14 min read
—
Resolve every duplicate sale record to one cited comp, normalize sales, lease, cap-rate, operating and financing comps, and put the inclusion rules in writing before IC.

Casimir Rajnerowicz
Content Creator
V7 Go
A smarter way to manage due diligence and underwriting
Two documents on the desk describe one sale of a grocery-anchored neighborhood center. Page 6 of the September 2024 appraisal reports a pending contract at $21,400,000, signed in August 2024. The comp sheet in a broker opinion of value (BOV) from 2026 lists the same center at $20.9 million, closed in November 2024. Real estate comps analysis has to settle which record the investment committee sees. The IC memo cites the BOV at a 7.40% cap rate, and a VP has the appraisal open at 6.90%.
The answer sits in the firm’s own files. The team bid on this center in 2024, ordered the appraisal and lost. Its lost-bid folder still holds the recorded deed, the closing statement and the inspection amendment that cut the contract by $500,000 for a roof replacement. Those three documents trace the 50 basis point gap to two measurements of one sale, taken at different moments on different NOI bases.
No licensed database holds that combination for this center. The deed is public. The amendment, the appraiser’s stabilized NOI and the rent roll the team verified during its bid exist only in the firm’s records. A record from those files counts as a comp only after it has been resolved to one property, one transaction and one cited source. An analyst then has to approve it against written inclusion rules. The center and every figure built on it here are hypothetical.
Why the IC memo and the 2024 appraisal price one sale 50 bps apart
The appraisal, the BOV and the deed describe one transaction, and the matching keys prove it before anyone argues about price. The deed carries the assessor’s parcel number (APN), the recorded date and the grantor and grantee. The appraisal repeats the address, APN and seller. The BOV gives only an address and a closing month.
Source and page | Document date | Price | Date type | Area basis | Cap rate and NOI basis | Matching keys present | Reason for difference |
|---|---|---|---|---|---|---|---|
Appraisal, pages 6 and 41 | Sept 2024 | $21,400,000 | Contract, Aug 2024 | 98,200 SF GBA | 6.90% on stabilized NOI of $1,476,600 | Address, APN, grantor | Predates the roof amendment. Stabilized NOI nets out vacancy and reserves |
Broker BOV, comp sheet (page 3) | 2026 | $20,900,000 | Closing, Nov 2024 | 96,400 SF GLA | 7.40% on in-place NOI of $1,546,600 | Address only | Interested-party source. In-place NOI |
Deed and closing statement, lost-bid folder | Nov 2024 | $20,900,000 | Recorded, mid-Nov 2024 | Not stated | Not stated | APN, recorded date, grantor and grantee, address | Reflects the $500,000 roof amendment |
Approved comp | Approved [date] | $20,900,000 recorded | Closing, Nov 2024 | 96,400 SF GLA | 7.40% in-place and 7.07% stabilized, each stored with its basis | All four keys matched | Roof amendment cut $500,000. Appraisal predates it. Analyst [initials], [date] |
The cap-rate gap splits into two parts once both rates sit on the recorded price. Restating the appraisal’s stabilized NOI over $20.9 million moves the rate from 6.90% to 7.07%. Swapping in the BOV’s in-place NOI, which carries no vacancy allowance or replacement reserve, moves it to 7.40%.
Step | Cap rate | Change |
|---|---|---|
Stabilized NOI over the $21.4M contract price | 6.90% | |
Stabilized NOI over the $20.9M amended price | 7.07% | +16.5 bps (price basis) |
In-place NOI over the $20.9M amended price | 7.40% | +33.5 bps (NOI basis) |
Price per square foot makes the two records look closer than they are. The recorded price over GLA gives $216.80, and the contract price over GBA gives $217.92. The gap is only $1.12 because the higher contract price is spread over the larger gross building area.
Cannon and Cole’s study of NCREIF sales from 1984 to 2009 found appraisals deviated from subsequent sale prices by 11% to 13.5% in absolute terms. Disagreement is the normal case, so the analyst keeps all three records and writes down why each gap exists.
The deed outranks the BOV on price because of who produced each document. Fannie Mae’s residential Selling Guide B4-1.3-07 accepts comp data from a party with a financial interest but requires verification with a party that has none. A BOV comes from a broker pitching for an assignment, so its price gets checked against the deed before approval. Each comp type repeats this resolution problem in its own form.
