Writing

Field notes

Short pieces written for the practice's own clients and published as they are. They are working notes, dated 2026, kept here so a conversation on a Tuesday call has somewhere to point.

4 February 2026

The thirteen-week forecast is the only forecast a small company needs

Most owner-run companies already have an annual budget, or at least a tax projection their accountant built last spring. Those files answer a different question. They ask whether the year, taken as a whole, will land near a number you can live with. They do not tell you whether the bank account will clear payroll in the third week of March. Cash is a sequence of days. A year is an average of those days, and averages hide the weeks that hurt.

Thirteen weeks is long enough to see a gap forming and short enough that the inputs are still mostly known. You can name the invoices that will be paid, the suppliers who will be paid, the payroll runs, the rent, the GST or HST instalment, and the loan payment. Beyond that horizon the list becomes a set of hopes. People still publish twelve-month cash forecasts. They are useful as a conversation about shape. They are a poor place to park a decision about a facility or a supplier term.

The method is plain. Start with the bank balance on Friday. List receipts you can defend, by week, from open receivables and known retainers. List payments the same way. Keep payroll, rent, tax and debt on their own lines so a quiet sales week does not hide a statutory payment. Roll the close of week one into the open of week two. When the close goes negative, you have a date. The date is the whole point of the file.

Companies that do this well treat the forecast as a weekly habit, not a project. Each Monday someone updates receipts that arrived and payments that cleared, then looks again at week nine. If week nine has moved, they ask why. Often the reason is a single customer who slipped from net 15 to net 45, or a supplier who pulled a delivery forward. Those are phone calls, not strategy sessions. The forecast earns its keep by making the phone call happen while there is still a week to spare.

A common objection is that the business is too seasonal for a short window. Seasonality is exactly why the short window helps. A food wholesaler we have seen in composite work had a nine-day hole every February. The annual budget showed a profitable year. The thirteen-week file showed the same hole arriving on roughly the same date. The fix was terms, not a new operating line. You only find that if you are willing to look at February as February, rather than as a twelfth of a good year.

Another objection is that the numbers change too fast to bother. They do change. That is an argument for a short file you can rebuild in an hour, not for giving up. A model with forty tabs and a colour-coded dashboard will not be touched on a Tuesday morning. A single sheet with thirteen columns will.

We still build annual budgets. They have a job in October, when you are naming hires and leases. For the question “will we make payroll and the CRA instalment without surprising ourselves,” thirteen weeks is the file we open first.

Takeaway: keep a thirteen-week cash sheet current, and treat a negative week as a date you can still act on.

18 March 2026

Your best location is probably subsidising the other two

Combined profit and loss statements are a kindness. They take three clinics, three chairs, or three shops and present a single healthy margin. Owners like them because the year looks fine. The kindness has a cost. Rent, supervisor time, shared inventory and the hours of a specialist who travels between sites are sitting in a heap called overhead. The heap is then spread by a rule that felt fair when someone set it up, or it is not spread at all.

When we allocate those costs back to the locations that used them, the ranking often changes. The busy site that “carries the group” may be the only site that can carry its own rent and still contribute. The quiet site may be covering its materials and little else. This is ordinary. It does not mean the quiet site should close tomorrow. It means the story you tell about which location is best needs a second pass.

A workable method is to pick three costs that actually move with a site and assign them on a basis you can explain in a sentence. Rent goes to the site that signs the lease. Specialist hours go to the site where the hours were worked. Delivery or lab fees go to the site that generated the order. Shared administration can stay shared, with the method written down. Perfection is not available. A rough allocation you can defend is better than a combined page that no one can interrogate.

In an illustrative dental group of twenty-two chairs, the combined file showed a tidy year. Once rent and hygienist hours sat on each location, one site was covering about CAD 8,400 a month of cost the other chairs were not earning. The owners had been planning a fourth operatory at the busy site. After the allocation they kept the third site open, changed hygiene rostering, and delayed the extra chair. The busy site was already doing the labour of two. Adding capacity there would have raised rent and wages without raising the hours the quieter sites could fill.

