LTM vs NTM Multiples and Calendarization

Every valuation multiple hides a second question inside it: over what period? A company trading at "12x EBITDA" means very little until you know whether that EBITDA is the last twelve months of actual results or the next twelve months of analyst hope. The period in the denominator matters as much as the metric itself, and mixing periods is one of the fastest ways to corrupt a comps table or trip yourself up in an interview.

LTM (last twelve months) and NTM (next twelve months) are the two workhorse conventions. LTM looks backward at reported, filed numbers. NTM looks forward at estimates. Both describe a rolling twelve-month window, yet one is history and the other is forecast, and that single difference changes where each belongs, how you build it, and how a fast-growing company's multiple should be read.

This post is about the mechanics: how to construct an LTM figure from a 10-K and a 10-Q stub with real numbers, where forward estimates come from, and how to calendarize an off-cycle company so its multiple lines up with a December-year-end peer. If you want the catalog of which multiples exist, our guide to common valuation multiples covers EV/EBITDA, P/E, and the rest. Here we focus on the time dimension sitting underneath all of them.

What LTM (Last Twelve Months) Captures

LTM is the sum of a company's most recent four reported quarters. It is a rolling window that moves forward every time a new quarter is filed, so the LTM figure in March differs from the LTM figure in June even though the fiscal year end has not changed. Because it is built entirely from filed results, LTM is verifiable: anyone with the 10-K and the latest 10-Q can reproduce it.

The strength of LTM is that nobody can argue with it. The numbers happened. The weakness is that they already happened, which makes LTM a poor guide to a business that looks nothing like it did a year ago.

Last Twelve Months (LTM)

The sum of a company's financial results over the most recent four reported quarters, also called trailing twelve months (TTM). LTM figures are built from actual filings rather than estimates, which makes them the standard for precedent transactions and credit analysis.

What NTM (Next Twelve Months) Captures

NTM is the mirror image: the sum of the next four projected quarters, drawn from analyst estimates or a management plan. It is inherently an opinion about the future, so two analysts can produce two different NTM figures for the same company on the same day.

Forward multiples exist because valuation is about what a buyer receives going forward, not what a seller earned in the past. A subscription software business growing 30% a year will always look expensive on trailing numbers and more reasonable on forward numbers, which is exactly why growth sectors quote NTM by default.

Before we go deeper, here is the distinction on one page.

In practice, bankers rarely pick one and discard the other. A finished comps output usually shows both LTM and NTM columns side by side, because each answers a different question. What changes across deal types is which column the conclusion actually leans on.

Precedent Transactions Lean on LTM

A precedent transactions analysis looks at what acquirers paid for similar businesses in the past. The multiples are computed on the target's LTM figure as of the announcement date, because that is the hard, disclosed number both sides negotiated against. Buyers do not want to pay for a seller's optimistic projections, and the LTM figure is the neutral, filed anchor. Our walkthrough of precedent transactions analysis explains how these deal multiples are assembled and cleaned, but the period convention is almost always trailing.

Trading Comps Lean on Forward Estimates

Trading comps value a company against where its public peers trade today. Since public markets price in expected performance, the conclusion usually leans on forward multiples, most often NTM or the next fiscal year. A stock priced at 20x LTM earnings but 15x NTM earnings is telling you the market expects roughly a third more earnings next year. The full mechanics of assembling a peer set live in our guide on how to build a comparable company analysis; the point here is that trading comps and precedent comps often disagree on period on purpose.

Credit and Covenants Live in LTM

Lenders think almost entirely in LTM. A leverage ratio of $\frac{\text{Net Debt}}{\text{LTM EBITDA}}$ uses trailing EBITDA because a covenant has to be tested against numbers that actually exist, not against a forecast the borrower controls. When a credit agreement says leverage may not exceed 4.5x, it means 4.5x the last twelve months of EBITDA, measured each quarter. Forward EBITDA would hand the borrower a lever to loosen its own covenant, which no lender allows.

The most common technical task involving LTM is constructing it when a company has reported partway through its fiscal year. You cannot just grab the latest 10-K, because that captures a fiscal year that may have ended months ago. You have to roll it forward using the interim filings.

