Why House Flips Lose Money: Budget vs Actual and the 7 Usual Causes

Why do house flips lose money? Rarely because of one disaster. Most lose it to three ordinary misses that arrive together: an ARV set from the best comp, a rehab figure that was a guess, and a schedule that doubled. On the example house a $35,000 profit becomes $19,960 with the ARV 5% lower, $24,440 with the rehab 20% higher, $30,500 with three more months, and $4,900 with all three at once. Here are the seven usual causes, and what a budget-versus-actual view does about them.

The arithmetic of a flip that went over budget

A flip's profit is a small number left over from large ones.

Profit = ARV − Price − (Rehab + Contingency) − Purchase costs − Financing − Holding − Selling

On the example house (1,600 sq ft, built 1978, bought for cash):

$35,000 is 10.9% of the ARV. A 5% move in the ARV is $16,000; a 20% miss on the rehab is $10,560 with contingency; three months of holding is $4,500. None of those numbers is unusual, and each one is a large share of the profit. That is the whole mechanism: house flipping risks are not exotic events, they are normal errors against a thin margin.

The 7 usual causes

1. An ARV set from the best comp

ARV (after-repair value) is what the house should sell for once the work is done, and the only way to know it is from comparable sales: similar houses, nearby, sold recently. The mistake is to pick the highest one. The best comp on the street had a bigger lot, a finished basement or a buyer who overpaid, and your house will be judged by the median, not the maximum. How to calculate ARV shows the range method: three to five comps, price per square foot, low, median and high, and a deal that works at the median.

2. Repairs estimated from a $/sq ft guess

A per-square-foot figure is a way to check an estimate, not a way to make one. "Medium rehab, $35 a foot, 1,600 sq ft, $56,000" is a range of roughly $40,000 to $72,000 wearing a single number. The actual cost is decided by the roof, the panel, the water heater, the kitchen and the baths, each of which is a line with a quantity and a unit price. A flip that went over budget by a third very often started as a per-foot guess that fell on the low side of its own range.

3. No contingency, or contingency spent on forgotten lines

Contingency is money for what you could not see: the rot behind the tub surround, the sewer line, the subpanel. The app's default is 10% of the repairs. The two failure modes are having none, and having 10% and spending it on dumpsters and permits that should have been lines of their own. When the concealed condition arrives, there is nothing left for it.

4. Months that double

The plan said six months. Permits took five weeks, the plumber had another job, the buyer's financing fell through and the house went back on the market in a slower season. Twelve months is not a rare outcome. Every month costs holding, and with a loan it costs interest too; on a $228,000 hard money loan at 11% that is $2,090 a month before taxes, insurance and utilities. Holding costs on a house flip works the full monthly figure.

5. Change orders without a reason

"While we are at it" is the most expensive phrase on a job site. A larger shower, a second coat of a better paint, recessed lights in the hall. Each one is small, none is written down, and at the end nobody can say where $14,000 went. The problem is not the change; sometimes it is right. The problem is a change with no reason attached, because a budget line that moved without a reason cannot be learned from on the next house.

6. Selling costs forgotten

Selling costs are paid out of the sale, so they are invisible until the settlement statement. Agent commissions and the seller's closing costs on the example house are $19,200 at a 6% default, the second-largest cost after the rehab. A flipper who counted "ARV minus price minus rehab" and called the rest profit has overstated it by roughly the price of a kitchen.

7. A fixed figure replacing an unfinished estimate

The contractor walks the house and says "fifty-eight six". The estimate had three lines still unpriced, so the round figure replaces it and the deal looks finished. But the bid did not include the roof, because the contractor does not do roofs, and the three unpriced lines are still unpriced; they have just disappeared from view. An estimate that is incomplete must stay visibly incomplete, even when a bid stands in for it.

The stress tests on the example house

Three tests, each changing one assumption, then all three at once. Selling costs are 6% of the ARV, so they move when the ARV does; everything else holds.

The arithmetic, line by line (price $200,000 and closing costs $4,000 in every row):

ARV −5%:       304,000 − 200,000 − 52,800 − 4,000 −  9,000 − 18,240 = 19,960
Rehab +20%:    320,000 − 200,000 − 63,360 − 4,000 −  9,000 − 19,200 = 24,440
+3 months:     320,000 − 200,000 − 52,800 − 4,000 − 13,500 − 19,200 = 30,500
All together:  304,000 − 200,000 − 63,360 − 4,000 − 13,500 − 18,240 =  4,900

Rehab +20% is $48,000 × 1.2 = $57,600, plus 10% contingency = $63,360. Three more months are 9 × $1,500 = $13,500.

Read the last row slowly. Nothing dramatic happened. The comps were a little optimistic, the rehab was a normal miss, the sale took one extra season, and the deal that looked like $35,000 is worth $4,900 before the first real surprise. With hard money the same row is a loss: $17,100 of points and interest for six months, more for nine. This is the test to run before the offer, not after the closing; the house flip profit calculator article shows the same table with financing in it.

How the original plan, the forecast and the actual stay apart

The stress tests protect the offer. What protects the project is keeping three numbers for every line and never letting them merge:

• Original plan: the budget and the schedule you bought on, frozen the day the project starts. It never changes again.
• Current forecast: the original, corrected with what you know today: the roofer's bid, the permit delay, the extra month on the schedule.
• Actual: receipts, invoices and the sale price. Money that has really moved.

A project managed from one number, "the budget", that gets edited whenever a bid comes in has no memory. By month four the budget says $63,000, nobody remembers it said $52,800, and the question "are we over?" has no answer. Three numbers side by side answer it instantly: original $52,800, forecast $63,360, actual to date $41,200.

Flip Investor OS stores the three separately and shows them together on the project overview, profit and rehab and schedule each with its variance. The original plan is frozen in the database, not only on the screen, so it cannot be quietly edited to match a bad outcome. A milestone that slips recalculates the forecast holding costs on its own, because months are money.

Every line that moved gets a reason

The second half of the discipline is the reason. In Flip Investor OS, when a budget line's forecast differs from the original, the line asks why, and the answer is one of a short fixed list: concealed condition, scope change, material price, code requirement, contractor, estimate was off, other. "What changed and why" on the project overview collects them. A line that moved without a reason stays visible as unknown and shows up in Needs attention, the list of things that change money, time or certainty.

This is not bookkeeping for its own sake. A flipper who finishes three projects with reasons on every variance knows something about the next house that no article can teach: that their kitchen estimates are fine and their electrical ones run 30% light, that concealed conditions on pre-1980 houses have cost them $6,000 each on average, that the one contractor's change orders are the problem. The app's history and insights screens are built from exactly these facts, estimate versus bid versus actual by work category, from the user's own projects and never from a single deal.

If you want to run your own numbers through the stress table, open the free calculator; the profit calculator is free, no card, and a saved deal analysis carries the stress tests, the frozen original and the reasons into the project.

What to do before the offer, in one list

• Take the ARV from the median of three to five comps, not the best one, and run the deal at the low end too.
• Build the rehab line by line, by zone, with a quantity and a unit price; use the per-foot figure only to check the total.
• Keep contingency for the unknown and give permits, dumpsters, utilities and staging their own lines.
• Count selling costs at 6% of ARV until an agent gives you a real figure.
• Run ARV −5%, rehab +20%, three more months, and all three at once. Decide what margin you need to survive the last row.
• Once the project starts, freeze the plan, forecast from bids and dates, record actuals from receipts, and write a reason on every line that moves.

None of this guarantees a profit. It does something more useful: it makes the loss visible in the deal analysis, where it costs nothing, instead of at the closing table.