Real estate comps analysis by comp type, and the trap in each
Each comp type answers a different underwriting question and fails in its own way. IVS 105 says the market approach fits best where there are “frequent and recent observable transactions in substantially similar assets”. It also expects the valuer to document qualitative and quantitative differences between each comparable and the subject. Internal deal files often explain those differences.
Sales comps: contract price, recorded price and the date you adjust from
A sales comp proves what a buyer paid for comparable property rights on a known date. The trap is choosing the wrong date or the wrong price.
August 2024 contract at $21,400,000 → October 2024 roof amendment, minus $500,000 → mid-November 2024 closing at $20,900,000 → mid-November 2024 deed recorded.
For this center, the time adjustment should run from the October amendment, when the $20.9 million price was agreed. The August date belongs to a price that no longer exists.
Standard Appraisal Institute teaching applies adjustments in a fixed sequence:
Property rights conveyed.
Financing terms.
Conditions of sale.
Expenditures immediately after purchase.
Market conditions.
Location and physical characteristics.
The roof amendment belongs to the fourth element. The buyer paid $20.9 million for a center that needed a $500,000 roof. An appraiser comparing it with a subject in sound condition would add the roof cost back. The comp record therefore stores the recorded price with a roof flag. Our piece on AI in property appraisal covers the appraiser’s side.

Comp selection has always rested on the appraiser’s judgment, so the firm should keep a record of which comps its 2024 appraiser chose and why.
In residential appraisals, the FHFA found that time adjustments were applied to fewer than 13% of comparables between Q3 2018 and Q4 2021. For 2021 it estimated that adjustments were warranted on 64% of comparables, yet they were applied on only about 25%. The same habit lets a retail sale that closed 19 months ago reach an IC memo unadjusted.
Lease and rent comps: face rent against net effective rent
A lease comp proves what a tenant agreed to pay after concessions, and the trap is comparing face rents, or asking rents, with executed deals. Take a hypothetical 10-year NNN inline lease at $24.00 per square foot.
Item | $/SF over the term | $/SF per year |
|---|---|---|
Face rent, 10 years | 240.00 | 24.00 |
Free rent, 6 months | −12.00 | |
Tenant improvements | −35.00 | |
Leasing commission, 6% of $240 | −14.40 | |
Net effective rent, undiscounted | 178.60 | 17.86 |
The lease nets $17.86 per square foot a year before discounting, 25.6% below face. An analyst who uses $24.00 as market rent overstates the rent the landlord keeps by $6.14 per square foot a year. Those concession terms come from the lease file. Our comparison of lease abstraction software covers tools that extract them and check them against the rent roll.
Cap-rate comps: survey rates, transaction rates and the NOI under each
A cap-rate comp proves how buyers priced income, and it is only as good as the NOI and the date under it.
Rate | Source | NOI basis | Date | What it measures |
|---|---|---|---|---|
Survey range | CBRE Cap Rate Survey | Stabilized, per CBRE’s definition | Survey wave | Professional estimates |
6.90% | 2024 appraisal, page 41 | Stabilized | Sept 2024 | Appraiser’s conclusion against a contract price |
7.40% | Deed plus BOV | In-place | Nov 2024 | Recorded price over in-place NOI |
The survey row is a collection of estimates. CBRE’s US Cap Rate Survey gathers about 3,600 cap-rate estimates across more than 50 markets from 200 to 250 CBRE professionals in each wave. It defines its stabilized rate as stabilized NOI over acquisition price. A memo that cites a survey range as a comp presents professional judgment as a closed sale. Our post on how NOI is calculated covers the income side, and the real estate underwriting guide covers the cap-rate arithmetic.
MSCI has observed that US commercial real estate cap rates tend to follow corporate bond yields on a four-quarter lag. If that pattern holds, a November 2024 transaction rate may still be catching up with bond yields from late 2023. A comp set that straddles a rate move needs a flag in each record, even when every comp sits inside a 12-month window.
Operating comps: one center, two expense ratios
An operating comp proves what comparable properties cost to run, and the trap is comparing expense ratios built on different lease structures and recovery treatments. The hypothetical center spends $6.80 per square foot on operating expenses, recovers $6.10 from tenants and collects $16.74 in base rent. Two honest ratios come out of those three figures.
Operating expenses over gross revenue including recoveries, $6.80 ÷ $22.84 = 29.8%.
Non-recoverable expenses over base rent, $0.70 ÷ $16.74 = 4.2%.
Mix the two conventions in one column and the center looks expensive or cheap depending on the source. The comp record states the ratio’s numerator, denominator and recovery treatment, and the T-12 analysis guide covers expense normalization.
