Real estate underwriting is how an acquisition team turns a deal packet into projected cash flow, returns, and a file someone else can review. A flyer can start that work. It cannot finish it. The rent roll, the trailing twelve-month statement, and the offering memorandum still have to be read, and the assumptions still have to be named, before a return is useful.
Create an account and start a property report
This guide is for multifamily and commercial acquisition teams, brokers, owner-operators, and fund analysts. It covers the process, the inputs, the line between results and assumptions, and the review controls that should pass before a person relies on the model. It is not a mortgage guide, not a loan-committee checklist, and not investment, tax, or accounting advice. An evaluation can be wrong. Check it against the source documents.
Search results for the phrase often mix two different jobs. One is a lender deciding whether a borrower qualifies for a home loan. The other is a buyer building a model of a property. This page is the second job. If you need the document stack that sits beside the model, read the document review guide. If you need how a checked model becomes a decision packet, read real estate investment analysis.
What real estate underwriting is for
In an acquisition setting, real estate underwriting means making a projection of how a property might perform, then showing the path from the files to that projection. The output is a model. The point of the model is a conversation the team can audit, not a verdict the software hands down.
Three questions sit under almost every deal:
- What is the property producing now, in the files you actually have?
- What are you assuming will change, and who chose that number?
- What has to be true for the price, the debt, and the exit to still make sense?
Career guides often teach this as a model sequence. You enter property facts, in-place income and expenses, operating assumptions, debt, and a sale, then test value against a target return. That sequence is a useful skeleton. It is not a substitute for knowing which page each number came from.
Investors and partners tend to trust a file they can retrace. A polished summary with no source trail asks them to take the analyst's word for it. A model that names the rent roll page, the T-12 line, and the assumption owner asks a narrower question: is this input still right?
Underwriting is also not an appraisal. Appraisal practice has its own standards, including the Appraisal Institute Standards of Professional Practice. A buyer's model can inform an internal discussion. It does not replace an appraisal, a lender review, or a legal close.
Separate in-place results from assumptions
The most common modeling mistake is treating a hope as a fact. In-place results are what the documents say happened or what is contracted now. Assumptions are what you think will happen next. They belong in different places in the file, with different owners.
| Layer | What belongs here | Typical source | Review question |
|---|---|---|---|
| In-place income | Unit mix, contract rents, vacancy, and concessions. | Rent roll. Leases when you have them. | Does the model match the roll, not the flyer? |
| In-place expenses | Operating costs for the trailing period, by category | T-12 or operating statement | Is each line classified the way the team defines it? |
| Operating assumptions | Rent growth, expense growth, stabilized vacancy, other-income changes | Team policy, not a broker pro forma | Who set the number, and is the reason written down? |
| Capital and exit | Reserves, renovation budget, hold period, exit cap rate | Scope notes and the team's assumption table | Is the budget itemized, or is it one round number? |
| Debt | Rate, proceeds, interest-only period, amortization | A term sheet, or a labeled placeholder. | Is this a quote, or a guess marked as a quote? |
Keep the broker's pro forma out of the in-place columns. A marketing package is allowed to show a cleaned-up future. Your model should show the trailing period first, then the changes you are willing to defend. If a line in the seller's pro forma sits far from the trailing average, flag it. Do not silently adopt it because the deal then looks better.
The rent roll is one input, not the whole process. For what a roll contains and how deal teams read one, use what a rent roll is. For the fields worth checking before they enter a model, use the rent roll template notes. This page does not repeat those guides.
A practical sequence for the file
Teams work at different speeds. The order below is a review sequence, not a promise that a deal takes a set number of days. Skip a step only when you write down why it does not apply.
- Name the asset and the question. Property type, unit count or square footage, location, and the decision in front of you: screen, bid, or pass. A one-page flyer can open this step. It should not be the only file in the folder.
- Lock the in-place inputs. Enter income and expenses from the rent roll and the T-12. Note missing months, odd categories, and any total that does not foot. Leave a blank rather than invent a line.
- Write the assumption set in one place. Growth, vacancy, other income, reserves, renovation scope, hold period, and exit cap rate. Each line needs an owner. "Broker said so" is not an owner.
- Add debt as a labeled case. If you have a term sheet, cite it. If you do not, mark the rate and proceeds as a placeholder. Do not let a placeholder look like a commitment.
- Run a small set of cases. A base case, a case with a wider exit cap rate, and a case with slower rent growth are usually enough to see if the result is brittle. Stop adding cases once the story is clear.
- Hand the file to a second reader. That person checks inputs against sources. They do not retype the model from scratch unless a source is wrong.
This sequence stops at a reviewable model. It does not size a loan, write an investment-committee memo, or tell you to buy. Those are later jobs, and they fail when this sequence was skipped.
Review controls before anyone relies on the model
A control is a check with a named owner and a source. It is not a vibe that the model "looks right." Acquisition teams that skip controls tend to find the error in the meeting, which is the most expensive place to find it.
