Why Your Underwriting Never Gets Better Without a Feedback Loop
You projected a $295,000 ARV, listed at $299,900, and accepted $281,000 after 41 days on market. That $14,000 gap probably felt like bad luck. But if you never trace it back to the comps you picked at offer time, you will make the same call on the next house and the one after that.
Want to skip the spreadsheet? FlipSmrt pulls comps, estimates ARV and rehab, and runs the deal math from just an address. Your first 3 analyses are free.
Most investors run fast analyses but almost never compare them to closed results. They underwrite, buy, renovate, sell, and move straight to the next deal. So the same biases repeat on every offer: comps pulled from the nicer side of the subdivision, a kitchen budget that always runs light, or a four-month hold that keeps turning into six. Without a written record, those patterns stay invisible.
A post-mortem fixes this, but only if you capture the right baseline. At the moment you make the offer, snapshot the full analysis:
- ARV and the specific comps used to support it
- Rehab budget broken out by trade
- Projected hold months
- Buying, holding, and selling costs
- Your MAO and target profit
The rule that matters most is simple: never overwrite that snapshot. Save it, date it, and leave it alone. Then at closing, log the actuals line by line beside the original numbers. A $9,000 rehab overrun only teaches you something if you can see it came from electrical and flooring, not from the whole budget at once.
This is where a living spreadsheet fails you. You open the same file during the rehab, update the roofing quote, nudge the ARV after a new sale, and by closing the "projection" quietly matches reality. Saving or exporting a FlipSmrt report at offer time avoids that drift. You get a fixed, line-item baseline with the adjusted comps, trade-level rehab priced for the county, and the MAO math as it stood the day you offered. That frozen record is what you grade yourself against later.
Worked Example: Breaking a $24K Profit Miss Into Its Causes
Take a flip bought at $180,000. Your underwriting projected an ARV of $295,000, rehab of $48,000, buying costs of $3,700, five months of holding at $2,400 per month ($12,000), and selling costs at 8% of the sale price ($23,600). Subtract everything from the ARV and the projected profit comes to $27,700. On paper, that is a solid deal.
Here is what actually happened. The house sold for $281,000, rehab came in at $54,500, buying costs held at $3,700, and the project ran seven months instead of five, pushing holding costs to $16,800. Selling costs landed at $22,480 because 8% of a lower price is a smaller number. The math: $281,000 minus $180,000, $54,500, $3,700, $16,800, and $22,480 leaves an actual profit of $3,520. That is a $24,180 miss.
The useful part is splitting that gap into causes you can act on. Each line below compares projected to actual:
- ARV miss: the sale came in $14,000 below projection, but selling costs dropped by $1,120 ($23,600 minus $22,480). The net hit is $12,880.
- Rehab overrun: $54,500 minus $48,000 equals $6,500.
- Extra hold: two additional months at $2,400 equals $4,800.
Add them up: $12,880 plus $6,500 plus $4,800 equals $24,180, which matches the total variance exactly. Buying costs contributed nothing, so they need no adjustment. If your pieces do not sum to the total, you have missed a line item or double-counted one, and the analysis is not finished yet.
Now the miss has a shape. More than half of it (about 53%) came from the ARV, roughly 27% from rehab, and about 20% from time. That tells you where to look first on the next deal: your comp selection and adjustments, not your contractor. Without this breakdown, most investors simply blame the market and change nothing.
Worked Example: Turning Four Post-Mortems Into Calibration Factors for a New Offer
One post-mortem tells you what went wrong on one house. Four post-mortems tell you how you tend to be wrong, and that pattern is what you can actually correct for. Line up your last four flips and calculate three ratios for each: actual sale price divided by projected ARV, actual rehab divided by estimated rehab, and actual hold minus projected hold.
Here is what that looks like for a hypothetical investor with four completed flips:
- Flip 1: ARV at 95.3% of projection, rehab at 108% of estimate, hold 1 month long
- Flip 2: ARV at 97.8%, rehab at 115%, hold 3 months long
- Flip 3: ARV at 96.1%, rehab at 110%, hold 2 months long
- Flip 4: ARV at 96.8%, rehab at 115%, hold 2 months long
Averaged out, actual ARV ran 96.5% of projected, rehab ran 112% of estimate, and holds ran 2 months longer than planned. None of those misses looks dramatic on its own. Stacked together on every deal, they quietly drain profit from each flip.
Now apply them to a new deal with an analyzed ARV of $260,000 and a rehab estimate of $40,000. The unadjusted MAO is ($260,000 x 0.70) - $40,000 = $142,000. Calibrated, the ARV becomes $260,000 x 0.965 = $250,900 and the rehab becomes $40,000 x 1.12 = $44,800. That gives ($250,900 x 0.70) - $44,800 = $175,630 - $44,800 = $130,830.
The hold overrun still needs to come out. At $2,100 per month in loan interest, taxes, insurance, and utilities, two extra months cost $4,200, which brings your calibrated offer to about $126,630. That is roughly $15,000 below the unadjusted number. This lower offer is not pessimism. It reflects your real track record, and it is the number that would have protected your margin on each of those four flips.
Run these numbers on a real deal in under 5 minutes
FlipSmrt turns an address into ARV, a rehab budget, and your Maximum Allowable Offer automatically.