The LTM Formula and Stub Periods

The identity every analyst memorizes is:

You start with the most recent full fiscal year, add the year-to-date results reported in the latest 10-Q, and subtract the same year-to-date period from one year earlier. The subtraction strips out the portion of the old fiscal year that the new stub has already replaced, leaving a clean twelve-month window ending on the latest quarter date.

Stub Period

A partial-year reporting period, such as the three or nine months covered by a 10-Q, that sits between full fiscal years. Building an LTM figure means adding the current stub and subtracting the year-earlier stub so the result still spans exactly twelve months.

The mechanics are always the same regardless of the metric:

Worked Example: Coca-Cola's LTM Revenue

Take The Coca-Cola Company, which reports on a December fiscal year. For its fiscal year ended December 31, 2024, Coca-Cola reported net operating revenues of roughly $47.1 billion (precisely $47,061 million). Suppose you are valuing Coca-Cola as of the third-quarter 2025 filing. From the Coca-Cola Q3 2025 10-Q, net operating revenues were $36,119 million for the nine months ended September 26, 2025, versus $35,517 million for the nine months ended September 27, 2024.

Plug those into the identity:

So Coca-Cola's LTM revenue as of late September 2025 is about $47.7 billion, not the $47.1 billion you would have used if you lazily pulled the last 10-K. The $602 million difference is exactly the year-over-year growth in the first nine months of 2025, and ignoring it would understate the denominator and overstate the multiple.

The same build works on any income-statement line. The identical arithmetic works for EBITDA, EBIT, or net income; you simply run the FY plus current stub minus prior stub on that line instead of revenue. The one wrinkle is that EBITDA usually needs normalizing adjustments (one-time charges, non-recurring gains) applied consistently across all four quarters, so a stub with a big impairment can distort the LTM figure if left raw. Apply the same add-backs to every quarter before you build the trailing number, or the window will not be comparable.

Trailing multiples reward reps more than reading: Practice building LTM figures and other valuation math with worked answers, start practicing interview questions for free and find the gaps before an interviewer does.

If LTM is a construction task, NTM is a sourcing task. You are not computing forward numbers from filings; you are pulling them from somewhere, and where you pull them from determines how much you should trust them.

Consensus and Management Estimates

The most common source is consensus, the mean or median of sell-side analyst estimates aggregated by data vendors. On a public company, consensus NTM revenue and EBITDA are what trading comps use, because they represent the market's shared expectation and are updated continuously as analysts revise. For private or thinly covered names, the forward numbers usually come from management projections in a model, a lender presentation, or a confidential information memorandum. These are useful but self-interested: management has every reason to present an ambitious plan, so a banker discounts them accordingly.

Vendors like S&P Capital IQ, FactSet, and Visible Alpha matter here because they also calendarize the estimates for you, standardizing companies with different fiscal year ends onto a common calendar basis. Knowing where the numbers are sourced is part of being able to read SEC filings and financial data critically rather than trusting whatever a screen spits out.

The Pitfalls of Forward Numbers

Forward estimates carry risks that trailing numbers do not. Consensus can be stale in a fast-moving situation, clustered around a herd view that turns out wrong, or thin for a company only two analysts cover. Management cases are optimistic by construction. And a single quarter that misses can swing the entire NTM figure, because you are extrapolating from a small, recent base.

Even after you settle on LTM or NTM, one problem remains. Peers in the same industry often close their books on different dates, and comparing a June-year-end company's results against a December-year-end peer compares different slices of time. Calendarization fixes that.

Calendarization

The process of adjusting a company with an off-cycle fiscal year end so its financials line up with a common calendar period, usually the calendar year or a calendar-based next-twelve-months window. It is done by blending the two fiscal years that straddle the target period in proportion to the months each contributes.

Why Fiscal Year Ends Break Comparability

Imagine a comps set where one company closes in December, another in June, and a third in September. Their most recent annual figures span three different twelve-month windows, potentially crossing different demand cycles, currency moves, or input-cost environments. Dropping those raw annual numbers into one table produces multiples that are not comparable, even though every cell looks tidy. Calendarization rebuilds each off-cycle company's figures as if it, too, reported on the calendar year, so the multiples finally describe the same span of time.