Financing comps: the same loan amount, two DSCRs
A financing comp proves what lenders would advance against similar income, and the trap is reading interest-only bridge, floating-rate and fixed amortizing loans as interchangeable. A hypothetical $14 million loan on the center, 67% of the $20.9 million price at 6.50%, produces two coverage ratios from one NOI of $1,546,600.
Loan structure | Annual debt service | DSCR | Debt yield |
|---|---|---|---|
Interest-only | $910,000 | 1.70x | 11.0% |
30-year amortizing | $1,061,874 | 1.46x | 11.0% |
The debt yield is identical because it ignores the payment schedule. The DSCR moves from 1.70x to 1.46x on amortization alone, which is the difference between clearing a 1.50x covenant and missing it. External sources rarely carry rate type, amortization, recourse and covenant package together. The firm’s own closed loans and term sheets are its best financing comps.
Retrieving approved comps in V7 Go, with the appraisal page attached
V7 Go runs the comp workflow against what the firm already knows. It resolves prior appraisals, BOVs, deeds, rent rolls and loan term sheets to properties and transactions. The 2024 center exists once, with all three source records attached. When a new grocery-anchored deal arrives, retrieval returns only approved comp records that match the subtype, submarket and date window the analyst sets. V7 Go works on evidence the firm owns or licenses and supplies no market data of its own.

The sample is a fund report, and a comp value works the same way, opening on the appraisal page it came from.
Each comp record follows the same defined steps:
Extract price, dates, area, NOI and cap rate from each appraisal or BOV. Cite every value to its page, such as the income approach on page 41.
Match records on four keys, which are APN, recorded date, grantor and grantee, and standardized address.
Flag conflicts, such as $21.4 million against $20.9 million, to the analyst with both source pages open.
Restate cap rates to a declared NOI basis in a Python step, which also computes net effective rent and price per square foot.
Route each proposed merge to the analyst, who approves or rejects it and writes the reason.
Write a typed comp record with every field, its basis and its citation.
The model behind the extraction step is one setting inside that step. The arithmetic runs in code, so 7.07% comes out of the same formula on every record. The analyst still judges whether the cited source deserves trust. The comparable analysis agent is a starting point for this route. The wider set of V7 Go workflows for real estate teams applies the same pattern to rent rolls, leases and offering memorandums.

The screenshot shows a fund fact sheet example, and the same extraction pattern maps an appraisal comp grid to typed comp fields.
The inclusion log: what the analyst accepts, rejects and explains
The inclusion log is the evidence that a person chose the comp set. Retrieval proposes candidates, and the analyst decides each one with a reason, a cited source and a date. Here is a hypothetical excerpt for the new grocery-anchored deal.
Comp | Decision | Reason | Source cited | Reviewer | Date |
|---|---|---|---|---|---|
C-01, grocery-anchored center, 96,400 SF | Include with time adjustment | Resolved from three records, arm’s-length, 19 months since closing, adjusted from the October 2024 amendment | Deed, closing statement, appraisal page 41 | [Analyst] | [Date] |
C-02, three-center portfolio | Exclude | Portfolio sale with an unallocated price | BOV comp sheet, page 4 | [Analyst] | [Date] |
C-03, unanchored strip, 31,000 SF | Flag | Wrong subtype, kept as rent evidence only | Appraisal rent comp grid, page 28 | [Analyst] | [Date] |
C-04, grocery-anchored center | Exclude | Related-party transfer between affiliated entities | Deed, grantor and grantee | [Analyst] | [Date] |
C-05, grocery-anchored center, closed 2025 | Include | Arm’s-length, same subtype, eight months old | Closing statement, prior IC memo comp grid | [Analyst] | [Date] |
The log also answers the VP’s original question in one line. The 7.40% in the memo traces to C-01, and its reason column states that the appraisal’s 6.90% sat on a pre-amendment price and stabilized NOI. Excluded records stay in the log too. The related-party transfer in C-04 will surface again on the next retail deal. The written reason keeps the next analyst from approving it by accident.
Every comp record also stores three timestamps:
The transaction date, taken from the deed or closing statement.
The date the record was collected.
The date it was last approved, with the approver’s name.
A comp approved in March 2025 under a 12-month age rule may fail the same rule on the next deal. The last-approved date tells the reviewer that the decision has to be made again.

The demonstration uses fund metrics, and an approved comp opens its source document the same way when a reviewer questions the price.
