April Is Too Late

My accountant called me in March a few years back to tell me I’d had a great year.

Then he told me what I owed. The number was roughly thirty thousand dollars higher than it had to be.

Nobody made a mistake. Every decision that could have moved that number had already been made, and almost all of them got made in the last ten weeks of the previous year, while I was busy closing deals and telling myself I’d deal with taxes when it was tax season.

That’s the part nobody says plainly. Tax season isn’t when you do tax planning. Tax season is when you find out what your tax planning was.

By April you’re a historian. You’re documenting decisions a guy made in October and November, and that guy was you, and he wasn’t thinking about it.

So today we think about it. You have about fourteen weeks before the wall.

Two things before we start, and I mean both.

I am not an accountant. None of this is advice for your situation, because your situation depends on your entity, your state, your income, whether you have employees, and roughly nine things I don’t know about you. What this is, is the list of conversations worth having and a calendar telling you when to have them. Take it to the person who signs your return.

Second, none of this is about paying less than you owe. It’s about not paying more than you owe because you were busy in November. Those are very different things and only one of them is legal.

Most owners think of taxes as one deadline in the spring. There are actually two kinds and only one of them is in the spring.

Funding deadlines are generous. You can often put money into things well into the following year, sometimes as late as October with an extension.

Decision deadlines are not. They close December 31 and they do not reopen. No extension, no letter, no exception.

Almost every dollar you’re going to save runs through a decision deadline. That’s the whole reason April is too late. In April you can still write checks, but you can’t go back and be a man who existed on December 30.

If you’re buying anything substantial for the business this year, the rules right now are about as good as they’ve been in a decade.

For 2026, Section 179 lets you expense up to $2,560,000 of qualifying property, with the phase out starting at $4,090,000 of total purchases. And 100 percent bonus depreciation is back and permanent for qualified property acquired after January 19, 2025. Between the two, most of what a normal business buys can come off this year instead of dribbling out over seven.

The difference that matters for planning: Section 179 is limited by taxable business income and can’t create a loss, while bonus has no dollar cap and can. Section 179 gives you precision, asset by asset. Bonus gives you volume.

Now the phrase that costs people real money.

Placed in service.

Not ordered. Not paid for. Not sitting on a truck somewhere between a warehouse and your building. Placed in service means installed, operational, and available for its intended use. A machine that arrives December 28 and gets wired up January 6 is a next year deduction, and the man who wrote the check in December is going to be extremely unhappy in March.

If there’s a purchase in your plan, the question to ask your vendor this week is not when it ships. It’s when it will be running. Then work backward.

One more piece people get wrong. Vehicles have their own ceilings. For 2026, a heavy SUV between 6,001 and 14,000 pounds is capped at $32,000 of Section 179, with the balance eligible for bonus, and passenger vehicles are capped far lower. All of it requires business use above 50 percent, recalculated annually, with recapture if it drops later. If your plan for a tax deduction involves buying a truck you don’t need, that isn’t planning. That’s shopping with a permission slip.

This is the one I see missed most often, and it’s the one that stings, because it’s free money left in the road.

A solo 401(k) generally has to be established by December 31 of the plan year. Not funded by then, established. The funding can come later, in some cases as late as your extended filing deadline. But if the plan doesn’t exist on December 31, the door is shut for that year.

The 2026 numbers: $24,500 in employee deferral, $72,000 total between deferral and employer contribution. Add $8,000 in catch up if you’re 50 or older, which puts the ceiling at $80,000. If you happen to be 60, 61, 62 or 63, the enhanced catch up is $11,250, which takes you to $83,250 and then goes away at 64.

Compare that to a SEP IRA, which has no employee deferral at all and tops out at the employer piece. The SEP can be opened and funded as late as your extended filing deadline, which makes it the only option for a man who wakes up in April. The solo 401(k) reaches a much bigger number at the same income, because of that deferral.

The practical read: if you don’t have a plan and you’re going to have profit, open the solo 401(k) before December 31 even if you have no idea yet what you’ll put in it. An empty plan that exists is an option. A plan you meant to open is nothing. If you have employees this is a bigger conversation and it starts in October, because plan documents and census data take time.

If you’re on cash basis, and most owners reading this are, you have more control over the calendar than you think.

Income counts when you receive it. Expenses count when you pay them. That gives you two levers.

You can push income into next year by invoicing in early January instead of late December. You can pull expenses into this year by prepaying things you were going to buy anyway. Insurance. Rent. A year of software. Contractor work performed and paid now.

The rule that keeps this honest is that the expense has to be real and the benefit generally can’t stretch too far out. Prepaying twelve months is ordinary. Prepaying three years and deducting all of it today is a different conversation and a worse one.

Here’s where men get it backwards. Deferring income only helps if next year’s rate is the same or lower. If this year is mediocre and next year is going to be a monster, you want the opposite: income now at the lower effective rate, deductions next year when they’re worth more.

So the first question isn’t “how do I lower this year.” It’s “which of my next two years is going to be bigger.” Then you decide which direction to push.

Run that with actual numbers. Build the two year picture, not the one year picture. I run the scenarios through a model before I take them to my accountant, mostly so I show up with a question instead of a shrug, and I use Galaxy so I can push the same set of numbers past two or three different models and see where they disagree. Where they disagree is exactly what I ask the CPA about.

If you run an S corporation, the number on your own W2 is a decision, and it’s a decision with two opposing forces pulling on it.

Lower salary means less payroll tax. Higher salary means more room in the retirement plan, since the employer contribution is a percentage of wages, and it can also matter for the qualified business income deduction if you’re anywhere near the wage limitations.