Worked Example: Calendarizing Nike onto Calendar 2024

Nike closes its fiscal year on May 31, which makes it a textbook calendarization case. Its fiscal 2024 (ended May 31, 2024) revenue was about $51.4 billion, and its fiscal 2025 (ended May 31, 2025) revenue was about $46.3 billion, per Nike's fiscal 2025 Form 10-K. Neither of those windows is calendar year 2024, so to compare Nike against a December-year-end peer on a calendar-2024 basis, you blend the two fiscal years by the months each contributes to that calendar year.

Calendar 2024 (January to December 2024) is made up of two pieces. January through May 2024, five months, falls inside fiscal 2024. June through December 2024, seven months, falls inside fiscal 2025. The weighted formula is:

Filling in Nike's numbers, with five twelfths of fiscal 2024 and seven twelfths of fiscal 2025:

That works out to roughly $21.4 billion from the fiscal 2024 slice plus about $27.0 billion from the fiscal 2025 slice, for a calendarized 2024 revenue near $48.4 billion. Now Nike's multiple can sit in the same table as a December-year-end peer without comparing mismatched periods.

Forward estimates calendarize the same way. To build a calendar-year NTM estimate from a company that reports fiscal-year consensus, you blend the current and next fiscal-year estimates by the months each contributes to the twelve-month window you care about. This is exactly what the data vendors automate, but you should be able to do it by hand, because interviewers love asking you to calendarize a simple two-year forecast on the spot.

Get the complete guide: Download our comprehensive 160-page PDF, access the IB Interview Guide covering valuation, modeling, and the technical frameworks interviewers actually test.

Building the periods correctly is only half the discipline. The other half is making sure the numerator and denominator of every multiple describe the same thing.

Match the Numerator and Denominator Periods

A multiple is a ratio, and a ratio is meaningless if its two halves cover different periods. If your denominator is LTM EBITDA, your numerator (enterprise value or equity value) must be measured as of the same date, and you must not quietly pair it with a forward figure elsewhere in the table. The classic failure is an EV built from today's share price divided by an EBITDA figure that is nine months stale. Keep enterprise value and its metric on the same clock, and if you are unsure how the numerator is assembled, revisit enterprise value versus equity value before you spread the comp.

Use Diluted Shares as of Today

Even in a trailing multiple, the share count in the numerator should reflect today's fully diluted shares, not the diluted count from the fiscal year you are pulling EBITDA from. Options, restricted stock, and convertibles change the diluted count over time, and enterprise value is a live, market-based number. The treasury stock method is how you convert options into net new shares, covered in our explainer on the treasury stock method and diluted shares. Using a stale share count while marking equity value to today's price is a subtle but common inconsistency.

Interviewers use LTM, NTM, and calendarization to probe whether you actually understand what a multiple means, or whether you just memorized "12x EBITDA." A few traps come up again and again.

Mixing LTM Metrics With Forward Multiples

The single most common mistake is applying a multiple built on forward numbers to a trailing metric, or vice versa. If comparable companies trade at 14x NTM EBITDA and you multiply that by your target's LTM EBITDA, you have valued the company on a mismatched basis and probably understated it if it is growing. The peers' forward multiple must be applied to the target's forward EBITDA. This trap also shows up as stale stub data, where an analyst forgets to roll the LTM figure forward after a new 10-Q drops and keeps valuing the company on last year's numbers.

Growth Companies Where LTM Overstates the Multiple

For a fast-growing business, the trailing multiple is mechanically inflated because the denominator is the smallest it will be. A company earning $100 million of LTM EBITDA but expected to earn $150 million next year trades at 20x on trailing numbers and about 13x on forward numbers at the same $2 billion enterprise value. Quoting only the LTM multiple makes the company look far more expensive than the market believes it is, which is precisely why growth and technology names are quoted on forward multiples. If an interviewer asks why a software company trades at 40x EBITDA, the right instinct is to ask whether that is trailing or forward before you call it overvalued.

The distinction between trailing and forward multiples is not academic. It decides how you build the number, where the figure belongs, and how you read a company's valuation.

Master these mechanics and you will avoid the errors that quietly wreck comps outputs and the interview traps built to catch candidates who never learned the difference. When you can build an LTM figure from a 10-Q, calendarize a June-year-end peer, and explain why a growth company looks cheaper on forward numbers, you are thinking like an analyst who understands valuation rather than one who memorized a multiple.