Retail groups show the same pattern with inventory. The flagship store holds stock that the smaller shops draw from. If the flagship is charged for the whole warehouse and the smaller shops only show a transfer at cost, the flagship looks worse than it is and the smaller shops look like heroes. Reverse the charges and the conversation about which lease to renew becomes a different conversation.

The point of the work is not a league table. It is a set of decisions that match the economics. You might keep a quiet site because it feeds the brand, or because a partner lives nearby, or because closing it would dump patients onto a wait-list you cannot staff. Those are fine reasons. They should be named as reasons, not hidden inside a combined margin that makes every site look the same.

Do this once a year at least, and after any lease renewal or hire that is tied to one address. The spreadsheet is simple. The argument it starts is the valuable part.

Takeaway: allocate rent and the hours that actually travel, then decide about each site with those figures on the table.

6 May 2026

Reading a receivables ageing report in four minutes

Ageing reports arrive as long PDFs. Most of the lines do not need you. Four checks will tell you whether collections are a story you already know or a problem that has grown a new limb. The rest of the file can wait for the bookkeeper and the person who actually rings customers.

First, look at the total and at the share that is current. If current is still most of the book, the machine is working. If more than a third of the book has slipped past terms, you are financing your customers. That is a pricing and a process question as much as a collections question. A company on net 30 that is routinely paid on day 52 has given away three weeks of cash without putting it on an invoice.

Second, look at concentration. Sort by balance and read the top five names. If two customers are half the overdue book, the next action is two phone calls, not a new policy. If the overdue book is a hundred small balances, the next action is a process: statements on a fixed day, a stop-credit rule you will actually keep, and a person who owns the list. Concentration and scatter need different medicine.

Third, look at the oldest column. Balances that have sat in 90-plus for two cycles in a row are often not going to be collected in the ordinary way. They may be disputes, they may be related-party amounts, or they may be work that was invoiced before it was finished. Write them down as a separate question. Leaving them in the main total makes the current book look worse than it is and makes the old book look more alive than it is.

Fourth, compare this ageing to the one from last month and from the same month last year. A single report is a photograph. Two reports are a direction. If the 60-day column is growing while sales are flat, someone has stopped chasing. If the 60-day column is growing while sales are up, you may only be looking at a busy month that has not yet been collected. The comparison tells you which sentence is true.

Credits and unapplied cash confuse every ageing. A credit sitting in current against an invoice sitting in 90-plus is a bookkeeping problem, not a collections win. Ask for those to be matched before you hold a meeting about “the debtors.” Ten minutes of matching will remove a surprising amount of heat from the conversation.

We put a short ageing watch on the monthly pack: total, share current, top names, and anything that aged another bracket. The full PDF stays in the working papers. Four minutes is enough for the owner. The rest of the hour should go to the two or three names that would actually move cash this month.

Takeaway: read total, concentration, the oldest column, and the change from last month, then spend the hour on the names that would move cash.

23 June 2026

Price increases: how to model the volume you can afford to lose

Owners often talk about a price increase as a feeling. Customers will be upset. A competitor is cheaper. The last increase was only last year. Those sentences may all be true and still leave you without a number. The useful question is narrower. If we raise this price by this amount, how much volume can we lose and still be better off than we are today?

The arithmetic is older than any of us. Contribution margin is price minus the costs that move with the unit. After a price rise, contribution per unit goes up. You can then lose a certain number of units and still match the old total contribution. That number is the volume you can afford to lose. If you believe you will lose less than that, the increase is worth doing. If you believe you will lose more, the increase needs a different size, a different product, or more time.