Source tie-out
Every in-place figure in the model should point at a page, a tab, or a line in a source file. Unit mix and in-place rent tie to the rent roll. Operating expenses tie to the T-12. If the offering memorandum disagrees, the disagreement stays visible. You do not average the two numbers and move on.
Scans and handwriting need a slower read. A model can extract a number from a poor image and still be wrong. A person has to look at the page. The same is true when a column is cut off or a total does not match the lines above it.
Classification check
Expense names are not standard across owners. One seller's "contract services" is another seller's payroll plus landscaping. Before you compare a deal with the last one, confirm that categories mean what your team thinks they mean. Reclassify in the model and keep the original label in a note.
Assumption ownership
Growth rates, vacancy, and the exit cap rate should come from the team's current table unless someone writes an asset-specific reason. A number that appears only because it makes the deal clear a hurdle is a review failure, even if the spreadsheet runs.
Second reader, then the deal lead
The second reader checks inputs against documents. The deal lead checks whether the assumption set matches team policy and whether the risks are named. Those are different jobs. Asking one person to do both at the end of the day usually means the source check gets skipped.
A practical close looks like this:
- In-place income and expenses match a named file, or the gap is written down.
- Assumptions have an owner and a date.
- Debt terms are either cited to a quote or marked as placeholders.
- At least one downside case is in the file.
- A second person has looked at the sources, not only at whether the file opens.
None of those checks is a professional opinion. They are the minimum before a teammate treats the output as a starting point for a price talk.
Where Slung fits, and where it stops
Slung is an underwriting product for multifamily acquisition teams. You upload a flyer, an offering memorandum, a rent roll, or a T-12, or you email those files to sam\@slung.ai. Documented uploads are Excel, CSV, and PDF. Exports from systems such as Yardi, RealPage, AppFolio, Entrata, MRI, and Rent Manager are examples, not a promise that every file type will parse.
A flyer can open a property report. It does not fill in a rent roll or a T-12 you do not have. After extraction, you review line items, classifications, and unit mix against the source, then change assumptions such as cap rates, growth, financing, and renovations. The web model can be exported to an editable Excel workbook for that deal. That workbook is not the free v53 template, which is a blank starting file and not the app.
Email to Sam is an entry path. New senders may be asked to create an account, and email underwriting can depend on the plan. Plans named on the site are Free Analyst, Principal monthly, and Principal annual. A trial exists only when one is offered. Public pricing is call for pricing. See the pricing page and the underwriting docs before you assume a feature is on every tier.
Slung does not decide the deal. It does not certify the numbers. It does not replace a second reader. If you are comparing tools rather than walking the process, the separate note on automated underwriting software covers what to test in a demo. Security questions belong on the security page. That page says the SOC 2 examination is not complete, so this article does not claim a certification.
Teams that want workflows built around their own files and process, rather than the underwriting product alone, start with Slung Enterprise. That is a scoped working relationship. It is not a package you can infer from this guide.
What this process does not decide
A finished model is still a model. It can support a screen, a bid range, or a pass. It cannot tell you the property is a good investment, that a lender will approve debt, or that the tax treatment will match the spreadsheet. Those judgments sit with the people accountable for them, and with the advisors they hire.
It also does not finish diligence. Title, survey, environmental questions, and the lease file can change the story after the cash-flow model looks stable. Use the model to see the financial question clearly. Use the rest of the file to see whether the financial question was the right one.
If the sources are thin, say so in the model. A partial T-12, a rent roll with blank lease-end dates, or an offering memorandum with no support should stay marked as partial. Filling the gaps with round numbers makes the file look finished. It does not make it reviewable.
If you want the product path, book a demo as a request, or create an account and upload the files you already have. Either way, check the output against the source documents before you rely on it.
Frequently asked questions
What is real estate underwriting?
For an acquisition team, real estate underwriting is the process of turning property documents into a projection of cash flow and returns, with the assumptions written down. It is not the same job as a mortgage lender's credit review, and it is not an appraisal.
What inputs does a deal team need before the model is useful?
A flyer can start the file. A useful model still needs in-place income, usually from a rent roll, and in-place expenses, usually from a trailing twelve-month statement. Debt terms and an exit case can be labeled assumptions until you have a quote and a team policy for the cap rate.
How is underwriting different from due diligence?
Underwriting builds and reviews the financial model. Due diligence tests the broader file, including leases, title, and whether the marketing package matches the documents. You can run them in parallel. You should not treat a finished model as a finished file review.
Who should review the model before a price talk?
Someone other than the person who built it should tie in-place figures back to the source documents. A deal lead can then check that assumptions match team policy. One person rushing both checks at the end is how source errors survive.
Does Slung replace that review?
No. Slung can extract figures from a flyer, OM, rent roll, or T-12 and hand back a model you can edit and export to Excel. You still have to correct classifications and unit mix, especially on scans, and you still have to decide whether the assumptions are yours. The output is not financial, tax, or investment advice.
