Every offer you make rests on one number: what the property will be worth once it is fixed. Get that number right and the rest of the math is arithmetic. Get it wrong and no amount of negotiating skill saves the deal, because you were solving the wrong problem from the first phone call.
After repair value, usually written ARV, is your estimate of the resale price of the property in restored condition, in its actual market, today. Every word in that sentence is doing work, and most bad ARVs come from quietly ignoring one of them.
ARV Is an Estimate of a Buyer's Behavior
The most common framing error is treating ARV as a property fact waiting to be discovered. It is not. It is a prediction about what some future buyer will pay, and the only evidence available is what similar buyers recently paid for similar houses.
That is why comparable sales carry the whole weight. Not listings, which are asking prices and therefore opinions. Not automated valuation estimates, which are useful for triage and were never designed to underwrite a purchase. Closed sales, because those are the only prices anyone actually agreed to.
Holding that distinction changes how you read the numbers. A house listed at a high price for five months is evidence that the price is wrong, not evidence of value.
The Comp Rules That Do the Work
Comp selection is where ARVs go wrong, and the rules are conventional for a reason. They are approximations of how an appraiser will look at the same property later, and your ARV is much more useful when it anticipates that.
Start tight and only loosen when you have to. Within roughly half a mile in a suburban market, and considerably tighter in a city where value can change block to block. Sold within the last six months, stretching to a year only if inventory forces it and only with an eye on whether the market has moved since.
Match the physical characteristics that buyers actually price: square footage within about ten to fifteen percent, the same bedroom and bathroom count, comparable lot size, similar age and construction style. A three bedroom is not a proxy for a four bedroom, and finished basements and garages are priced very differently across markets.
Above all, stay inside the boundaries buyers care about. School attendance zones, subdivisions, and the far side of a highway or rail line are all real price walls that a radius search will happily cross without telling you. Half a mile that crosses a boundary is worth less than a mile that does not.
Three to five genuinely comparable sales beat a dozen loose ones. When they cluster tightly, you have a defensible number. When they scatter widely, that scatter is information: the neighborhood is not homogeneous and your confidence should drop accordingly.
Adjustments, and Their Limits
No comp matches exactly, so you adjust. If your comp has an extra bathroom, subtract what a bathroom is worth in that market. If yours has a bigger lot, add.
The discipline is to adjust for the handful of features that materially move price and stop there. Once you find yourself making six adjustments to force a comp to fit, the honest conclusion is that it is not a comp. Every adjustment is an estimate layered onto an estimate, and stacking them produces a number with a false air of precision.
Repair condition is the adjustment people most often miss. If your comps sold in renovated condition and you are estimating a renovated resale, they align. If your comps sold as distressed properties, they are telling you about the as-is market rather than the ARV. Mixing the two is one of the most expensive mistakes in this business.
Where the Number Comes From Matters
Public records, paid data services, and agent-pulled MLS data all give you sold comps, and they do not all give you the same picture. MLS data carries the photographs and remarks that tell you what condition a property sold in, which is precisely the thing a raw public record cannot tell you.
If you are working remotely, this gap widens. Virtual wholesaling depends heavily on having someone who can confirm what the data cannot show, whether that is an agent, a contractor, or a boots-on-the-ground partner. An ARV built entirely from spreadsheets in an unfamiliar market is a guess wearing a suit.
Automated estimates from consumer real estate sites deserve one specific use: fast triage on a large list to decide which properties merit real analysis. They are not an underwriting tool, and their published accuracy varies enormously between dense urban markets and thin rural ones.
Ask the Buyers Who Will Actually Buy It
The most underused ARV check is free. If you are wholesaling, the people who will buy the contract are pricing this exact neighborhood every week.
Sending a property to a few trusted buyers and asking what they would pay produces a market-tested number rather than a desk-derived one. It also surfaces local factors no dataset contains: the street everyone avoids, the flood history, the pending development. That is one of the compounding benefits of maintaining a real cash buyer list rather than assembling one when you already have a contract.
Where the number should live is on the deal record itself, alongside the comps you used and the date you ran them. ARVs age, and an estimate from four months ago in a moving market is a liability if nobody remembers when it was calculated. The case for keeping deal data on one record is in the guide to the real estate investor CRM.
Once you have a defensible ARV, the offer follows from it, which is where the 70 percent rule comes in, and why it is a starting point rather than an answer.
This is one piece of a larger map. The full breakdown of off-market lead sources is in the guide to motivated seller niches.
Treat scatter as information. When your comparable sales cluster tightly you have a defensible number; when they disagree widely, that disagreement is telling you the neighborhood is not homogeneous and your confidence should fall accordingly.