How to Run Comps Like an Investor

Comps decide everything downstream. The after-repair value you pull from comparable sales feeds the maximum offer, the rehab decision, the buyer's underwriting, the lender's appraisal, and eventually the resale price. A sloppy comp set does not just produce a wrong number — it produces a wrong number that every later decision trusts. Learning to run comps properly is therefore the highest-leverage skill a new investor can build, ahead of negotiation, ahead of marketing, ahead of everything that happens after the value is set.

This guide teaches a repeatable method: define the subject clearly, pull a wide set of recent nearby sales, filter ruthlessly, adjust carefully toward the subject, weight the best evidence, and land on an ARV with a range and a written rationale. It also covers the traps — renovated comps for tired houses, asking prices dressed up as values, and geographic shortcuts that cross lines buyers actually care about. Practice this method on houses you do not intend to buy until the estimates start landing close to eventual sale prices. That calibration loop is how judgment is built.

1. Define the subject before you search

Write down the subject's facts first: address, beds, baths, living area, lot size, age, construction type, garage, and — critically — the condition it will be in when the value is supposed to exist. For a flip or wholesale ARV, that is finished condition. For an as-is value, it is current condition. Mixing the two is the most common beginner error: pulling renovated sales to value a house you intend to sell unrenovated. Note the neighborhood boundaries buyers actually use, including school zones and obvious physical dividers, because your search radius must respect them. A precise subject definition turns the comp search from browsing into measurement, and it gives you the checklist every candidate sale will be judged against.

2. Pull wide: recency, distance, and similarity

Start with sold properties — never listings — from roughly the last six months within about a half-mile to a mile in a city, widening in suburbs and rural areas where sales are sparse. Keep houses within a reasonable band of the subject's size, beds, baths, and age. Use public records, MLS access if you have it through an agent partner, and reputable data sources; Wise Revival's own comp work uses real sold data for the same reason your underwriting should. Sold data in non-disclosure states can be partial, which is a reason to widen carefully and weight agent and appraiser input, not a reason to substitute asking prices. Copy every candidate into a worksheet with price, date, distance, and key features before judging any of them.

3. Filter ruthlessly: most sales are not comps

Remove non-arm's-length transfers (family sales, foreclosure auctions, related-party deeds) unless you are specifically studying them. Remove houses that are a different product: new construction against 1960s stock, condos against detached houses, a different school zone across the freeway. Remove sales so old or so far that the market they prove no longer exists. What you want surviving is three to five sales a skeptical buyer would accept as evidence. If fewer survive, widen one dimension at a time — distance first in uniform neighborhoods, time first in distinctive ones — and record that the evidence is thinner, which will matter when you choose where in the range to price.

4. Adjust toward the subject, modestly and consistently

Adjust each comp's price for differences from the subject: living area at a consistent per-square-foot figure drawn from your market, beds and baths at consistent amounts, garage, lot, and condition. Adjust the comp, never the subject — you are estimating what the comp would have sold for if it were the subject. Keep individual adjustments small; a comp needing huge adjustments is a comp to discard, not to torture into agreement. Condition adjustments deserve special honesty: the gap between a tired house and a renovated one is the rehab budget, and double-counting it in both the comp adjustment and the repair line quietly invents margin that will not exist at resale.

5. Weight the best evidence and land on a number

Rank the adjusted comps by similarity, recency, and proximity, and let the best two or three dominate the conclusion. State the ARV as a point for underwriting plus a range for honesty: tight in a tract subdivision with abundant sales, wide in a mixed or rural area. Write two or three sentences of rationale naming the comps you weighted and why. That paragraph, attached to every offer and buyer packet, does more for your credibility than any claim about experience. When an appraiser or buyer later lands near your number, note it; when they land far away, study why. Calibration is a record, not a feeling.

6. Use listings and pendings for what they actually prove

Active listings are not value evidence, but they are competition evidence: they show what a buyer can choose instead of your house, and at what ask. Pendings hint at where the market is moving. Use them to sanity-check an ARV — if similar houses sit unsold at your ARV, the number is theoretical — and to set selling strategy in Step 12 of the Learn path. Never present a listing price to a seller, buyer, or lender as a comparable sale. The moment your evidence mixes opinions with transactions, every number you produce becomes negotiable in the worst way: people stop believing the real ones too.

7. Know when the comps say to walk away

A comp set can honestly conclude three uncomfortable things: the value is lower than the seller needs, the evidence is too thin to support any confident value, or the finished value cannot carry the rehab the house requires. All three are walk-away signals, and honoring them early is profit protection, not timidity. The deals that damage beginners are rarely the ones they analyzed badly — they are the ones they analyzed correctly and pursued anyway because they had time invested or a seller they liked. Let comps kill deals at the desk. The market will kill the survivors at resale, at a much higher tuition.

Worked example

Where beginners go wrong on this — and the fix for each

Three mistakes account for most of the damage we see on this topic. First, skipping the written step. Operators do in their heads what this guide does on paper — a value, an estimate, a decision — and the head version quietly edits itself mid-negotiation. The fix is boring and total: one worksheet, filled in before the conversation that could change it, re-run only when a fact changes. Second, borrowing confidence from other people's results. A stranger's market, exit, and risk tolerance are not yours, and copying their numbers without your inputs is how sound methods produce unsound deals. The fix is to treat every outside example — including the worked example above, which is labeled illustrative for exactly this reason — as a structure to fill with your own verified local numbers. Third, doing this once and calling it learned. The first attempt at any skill in this guide will be your worst one. The fix is reps with records: do it on practice properties, log the estimate against the eventual reality, and let the gap teach you. Within a handful of logged repetitions, your judgment stops being borrowed and starts being calibrated — and calibrated judgment, not any single technique, is what this entire library is trying to build in you.

Your first week with this guide

Do not try to master this in a weekend; instead, schedule five small, finished reps. Day one, re-read this guide and write the three numbers or decisions it asks for, in your own words, on one page. Day two, run the linked tool — the 70% Rule Calculator — on a real property currently listed or recently sold in your market, and save the inputs. Day three, repeat on a second property and compare what changed and why. Day four, find the gap in your own knowledge the reps exposed — a term, a data source, a local rule — and close it in the glossary or the Learn step for this subject. Day five, write the one-page version of your process that future-you will follow, and file the week's worksheets together. Five finished reps teach more than fifty pages of highlighting, and they produce the paper trail your second week builds on. When the week is done, continue to the Learn step matched to this guide — the path orders these skills the way a deal does, and the next skill assumes this one is practiced, not just read.

How we run the numbers: every formula and data source behind this guide is documented on How We Run the Numbers. Where a figure comes from an outside source, the page says so beside the figure.

Frequently asked questions