Work a small example. A shop sells a service at CAD 180. Direct labour and materials are CAD 70. Contribution is CAD 110. A CAD 15 increase takes the price to CAD 195 and contribution to CAD 125. The shop can lose about 12 percent of volume and still match the old contribution (110 divided by 125). If the owner thinks three regulars out of thirty will leave, that is 10 percent. The increase still helps. If the owner thinks eight will leave, the increase is too large for this service, or it needs to be paired with a cheaper option so the eight have somewhere to go besides the door.

Two mistakes show up often. The first is using gross margin after overhead instead of contribution. Overhead does not leave the building when a customer does, at least not in the first quarter. Using the fully loaded margin makes the “affordable loss” look smaller than it is and talks people out of increases they could have lived with. The second mistake is applying one percentage to the whole list. A line that is already tight may need more than 4 percent. A line that is a traffic-builder may need none. The model should be run by line, or at least by the five lines that pay the rent.

Volume loss is not only customers who leave. It is also customers who buy less, or who move to a smaller package. Put that in the model as a mix shift if you can see it coming. A clinic that raises the comprehensive visit and holds the short visit will see hours move. The cash result depends on which visit carried the margin.

Write the assumption down. “We can lose 8 percent of this line and still be ahead” is a sentence you can check in ninety days. “Customers will understand” is not. After the increase, compare units and contribution to the model. If you lost more than you could afford, you have a fact, and you can reverse or reshape the next change. If you lost less, you have room you did not use.

We run this sheet as part of a margin read. It is one page. It does not need a pricing consultant. It does need the current price list, a honest view of direct cost, and a willingness to name the loss you would accept.

Takeaway: raise a price only after you have written the volume you can lose and still match today’s contribution.

11 August 2026

What a lender is actually looking at in your statements

Owners prepare for a bank meeting by polishing the profit and loss. Lenders read that page. They start somewhere else. They want to know whether the company can service the debt under a duller year than the one you just had, and whether the security still covers the facility if things go badly in a quiet way rather than a dramatic one.

Cash from operations is the first real page. Profit that sits in receivables or inventory is not yet available to pay a loan. A company that grew revenue 18 percent and grew working capital faster than that will look busy and still feel tight. The lender will ask why the cash line did not follow the sales line. Have an answer that names the customers or the stock, not an answer that names “investment in growth” as if that were a figure.

The second page is leverage and coverage. How much debt sits against earnings, and how many times those earnings cover the payments. Covenant definitions vary. Some add back owner wages, some do not. Some treat a related-party loan as debt, some treat it as equity because a parent has subordinated it. Read the definition in your facility letter before you build the schedule. A homemade ratio that uses a friendlier definition will be rebuilt by the credit team in an afternoon, and the rebuild will not favour you.

The third page is concentration and related parties. A book that depends on two customers, or a balance sheet that is tangled with a holding company, a spouse’s property, and an intercompany receivable, will draw questions. Those questions are fair. Write the structure down in a short appendix before the meeting. Surprise is expensive in a credit file. Clarity is cheap.

Quality of earnings is the fourth. One-off sales of equipment, a wage that was skipped, a grant, a large deposit recognised early: these are the items that make last year look stronger than the year you are walking into. Name them. A trailing-twelve that quietly includes a one-off will be discovered. Naming it yourself is the difference between a careful borrower and a file that needs another round of questions.

Security and personal guarantees sit beside the statements. We do not give legal advice on those. We do put the loan balances, the next test dates and the remaining room on a single schedule so you can see the same page the analyst will print. If a test is close, say how close and what you would do in the next quarter. Hoping the analyst will not notice a tight covenant is not a plan.

The pack we build for a lender is the monthly pack plus those extra pages. It uses the same close. It does not invent a second set of books for the bank. If the monthly numbers are honest, the bank pack is a matter of arrangement and a few more sentences. If the monthly numbers are still being argued about at home, the bank meeting is the wrong place to settle the argument.

Takeaway: take cash, covenants, concentration and one-off items to the meeting, written on the same close you use yourselves.

If a note here is the conversation you want to have

Write to us and name the decision. We will tell you whether a first read is the right next piece of work.

Write to us