable coverage while hiding three months where cash goes negative, and for seasonal businesses, contractors, landscapers, retailers, anyone with a slow quarter, the monthly view is precisely what the lender wants to see managed. Show the trough, and show the operating line or cash buffer that carries it.
Cash flow is the statement that does the work, because profit and cash are different animals on a monthly horizon. Revenue booked in March may collect in May; inventory is paid for before it sells; HST collected is not yours and HST remittances leave in lumps; and loan principal is a cash outflow that never appears as an expense. Growth itself consumes cash, since every new dollar of revenue drags receivables and inventory behind it. A projection that converts profit to cash correctly, month by month, is rarer than it should be, and lenders notice.
The monthly cash flow also sizes the operating line, often the second facility in the same application. The deepest cash trough in the projection, plus a margin, is the limit you should request, and lenders typically margin operating lines against receivables and inventory, so the projected balance sheet has to show enough of both to support it. Asking for a line with no trough analysis behind it invites the lender to size it for you, smaller.
The projected balance sheet is the statement owners skip and analysts do not. It shows leverage after the new debt, the working capital position through the year, and whether the equity base stays credible. If the lender's covenants will be tested on balance sheet ratios, this is where you find out in advance whether you pass.
The assumptions page is the part that actually gets read
Every line of the projection should trace to a written assumption, and every assumption to a piece of evidence, because the analyst's job is to test the assumptions rather than admire the totals. Revenue is the assumption that decides the file, and it should be built bottom-up, from units, jobs, contracts, capacity or customers, never as a smooth percentage growth on last year. A contractor projects from signed and probable-weighted backlog; a clinic from practitioners times appointment capacity times realistic utilization; a distributor from accounts and order history. Top-down growth percentages are the single most common reason projections get discounted.
Evidence is anything the lender can verify: signed contracts and purchase orders, historical collection patterns, supplier quotes for the equipment being financed, the signed lease behind the rent line, market rates behind wage lines, and utilization data from your own past results. Where an assumption cannot be evidenced, say so and be conservative; a visibly cautious guess builds more credibility than a confident one with nothing under it.
Costs deserve the same discipline in less space: tie variable costs to the revenue drivers at your actual historical margins, not aspirational ones. Step fixed costs when thresholds are crossed, since the second crew needs a second truck and the new location needs staff before it needs customers. Build in the full cost of the loan itself, interest and principal, plus insurance, professional fees and the payroll burdens on any new hires. Underpriced costs are as damaging to credibility as overpriced revenue.
Do not forget the government lines, because analysts look for them specifically. HST collected and remitted should flow through the cash flow on your actual filing frequency, payroll remittances should track the wage lines, and corporate tax instalments belong in any profitable year's projection. Models that treat pre-tax cash as spendable overstate coverage, and that is exactly the kind of overstatement a credit analyst is trained to find.
Tie the projections to your history, and normalize honestly
The first credibility test is a bridge: this year's actuals sit beside year one of the projection, and any jump between them has a named cause. If revenue has grown modestly for three years and the projection shows a step change, the analyst needs the reason in one sentence: a signed contract, a second location opening in month four, added capacity from the financed equipment. No named cause reads as hope, and hope gets discounted to history.
Normalization runs both directions and belongs in the open. If the owner has been underpaid, restate compensation to market before computing coverage, because the lender will. If last year carried one-time costs, a lawsuit, a move, a bad-debt writeoff, adjust them out and label them. Do the lender's adjustments before the lender does; every one they find first costs more credibility than it should.
The historical statements underneath matter as much as the projections on top. Most lenders want CPA-prepared statements, at minimum a compilation engagement, alongside filed tax returns for two or three years, and the projections must reconcile to those documents, not to an internal spreadsheet with different numbers. This page is one panel of a larger submission; what a lender-ready financial package contains covers the full set, and Compilation Engagements is where the historical side gets built.
Three years of history is the standard anchor, but a shorter track record is not fatal if the file admits it. A newer business leans harder on contracts, pre-orders, the owner's verifiable industry history and a larger equity contribution. What does not work is dressing a two-year-old company in five-year projections with hockey-stick confidence; the thinner the history, the more conservative the base case has to be to be believed.
