Deal analysis is where investors lose money before they have spent any. The offer is a consequence of the numbers, and if the numbers are wrong the offer is wrong regardless of how well the conversation went.
What makes it difficult is not the arithmetic, which is simple. It is that three of the inputs are estimates and each one is systematically biased in the same direction.
The Four Numbers
Every deal reduces to these, whatever the strategy.
What it is worth in its current condition. Rarely the number that matters and useful as a floor.
What it will be worth once work is done. The after-repair value, and the input most likely to be optimistic, per how to calculate ARV.
What the work costs. The input most likely to be light, and the one that kills deals at a buyer's walkthrough, covered in estimating repairs.
What it costs to hold and transact. The input most likely to be forgotten entirely, set out in holding costs investors forget.
Get those four right and the offer follows mechanically. Get any one wrong by a meaningful margin and the deal was never what you thought.
The Bias Runs One Way
This is not random error, which is what makes it correctable.
Investors overestimate the after-repair value, because they want the deal. They underestimate repairs, because they have not opened the walls. They underestimate the timeline, because the plan assumes nothing goes wrong. They forget holding costs, because those accrue quietly rather than arriving as an invoice.
All four errors push the same way, which is why a deal that looked like it had margin turns out not to. The errors do not cancel.
The correction is not pessimism. It is building the bias into the process: a contingency on repairs, a conservative resale figure, a timeline with slack, and holding costs calculated on the longer timeline rather than the planned one.
Analysis Differs by Exit
The same property produces different numbers depending on what you intend to do with it, and analyzing for the wrong exit is a common error.
Wholesale. You are estimating what an end buyer will pay, which means their required margin is one of your inputs, worked through in how to price a wholesale deal.
Flip. Resale value minus full renovation minus holding minus selling costs minus your profit.
Wholetail. Light work and a retail sale, which sits between the two and is frequently the best answer for a property that is habitable but dated, detailed in wholetail and light rehab.
Rental hold. Priced on income and financing rather than resale, which is a different calculation entirely, per rental property analysis.
BRRRR. A hold funded by a refinance, where the binding constraint is what the property appraises for afterward rather than what you paid, explored in the BRRRR analysis.
Deciding the exit before analyzing is what prevents the common failure of running flip math on a property that only works as a rental, discussed in choosing the exit.
The Question the Analysis Is Answering
Investors run the arithmetic without deciding what it is for.
The analysis is not producing a valuation. It is producing a maximum offer: the highest number at which this deal still works for you, after everything.
That distinction matters because a valuation is a single figure and a maximum offer is a decision that depends on your cost of capital, your required return, your timeline and your risk tolerance. Two competent investors analyzing the same property will arrive at different maximum offers, and both can be right.
It also clarifies what to do when a seller wants more. You are not arguing about what the property is worth. You are explaining that above a certain number this stops being a deal for you specifically, which is a much easier conversation, described in when a seller wants more than you can pay.
Where the Estimates Come From
Each of the soft inputs has a method, and each has a failure mode.
Resale value comes from comparable sales, and the rule that does most of the work is matching condition rather than only size and location. A renovated comparable used to value an unrenovated property double-counts the renovation.
Repairs come from a scope of work rather than a per-square-foot guess, covered in scope of work for a rehab, and confirmed by contractor pricing where the numbers are large, per getting contractor bids.
Rent comes from actual comparable rentals rather than from an online estimate, set out in estimating rent.
Timeline comes from your own history rather than from optimism, which is why recording actual project durations matters.
Tools Help and Do Not Decide
Automated valuation and analysis tools have improved and they remain a screening instrument rather than an answer.
They cannot see condition, and that is the variable that moves a valuation most. They cannot know that the comparable three doors down sold under unusual circumstances. And they are trained on retail transactions rather than on the distressed ones you are actually pricing.
Used for triage they save real time. Used to underwrite they produce confident errors, worked through in real estate comps software and AI property analysis.
The Walkthrough Is Part of the Analysis
Investors treat the property visit as a social occasion and the analysis as something done afterward at a desk. The visit is where most of the analytical information is available.
What the roof looks like from the ground and from the attic. Whether the electrical panel is original. The age on the furnace label. Water staining, foundation movement, the grade around the exterior.
Photographing and recording those systematically produces a scope of work rather than an impression, the difference between an estimate that holds and one that gets corrected by a buyer.
The Numbers That Are Facts and the Ones That Are Guesses
A distinction worth holding explicitly, because they deserve different treatment.
Facts. The purchase price, closing costs, the loan terms, the tax bill, the insurance quote. These are knowable and should be exact.
Estimates with narrow ranges. Comparable sales, rent for a known property type in a known area. Reasonably tight if the work is done properly.
Estimates with wide ranges. Repairs on an unrenovated property, timeline, and what the market will do over the holding period.
The mistake is presenting all three with the same confidence, which produces a spreadsheet that looks precise and is not.
The practical response is to run the wide-range items at their unfavorable end rather than at their expected value. A deal that works with repairs at the top of your range and the timeline at the long end is a deal that works, and one that only works at the midpoint is a bet.
Grading Yourself Afterwards
The habit almost nobody has, and that is the only route to accurate estimates.
After every project, record what you estimated against what it actually cost and how long it actually took. Then look at the pattern rather than the individual deal.
Most investors discover a consistent bias of a knowable size, at which point the correction is arithmetic rather than judgment. An investor who knows they run twenty percent light on repairs can simply add twenty percent, detailed in grading your own numbers.
Without that record, every deal is analyzed with the same bias as the last one and the error never closes.
What Makes an Analysis Defensible
Not precision. Being able to show the basis.
Which comparables, and why those. What the repair estimate consists of, line by line. What timeline it assumes. What it looks like if the exit takes longer and sells for less.
That last scenario is the one that separates a robust deal from a fragile one, and that is the analysis most investors skip because the base case looks fine, explored in when not to borrow.
The Analysis That Takes Ten Minutes
Most properties do not deserve a full analysis, and the screening pass is what protects your time.
Pull the tax record for ownership, assessed value and sale history. Look at the property from the street where imagery exists. Pull three comparable sales and check whether they were renovated. Apply a rough per-square-foot repair figure by condition band. Run the arithmetic at a conservative resale figure.
That is ten minutes and it eliminates most properties. What survives gets the full treatment: a written scope, verified comparables, and a timeline.
The error investors make is inverting this, either analyzing everything thoroughly and running out of hours, or analyzing nothing thoroughly and making offers on screening numbers.
The screen decides whether to drive. The full analysis decides what to offer, and the two should never be confused, per pre-qualifying before the appointment.
The First Three Changes to Make
If your analysis is currently an estimate in your head, three changes produce most of the improvement.
Write the scope of work down rather than holding a repair number. Add a contingency and do not spend it in the model. And calculate holding costs on a timeline longer than you expect.
Then start recording estimated against actual on every deal, because in a year that record is worth more than any tool you could buy.