Analyze a Property FreeNuances: Separating Your Errors From Market Moves and One-Off Surprises
Before you blame your comps for a sale that came in low, time-adjust the projected ARV. Say you projected $295K in March and sold at $281K in October, and recent sales in that pocket of the neighborhood dropped about 3% over those seven months. Your market-adjusted projection is $295,000 x 0.97 = $286,150. That means your true comp error is roughly $5K, not $14K, and only the $5K belongs in your ARV calibration.
Rehab overruns need the same split. If you chose to add quartz counters or a second bathroom mid-project because buyers in that area expected them, that was a decision, not an estimating miss. Log chosen scope changes in their own column. A $6K upgrade you picked on purpose should never inflate the factor you apply to your next bid.
One-off surprises belong in contingency, not in a bias factor. A $9K cracked sewer line under the slab tells you to scope sewer lines on older homes and maybe raise contingency from 8% to 10%. If you fold it into a rehab multiplier, you will overpay for every future scope item. A new general contractor is a similar break in the data, so restart your rehab variance tracking when you switch crews.
Be careful with small samples. Two or three flips can't separate a real pattern from noise, because one bad deal can swing the average by thousands. A workable approach is to apply only half of any calculated factor until you have five or six comparable flips. Once that history exists, apply the full factor.
Finally, track the deals you passed on. Note your MAO, then check what each property sold for after its rehab six to nine months later. If houses you rejected keep reselling well above your projected ARV, your offers may be too conservative. That error costs you deals without ever appearing in a post-mortem.
Common Post-Mortem Mistakes That Bake In New Errors
Logging only final profit. A note that says "projected $38K, made $14K" tells you something went wrong, but not what. You can't fix an ARV assumption or a holding-cost estimate from one net number. The fix is to record projected and actual side by side for every bucket: ARV, rehab by trade, holding months, buying costs, and selling costs.
Editing the original projection. When a change order comes in, it's tempting to update the budget so the spreadsheet "stays accurate." Do that and the baseline disappears, and your miss shrinks to zero on paper. Freeze the projection the day you sign the purchase contract and save it as a separate tab or PDF. Track revisions in a new column so you can see both what you assumed and what actually happened.
Applying one overrun factor to every trade. Say rehab came in 18% over budget: $59K against $50K. If $7K of that $9K miss came from a panel upgrade and repiping, padding flooring and paint by 18% next time just inflates your rehab estimate and costs you deals. Calibrate per trade instead. Electrical and plumbing get the bump, and cosmetic lines stay where they were.
Stacking a factor on an already padded contingency. If you carry a 10% contingency and then add a 12% calibration factor on top, you're budgeting for the same risk twice. On a $50K rehab, 10% adds $5,000 to reach $55,000, and 12% on that adds another $6,600, for $11,600 of cushion. That much padding can drop your MAO enough to lose a deal that actually worked. Fold the calibration into the line items and keep contingency only for true unknowns, or keep the contingency and skip the factor.
Letting one bad flip swing the factors. A single house with hidden foundation damage can make every future offer too timid. Wait until you have three or four comparable flips before you set a factor, and flag one-off disasters so they don't dominate the average. Use the median miss across deals rather than the mean, because one outlier distorts a median far less.
FAQ: Running Post-Mortems on Your Flips
How many closed flips before calibration factors mean anything? Treat one or two deals as anecdotes, not data. Around four to six closed flips is where a pattern becomes worth acting on, especially if the misses point the same direction. If three of four projects ran rehab 10% to 12% over budget, that is a signal. If one ran 25% over because of a foundation surprise, it is noise until it repeats.
Should you track factors by price band or neighborhood? Start with one overall factor, then split it once you have enough deals to support the split. Price band usually matters first, because a $180K starter home and a $420K renovation attract different buyers with different finish expectations. Neighborhood splits make sense only when you flip repeatedly in the same few zip codes. Splitting four deals into four buckets gives you four guesses, not four factors.
How often should you reset them? Recalculate after every closed flip, using a rolling window of your last six to eight deals. Drop older projects so a stale sale price or an old contractor's pricing stops skewing today's numbers. Also revisit your factors whenever something structural changes, like a new general contractor or a new market.
Does this work for rentals and BRRRR? Yes. Compare projected rent, vacancy, repairs, and management to twelve months of actuals. If you underwrote $1,650 rent, leased at $1,575, and repairs run $90 a month higher than planned, projected cash flow was off by $165 a month, or $1,980 a year. For BRRRR, also compare your projected appraisal to the actual refinance value.
How does FlipSmrt help keep a consistent baseline? A post-mortem only works if every projection was built the same way. Running each lead through FlipSmrt produces the same structure every time: an ARV from adjusted sold comps within 0.5 miles, county-priced line-item rehab by trade, and an MAO, in under five minutes. Save that report, then compare it line by line to your closing statement and contractor invoices. Because the method stays constant, the gaps reflect the deal rather than how you built the spreadsheet that week, and your first 3 analyses are free.
Ready to analyze your next deal?
Get ARV, renovation estimates, and full deal math in under 5 minutes. Your first 3 analyses are free.
Start Free Analysis