Stress the downside, because projections become covenants
Submit a base case you would sign your name to and a downside case that shows the loan still gets paid, because the lender will run a downside anyway and it is better to author it than to receive it. The stress case cuts the assumptions that matter most: revenue lower by a plausible slice, collections slower by a month, the biggest customer smaller, margins compressed. If coverage survives, the case argues for approval better than any paragraph of prose. If it does not survive, you have found the real conversation: more equity in the deal, a longer amortization, a smaller ask, or security that lets the lender live with the risk.
There is also a reason not to inflate the base case: the projections you submit become the yardstick you are measured against after funding. Covenants, coverage ratios, leverage limits, reporting undertakings, are typically set off the projected statements, and lender reporting continues for the life of the facility: annual CPA-prepared statements, often interim statements, sometimes a fresh budget each year. Beating a sober projection builds the relationship that funds the next stage; missing an aggressive one starts covenant conversations no owner enjoys.
Expect the projections to be tested in due diligence, not merely read. Analysts verify against filed returns and notices of assessment, bank statements, aged receivable and payable listings, CRA account balances for HST and payroll, existing loan and lease agreements, and appraisals or quotes on what is being financed. Discrepancies between the model and the documents are found in this pass, which is why the package is assembled as one consistent whole rather than as documents from three different eras.
Presentation is part of the credibility. One workbook, with the three statements linked, the assumptions on their own page and the base and downside cases labelled, reads as management that understands its own numbers. Figures pasted between documents until they no longer agree read as the opposite. The analyst treats the state of the model as a proxy for how the business itself is run.
Leave time for the process itself. Commercial adjudication runs in weeks, longer where appraisals, environmental reports or program registrations are involved, and a package assembled under deadline pressure shows it. Starting the projections when the equipment is quoted, rather than when the deposit is due, is the cheapest credibility upgrade available.
| Component | What the lender does with it | The common mistake |
|---|---|---|
| Monthly cash flow projection | Tests debt service coverage through the weak months | Annual totals that hide a negative quarter |
| Assumptions page | Verifies each driver against evidence | Percentage growth with no named source |
| Historical statements and returns | Anchors the projection; prices the jump | Projections that do not reconcile to filed numbers |
| Projected balance sheet | Checks leverage, working capital and covenant room | Omitted entirely, or missing the new loan |
| Downside case | Confirms the loan survives a slow year | Submitting only the optimistic case |
What changes the projections, and how we build them
Six facts shape how the model should be built before a single cell is filled in:
- What the loan is for. Equipment, premises, working capital and acquisitions each imply a different structure, term and evidence set.
- Revenue visibility. Contracted or recurring revenue supports a tighter case than walk-in trade, and the assumptions page should trade on it.
- Seasonality. The deeper the trough, the more the monthly cash flow and the operating line matter.
- Existing debt and leases. Coverage is computed on everything, so the current obligations schedule shapes how much new debt fits.
- Owner draws. The compensation the household actually needs is a hard input, not a plug to make coverage work.
- The security on offer. Strong collateral buys patience on coverage; thin collateral means the cash flow case must stand alone.
Acquisition financing raises the bar on all of this. The projections have to present the target's historical results, normalized for the seller's compensation and one-time items, then layer on the acquisition debt and any vendor take-back to show combined coverage. Lenders also want the transaction structure visible in the model: price, equity injected, each financing layer and its servicing. A business purchase file is a projections file first and a negotiation second.
We build these files from the lender's side of the desk. Walla Assaf spent years in banking and corporate finance before founding the firm, which is rare among CPAs and shows up in how the package is framed: coverage computed the credit department's way, the ask structured before submission, the weak months addressed rather than hidden. The projections are one piece of business financing support for owner-managed businesses, and it is worth knowing what lenders need before approving business financing before you assemble anything. If you want the model, the package and the lender conversation handled by a business financing and projections CPA in Ontario, Financing Support starts with a free 15-minute discovery call, with the scope and fee in writing before work begins.