It also has to be reasonable compensation for the work you actually do. That’s a real standard, not a suggestion, and it’s one of the most commonly examined items in a small business return.

The Section 199A deduction is permanent now, and for 2026 the phase in ranges for the wage and property limits widened to $150,000 for married filing jointly and $75,000 for everyone else. Practically, the band where your salary decision changes your outcome got wider, so the math is worth rerunning even if somebody ran it two years ago under the old ranges.

The move is not to guess. The move is to put the question on your accountant’s desk in October, while payroll can still be adjusted, rather than in February when the W2 is already filed.

This one is underused because it got complicated, then uncomplicated, and most owners stopped paying attention somewhere in the middle.

Domestic research and experimental costs are fully deductible again instead of being spread over five years, and software development counts as research and experimental expenditure, which surprises people. If you built internal tools, customer facing software, or spent real money developing a product or a process this year, that’s worth a specific conversation. Foreign research still has to be amortized over fifteen years, so where the work happened matters.

Smaller businesses also had a window to amend 2022 through 2024 returns for amounts capitalized under the old rules. If nobody has looked at that for you, ask. Refunds for prior years are the rare kind of tax planning that pays you for work you already did.

Your fourth quarter estimated payment is due January 15. Miss it or underpay across the year and you get a penalty, which is just an expensive way to find out you weren’t paying attention.

The way out is the safe harbor. Generally, if you pay in at least 100 percent of last year’s total tax, or 110 percent if your income was above a certain threshold, you avoid the underpayment penalty regardless of how big this year turns out to be. That’s the whole point of it. You don’t have to predict a great year perfectly. You have to cover a number you already know.

Now the trick, and it’s a good one.

Estimated payments are credited when you make them. Withholding from a paycheck is treated as if it were paid evenly across the entire year, no matter when it actually came out.

So if you’re an S corp owner and you get to December and realize you’re short for the year, you can run additional withholding through a December payroll and it’s treated as though you’d been paying all along. It can clean up an underpayment that a January estimated payment cannot.

That’s a December move. It does not exist in April. Which is the theme of this entire piece.

Three things almost nobody claims, all of them unglamorous, all of them real.

Bad debt. If you’re on accrual and you recognized revenue from an invoice that will never be paid, and you’ve genuinely given up on it, it may be deductible. You need documentation that you tried. Collection emails, a write off decision, a date. On cash basis you never recognized it, so there’s nothing to write off, which is its own answer.

Dead inventory. Product that is genuinely worthless or obsolete may be written down. Not “it’s slow.” Worthless. The treatment depends on what you actually do with it, so this one needs your accountant specifically.

Abandoned projects. Money spent on something you formally shut down. Not paused, not resting, not circling back to in the spring. Abandoned, with a decision and a date attached.

The pattern in all three is the same. The deduction depends on a decision you made and can prove, and most owners never make the decision. They leave things in a state of vague maybe, and vague maybe is not a tax position.

This is the part to actually put on a calendar today.

This week. Pull a year to date profit and loss. Not a vibe, the actual report. Estimate where you land December 31. Everything below depends on that number existing.

By October 15. Meet with your accountant. Not a phone call, a working session. This is the single highest value item in this piece and roughly nobody does it, because most owners think of their CPA as a filer instead of an advisor and then wonder why they only get filing.

By November 15. Make the entity and compensation decisions while payroll can still reflect them. Place equipment orders with a confirmed install date before year end. Start plan documents if employees are involved.

By December 15. Open the retirement plan account even if it isn’t funded. Run the prepayment decisions. Decide, in writing, on bad debts, dead inventory, and any project you’re abandoning.

By December 31. Everything with a decision deadline is done. Equipment is installed and running. Plans exist. Payroll is run.

January 15. Fourth quarter estimate, using the safe harbor number your accountant gave you in October.

Send the agenda in advance. Three lines, and it changes the entire quality of the hour:

“I’d like to spend an hour on the next ten weeks rather than last year. I’ll bring a year to date P&L and my estimate for December 31. What I want out of it is a safe harbor number, a recommendation on entity and salary, and a list of anything with a December 31 deadline that applies to me.”

Then record it. You’re going to get forty minutes of detail about depreciation schedules and reasonable compensation, you’ll retain roughly none of it, and by November you’ll be guessing at what was said. I run Fathom on calls like this so I have the transcript instead of three vague notes and a feeling.

And if the reason you’re dreading that meeting is that your books are a pile of receipts in a glovebox, fix the pipe rather than the pile. I’ve got a Make scenario that takes anything landing in a specific email folder, drops the attachment in the right drive folder, and logs it to a sheet with the vendor and the amount. It took an afternoon to build and it permanently ended the March scramble in my operation.

Everything above is mechanics, and mechanics are worth money. But here’s the reframe.

You spend weeks a year fighting for margin. Chasing a price increase. Negotiating a vendor down. Squeezing two points out of an ad account.

A tax decision made in October is worth the same dollars as a deal closed in October, except it doesn’t require anybody else to say yes.

There’s no prospect to convince, no competitor, no procurement department. It’s you, a P&L, and a professional, for one hour, on a Tuesday. It might be the single least competitive dollar available to you all year.

And the only way to lose it is to do nothing until spring, which is exactly what you did last year and the year before that.

Your accountant is not going to call you in October. He’s got four hundred clients and he’s going to talk to the ones who call him. Be one of those.

Make the appointment today. Before you close the laptop.

Refined. Relentless. Unapologetic.

Marcus