Commercial real estate comps after IC: screening, model inputs and hold reviews
An approved comp record gets read again long after the memo that first used it, and each later use reads different fields.
Field in the comp record | Screening | Model inputs | IC memo | Hold and disposition review |
|---|---|---|---|---|
Subtype, submarket, transaction date | Yes | No | Yes | Yes |
Price per SF with area basis | Yes | No | Yes | Yes |
Market rent and net effective rent | No | Yes | Yes | Yes |
Cap rate with NOI basis | In-place | Stabilized | Both | Both |
Reason for difference and citations | No | No | Yes | Yes |
Last-approved date | No | Yes | Yes | Yes |
Screening needs only enough to decide whether a new center deserves a full underwrite. When a center fails the price-per-square-foot screen, the analyst can open the comps that set the threshold and check their dates. The model reads market rent, net effective rent and the stabilized cap rate, each with its basis declared. That is why the record stores 7.07% and 7.40% as separate fields. The real estate financial modeling guide covers how those values land on an input tab.
The IC memo reads the cited comp grid and the reason column. The investment committee memo agent shows how a memo draft pulls cited facts from approved records. Hold and disposition reviews read the record against current marks two years later, when nobody on the deal team remembers the roof amendment.

One approved comp record sits in the workflow layer and feeds screening, the model and the memo without anyone keying it in again.
What licensed market data covers that your deal files never will
Internal comps cover only the submarkets and years in which the firm bid, and they lean on appraisals, which lag the market. Licensed sources such as MSCI Real Capital Analytics, the CBRE Cap Rate Survey, CoStar, CompStak and NCREIF supply breadth and current transactions.
Evidence the firm’s files hold | Evidence only licensed or market sources hold |
|---|---|
Amended contract price and the reason for the amendment | Submarket transaction volume |
Rent roll verified against leases during the bid | Sales the firm never saw or bid on |
Operating expenses confirmed through estoppels and the T-12 | Current survey cap rates by market and subtype |
Loan terms the firm achieved, with amortization and recourse | Lease comps across buildings the firm never underwrote |
The appraisal lag is measurable. The Counselors of Real Estate reported that office appraisal cap rates fell 139 bps below transaction-based cap rates in 2Q22. The gap grew wider in later quarters. A comp set built mostly from a firm’s own office appraisals would have carried that lag straight into a 2022 memo. Licensed records also help resolution itself. A transaction export can supply the APN and recorded date that a BOV comp sheet leaves out. The internal record then matches on stronger keys. For the center, a 7.40% transaction rate checked against a current survey range is a stronger pair than either source alone. Where a licensed record and an internal record disagree, the same resolution steps apply, and the analyst records which source wins and why. The real estate data room analysis guide covers how prior deal files get organized so they are available to check.
The comp inclusion rules your analyst signs before IC
USPAP sets no fixed comp counts or time limits, as the Appraisal Institute’s guide notes make clear. The thresholds belong to the firm and need a signature. The rules here are firm-policy examples for a grocery-anchored retail strategy.
Rule | Threshold | Evidence required | Exception approver | Signed (name, date) |
|---|---|---|---|---|
Age | Up to 12 months as primary evidence. 12 to 24 months with a documented time adjustment. Over 24 months excluded unless no alternative exists. Comps that straddle a rate inflection are flagged. | Dated deed or closing statement | VP, acquisitions | [Name, date] |
Distance | Same trade area or submarket. A wider radius needs a written rationale. | Source document or analyst-entered location | VP, acquisitions | [Name, date] |
Asset subtype | Grocery-anchored compared only with grocery-anchored | Cited property description | VP, acquisitions | [Name, date] |
Tenancy | Anchor credit and WALT within a band the strategy sets | Cited rent roll or lease abstract | Head of acquisitions | [Name, date] |
Transaction context | Arm’s-length only. No distressed, related-party, entity-level or unallocated portfolio sales. | Flag recorded in the comp record | IC chair | [Name, date] |
Adjustment magnitude | Written explanation when net adjustments exceed 15% or gross adjustments exceed 25% | Adjustment grid | Valuation lead | [Name, date] |
The adjustment-magnitude row borrows a residential screen. Fannie Mae’s Selling Guide asks for an explanation when net adjustments exceed 15% or gross adjustments exceed 25% of a comparable’s sale price. A commercial firm can adopt the same trigger as its own policy. Under these rules, the 2024 center enters the new deal’s comp set at $20.9 million, 7.40% in-place and 7.07% stabilized. It carries the roof flag and a time adjustment from the October 2024 amendment. The rules take effect on the date of the last signature. They are re-signed each January, or sooner if the 10-year Treasury yield moves more than 100 bps within a quarter.
