Hold back 10 to 20 percent of the decorating total as a contingency, and go to the higher end in an older house where nobody can honestly tell you what's under the carpet or behind the plaster. Then guard a second, quieter leak that no percentage protects you from: the samples, the extra gallon, the hardware swapped twice, the return labels, the restocking fee. Those are the two places a decorating budget actually fails, and they need to be managed as one problem. Headline purchases rarely break the number. The number breaks in the gaps between decisions.
Come sit at the kitchen table with me on a Tuesday night. Laptop open. A gallon ziplock bag of receipts under one elbow. A fan of paint lids drying on a paper towel like poker chips, and three tile samples leaning against the fruit bowl because I still haven't looked at them in the right light. This is where I do the reconciliation, and it's the moment a budget tells the truth, months before the room does.
The top of the spreadsheet always looks fine. The floor, the fixture, the sofa: those got decided early and written down the day they got decided. It's the bottom of the sheet that bites. A sample pot. Return shipping on a box of tile that looked warm on a screen and gray in the hallway. A second gallon because the first one pulled greener than the lid. Individually, you would never call any of it a decision. Collectively, it's a line item nobody budgeted for, and in most decorating projects I've watched, it turns out to be the biggest uncontrolled line there is.
The reserve is for unknowns, not for upgrades
Ten to twenty percent is the working range in remodeling cost guidance. The reason it's a range is that the range is really a statement about the house, not about taste. A newer build with straight walls and standard dimensions sits at the low end. A hundred-year-old house with plaster, an unknown subfloor, and trim that was never quite square sits at the top of it, and sometimes deserves more.
What matters more than the exact number is what you allow yourself to spend it on. Contingency absorbs unknowns: the surface that turns out to be unsalvageable, the run of material that comes short, the thing behind the wall that has to be handled before anything pretty can happen. It does not fund upgrades. The moment you use the reserve to get the better faucet or the wider plank, you've spent your protection on a want, and the first genuine surprise now comes out of the room itself. Something else gets cut, or the project stalls half finished.
So the rule I keep is simple. Upgrades come out of the original allocation, or they wait for the next phase. If the upgrade doesn't fit inside the number you already have, the honest answer is that you can't afford it yet, and the reserve is not a loophole.
The leak that has no label on it
Here's what I notice at the bottom of the sheet. The drifting spend is almost all below the threshold where anyone asks for permission. There is a number, different in every household, under which a purchase doesn't require a conversation. A sample pot is under it. Return shipping is under it. A second box of cabinet hardware in a different size because the first size didn't line up with the existing holes is under it. Nothing that small gets discussed, so nothing that small gets counted.
Multiply under-the-threshold by four months and you have the real story of the budget. It isn't that one big thing went wrong. It's that forty small things went right through without anyone writing them down.
The fix is boring, which is exactly why people skip it. Give the small stuff its own line and its own cap: samples, shipping, returned goods, extra paint, fasteners, the second set of switch plates. Track it like a utility bill. When the cap is spent, you stop sampling and start deciding, which is the real purpose. Sample pots are supposed to end with an approved finish, not a shelf of maybes.
Two columns, not one
Most people track what they've paid. That's the wrong number.
What matters is committed spend: every deposit placed and every order confirmed, whether or not the card has moved yet. An order that's confirmed but not shipped is money that's gone. A deposit is money that's gone. If you only watch the cash, the sheet can look comfortable right up to the week when three orders land at once and you realize you crossed the line a month ago.
So keep two columns and add them. Paid. Committed. Total against the budget. Then subtract the reserve you promised yourself you'd protect, and look at what's left. When that figure goes negative, you are not getting close. You are already over, and you found out while there was still room to change something, which is the entire point of looking.
Fifteen minutes once a week, same night, same table. That's the discipline. It has never been a software problem.
Changing your mind pays twice for the same surface
Rework is one of the most common causes of overruns, and it's the one people describe as bad luck. It isn't. It's a decision made in the wrong order.
Change a finish mid-project and you pay for it twice: the material you already bought and the material that replaces it, plus labor to undo and redo, plus disposal or a restocking fee, plus the days the project sits still. Same surface, two purchases. That's the arithmetic.
Almost all of that is avoidable before the first order goes in, and it's avoidable with the cheapest tool in the whole project, which is the actual room at the actual hour. Take the tile off the screen and hold it against the trim at four in the afternoon, with the lamp you're actually going to own turned on. Look at the paint sheen on a real wall instead of a chip. Put the fabric sample on the floor where the sun lands at seven in the morning. If it fails in the room, it fails in the room, and you would much rather find that out for the price of a sample than for the price of a reorder, a restocking fee, and a week of nothing happening.
What happens to the part you don't spend
The reserve you never touch isn't leftover money, and it shouldn't be treated as found cash in the final weeks. That's when the last stretch of a project gets finished: the trim paint touch-ups, the outlet covers, the blinds that actually fit the window, the small repairs on the punch list. None of it is exciting. All of it is what makes a finished room feel finished instead of stopped.
The reserve that saves you is the one still sitting there at the end, waiting for the unglamorous stuff. Which is why I would rather hold back a little too much and hand some of it back than hold back too little and run dry in the last two feet.
How much to hold back does depend on scope. One room of paint and a rug doesn't need the same cushion as a bathroom full of plumbing you can't see. But the shape of the problem is the same either way: the reserve covers the unknown, the tracked small spend covers the drift, and the decisions you make slowly, in real light, before anything is ordered are what keep both of those small.
