Most stock screens run one model over every company, usually a discounted cash flow or a price-to-earnings ratio, and then produce a number that looks absurd on a bank or a REIT. The problem is not the math. It is that a bank, a landlord, a chipmaker, and a health insurer make money in ways that break a single model in different places. This guide is the map: the method MiniValuator uses for each industry it recognizes, what that method still cannot see, and the industries it does not yet handle well. For the full reasoning on any one row, each links to its own guide.
One thing belongs at the top, because it frames the whole map. Picking the right method for an industry tells you whether a stock is fairly priced for what it reports today. It does not tell you whether those reported numbers are about to turn. Every industry has a risk that lives in the filings rather than the data feed, a loan book going bad, a lease about to expire, a cycle about to roll over, and no screen reads those. So treat this as a map of methods and their blind spots, not a claim to see everything.
Why One Model Does Not Fit Every Industry
A discounted cash flow model assumes a company produces free cash flow you can forecast and discount. That holds for a software company and breaks for a bank, whose cash flow mixes lending and funding into a number that means little, for a REIT, whose depreciation buries the real cash its buildings generate, and for a mortgage REIT, whose earnings are a levered bet on interest rates.
A price-to-earnings ratio assumes this year's earnings resemble next year's. That is fair for a consumer staple and false for a memory chipmaker at the top of its cycle, where the lowest PE marks the most dangerous moment, or for a health insurer in a year when medical costs run hot.
The fix is not a smarter universal model. It is a different method for each industry, chosen to fit how that business actually earns. Here is the map.
The Map
| Industry | What fits it | What the method cannot see |
|---|---|---|
| Banks | Price to tangible book, adjusted for return on tangible common equity | Non-performing loans, capital ratios, charge-offs |
| Equity REITs | Price to funds from operations, not earnings | Lease quality, net asset value, occupancy |
| Mortgage REITs | One times tangible common book | Hedging, the duration gap, book value erosion |
| Memory and cyclicals | Normalized mid-cycle cash flow, not a peak year | Where in the cycle the business sits |
| Energy producers and deep cyclicals | Through-cycle cash flow, and no growth score | The commodity price deck, supply and demand |
| Health insurers | Normalized forward earnings on adjusted EPS | The medical loss ratio, reserve adequacy |
The pattern across the table is the same. For each industry there is a number the market reaches for that misleads, a bank's low PE, a REIT's depreciation-buried earnings, a cyclical's peak-year profit, and a method that fits better once you see how the business earns. And in every row the valuation is the easier half. The harder half, the asset quality, the lease, the cycle, the claims, sits in the filings, which is why the right-hand column matters as much as the middle one.
Even the methods on the map are screens, not full builds. The REIT row runs on a simplified funds from operations, net income plus depreciation, without the property-sale and impairment adjustments a full FFO would make. The mortgage REIT row anchors on tangible book calibrated for the agency names, and reads a commercial or servicing-heavy trust more roughly. Each is enough to place a stock and honest about being a starting point, which is the most a screen should claim.
The Industries I Have Not Solved Yet
A map is only honest if it marks the blank spaces. The clearest one is the master limited partnership, a structure still common across parts of the pipeline and midstream energy world, even as several large names have converted to ordinary corporations. An MLP issues units rather than shares, reports distributions rather than dividends, and hands investors a K-1 tax form instead of a 1099. Its per-unit figures do not line up with the per-share math a screen runs, and the measure that fits it is usually enterprise value to EBITDA rather than free cash flow, which tends to understate a pipeline's real worth. Until that method is built and tested, MiniValuator does not claim to value an MLP well, and saying so is better than handing you a confident number built on the wrong unit.
The same honesty applies to any business not on the map above. If an industry is not listed, the general model may not fit it, and the read deserves more skepticism, not less.
How MiniValuator Picks the Method
When MiniValuator reads a stock, it first recognizes the industry, then routes to the method built for it instead of forcing one model across all of them. It also turns off the red-flag checks that misfire on that industry, the interest-coverage warning on a bank, whose interest is raw material rather than strain, and the inventory check on a business that holds none. A bank goes to tangible book, an equity REIT to funds from operations, a mortgage REIT to tangible common book, a cyclical to normalized cash flow, a health insurer to forward earnings. When it uses one of these special methods, it labels the result as limited coverage so you know it stepped off the standard model, and it states plainly what the method does not cover. The verdict on a bank might say it valued the lender on tangible book while loan quality stayed out of reach, which is more useful than a flat not covered that reads like a downgrade. The label is a disclosure, not a black box. See how it reads any stock.
The Bottom Line
There is no single right way to value a stock, only a right way for each kind of business. A bank is its balance sheet. An equity REIT is screened on funds from operations, which adds depreciation back to earnings. A mortgage REIT is a levered bet on rates. A cyclical is a normalized average, not a single year. An insurer is a through-cycle earnings stream. Match the method to the business and the valuation gets honest. But the valuation is always the first half. The second half, the risk that lives in the filings, is yours to read, on every industry on this map and the ones not yet on it.
Frequently Asked Questions
Why can't you use one valuation model for every stock? Because different industries earn money in ways that break a single model in different places. A discounted cash flow fails on a bank, whose cash flow mixes lending and funding, and on a REIT, whose depreciation hides its real cash. A price-to-earnings ratio fails on a cyclical at its peak and an insurer in a bad claims year. The honest fix is a method matched to each industry, not one model forced across all of them.
Which industries does MiniValuator use a special method for? Banks, valued on tangible book and return on tangible common equity. Equity REITs, on price to funds from operations. Mortgage REITs, on one times tangible common book. Memory and other deep cyclicals, on normalized mid-cycle cash flow. Energy and cyclicals, with no growth score and a through-cycle read. Health insurers, on a normalized forward earnings multiple. Each has its own guide linked in the map above.
How does MiniValuator know which method to use? It recognizes the industry first, then routes to the method built for that business rather than running one model over everything. When it uses a special method, it labels the result as limited coverage and says what the method cannot see, so the read is a disclosure rather than a black box.
Why is valuing an MLP or pipeline harder? A master limited partnership issues units rather than shares and reports distributions rather than dividends, so the per-share math a screen runs does not line up. The measure that fits a pipeline is usually enterprise value to EBITDA rather than free cash flow. MiniValuator does not yet value MLPs well, and it says so rather than pretending otherwise.
What do all these methods have in common? Each one prices the business fairly for what it reports today, and none can see the risk that decides tomorrow. That risk, a bank's loan book, a REIT's leases, a cyclical's turn, an insurer's claims, lives in the filings as text and schedules, not in a numeric feed. The valuation is the first half of the work, and reading the filing is the second.
Does a special method mean the stock is worse? No. The limited coverage label is a disclosure that the tool used an industry-specific method and turned off checks that would misfire, not a judgment that the stock is weaker. It is more honest than forcing a standard model and reporting a number that does not fit.
Want to see the method that fits a given stock? Open MiniValuator and check any bank, REIT, cyclical, or insurer against the right one.