What is the difference between a sales comp, a lease comp and a cap-rate comp in commercial real estate?
A sales comp records what a buyer paid for comparable property rights on a known date, and it supports price per square foot and value conclusions. A lease comp records what a tenant agreed to pay under an executed lease, and it supports market rent, concessions and rollover assumptions. A cap-rate comp records the relationship between a price and the income behind it, so it is only useful when the NOI basis is stated. One transaction can produce all three. The grocery-anchored center in the worked example is a sales comp at $20.9 million and a cap-rate comp at 7.40% on in-place NOI or 7.07% on stabilized NOI. Its rent roll can supply lease comps for the inline space. Each use needs its own normalization: the recorded price and the date the price was agreed for the sale, net effective rent for the leases, and a declared NOI basis for the cap rate. Store them as separate fields with separate citations so a reviewer can check each one.
+
Why do an appraisal and a broker opinion of value report different prices for the same sale?
They usually record different moments or different measures of one transaction. An appraisal written while a deal is under contract reports the contract price, and any later amendment, seller credit or closing adjustment comes after it. A BOV prepared later often reports the closing or recorded price. Area can differ too, since one document may use gross building area and the other gross leasable area. The cap rate can rest on stabilized NOI in one and in-place NOI in the other. In the hypothetical example, a $500,000 roof amendment and a switch from stabilized to in-place NOI explain a 50 basis point gap. To settle it, match the records on APN, recorded date, grantor and grantee and address, then check the price against the deed and closing statement. A BOV comes from a broker with an interest in the assignment, so the recorded documents should outrank it on price.
+
Can a deal the firm bid on and lost be used as a comp?
Yes, and lost deals are often the best-documented comps a firm has. During the bid, the team usually collected an appraisal, a rent roll, a T-12, a property condition report and sometimes lender quotes. Once the sale closes to another buyer, the recorded deed confirms the final price and date. That combination shows the price, the condition the buyer accepted and the income behind it, which a database row rarely does. The limits matter. The firm's view of income dates from the bid, so the comp record should store the NOI basis and the date of each figure. The final price may differ from the last price the firm saw, so confirm it from the deed or a licensed transaction record. Confidentiality agreements signed during the bid may restrict how the material can be used, so legal should confirm the terms before a lost-deal file feeds a comp set shared beyond the deal team.
+
How old can a commercial comp be before it needs a time adjustment?
USPAP sets no fixed time limit for comparables, so each firm sets its own thresholds, and the useful ones are written down and signed. A common structure treats sales within 12 months as primary evidence, allows 12 to 24 months with a documented time adjustment, and excludes older sales unless nothing better exists. Age is only half of the test. A sale that closed eight months ago can still mislead if interest rates moved sharply in between, since cap rates tend to respond to bond yields with a lag. Comps that straddle a rate inflection deserve a flag even inside the age window. The starting date matters as well. If the price was renegotiated after inspection, the amended contract date is when the market set that price, and the time adjustment should run from there to the subject's valuation date. Record that date in the comp record so the next analyst does not adjust from the original contract.
+
Do AI real estate comps tools replace licensed comp databases?
The analyst who builds the comp set approves each inclusion and exclusion, and a VP or valuation lead signs the inclusion rules and approves exceptions. Transaction-context exceptions, such as using a distressed or portfolio sale, are worth escalating to the IC chair because they change what the comp proves. The comp record should show the property identity and its matching keys (APN, address, grantor and grantee, recorded date). It should also show the price with its date type, the area basis, the cap rate with its NOI basis, each value's source document and page, and the reason for any difference between sources. It should carry three dates as well: when the transaction happened, when the record was collected and when it was last approved. The inclusion log should keep rejected comps with their reasons, because the same portfolio sale or related-party transfer will come back on the next deal and the next analyst should see why it was excluded.
+
Who should sign off on comp inclusion, and what should the comp record show?
Go is more accurate and robust than calling a model provider directly. By breaking down complex tasks into reasoning steps with Index Knowledge, Go enables LLMs to query your data more accurately than an out of the box API call. Combining this with conditional logic, which can route high sensitivity data to a human review, Go builds robustness into your AI powered workflows.













