Backlog, book-to-bill, and modeling an industrial cycle

Three words interviewers expect you to define precisely

Industrials interviews turn on vocabulary faster than most sectors, because these words sound interchangeable and are not. Get them wrong in a superday and the rest of your answer stops counting.

An order is a customer commitment to buy. It is not a shipment, it is not revenue, and it is not always a signed contract in the legal sense. Companies book orders at different moments (a purchase order received, a contract executed, a notice to proceed issued on a construction job), and the booking policy is a disclosure choice, not a uniform standard.

Backlog is the stock of orders received but not yet delivered and recognized as revenue. It is a balance measured at a point in time, and it is not a balance sheet item in the audited sense. Most of what companies call backlog sits in a disclosure or an investor deck, alongside the narrower, standards-defined concept of remaining performance obligations.

Book-to-bill is orders divided by revenue over the same period. Above 1.0 means the company booked more work than it shipped, so backlog grew. Below 1.0 means it is shipping faster than it is booking, so backlog is shrinking. Exactly 1.0 means backlog is flat. It is a flow ratio, which makes it a better read on demand direction than the backlog level itself.

The roll-forward ties the flow to the stock, and interviewers do ask candidates to recite it:

Beginning backlog, plus orders booked, less revenue recognized out of backlog, plus or minus cancellations and scope changes, plus or minus currency translation, equals ending backlog.

A hypothetical single quarter, in millions:

Book-to-bill here is $1,100 divided by $950, or roughly 1.16, yet backlog grew only about 1.5 percent, because cancellations and a currency headwind ate most of the gain. That gap is what a good answer notices unprompted: book-to-bill tells you about gross demand, the roll-forward tells you what survived into the backlog.

Backlog quality: the part that separates a real answer from a rote one

Anyone can quote a backlog number. The candidates who stand out ask what is inside it. Five questions cover it.

Is it funded? On government and defense programs, a contract award and the appropriated money behind it are two different things. Funded backlog has dollars obligated against it. Unfunded backlog covers awarded but not yet appropriated work, including option years the customer may never exercise. Total backlog can be several times funded backlog, which makes the headline a statement about potential rather than contracted cash. That distinction, plus the vocabulary around contract vehicles and program of record risk, is core to aerospace and defense banking.

Is it cancellable, and what do the termination provisions do? Most commercial backlog is cancellable. What matters is what the customer owes on cancellation. A termination-for-convenience provision typically makes the customer pay costs incurred plus some margin on work performed, which protects the outlay but rarely delivers the planned profit. Cancellable backlog is worth the recovery, not the revenue. Ask whether cancellation is penalty-free, whether deposits or progress payments have been collected, and whether the equipment is standard (resellable) or bespoke (a write-down waiting to happen).

How long is it, and how much converts inside twelve months? Duration is the most useful quality metric and the one candidates skip. Two companies with identical backlog are in completely different positions if one converts 80 percent within a year and the other 25 percent. The twelve-month conversion figure is what feeds a near-term revenue build. Long-dated backlog also carries more cancellation, inflation, and customer-credit risk.

What margin is embedded in it? A record backlog booked at bad prices is a liability, not an asset, because the company is contractually obligated to deliver work it will lose money on. That is the classic engineering and construction failure mode: bidding aggressively to fill the book during a soft stretch, then absorbing cost overruns on fixed-price work for years, with loss provisions recognized the moment a contract goes underwater. Fixed-price versus cost-plus mix, and how project accounting turns a bad bid into an immediate charge, runs through building products and construction.

Does it escalate? Long-dated fixed-price backlog carries input-cost risk on steel, copper, castings, electronics, and labor. Escalation or index clauses pass some of that to the customer. Their absence leaves the company effectively short commodities and wages for the life of the contract.

Short-cycle versus long-cycle, and why backlog means opposite things

The same metric means different things depending on lead time, and interviewers use it to see whether you understand the businesses or just the ratio.

In short-cycle businesses (components, fasteners, consumables, aftermarket parts, distribution), lead times run from days to a few weeks. Backlog is thin by construction, often a few weeks of revenue. Orders are the demand signal in near real time and move almost in step with revenue, so book-to-bill carries little information the revenue line does not already have.

In long-cycle businesses (large turbines, locomotives, mining and construction equipment, defense platforms, process plants, marine), backlog spans years. An order booked today may not become revenue for two or three years, so the current order rate tells you about revenue in the outer years, not next quarter.

The practical consequence is what interviewers are driving at: a long-cycle company can post growing revenue out of backlog while its order book is already deteriorating, because it is converting work booked during better conditions. Book-to-bill is the leading indicator, revenue is the lagging one, and the gap can run for several quarters. A candidate who says "revenue is up, so demand is strong" about a long-cycle industrial has failed. The correct read is: revenue is up, book-to-bill is 0.85, backlog is draining, and unless orders recover, revenue rolls over roughly one backlog-duration out. Which machinery end markets are short-cycle and which are long is laid out in machinery and capital goods.

The other leading indicators, and the destocking trap

Orders are the primary signal but not the only one. What follows are categories rather than numbers, because the numbers move and the logic does not.

Capacity utilization across the customer base. Customers do not buy new capacity while existing capacity sits idle, so utilization has to climb toward practical limits before expansion orders appear. That is why the first leg of a recovery shows up as maintenance spending.

Fleet age and installed base. An aging fleet builds a replacement need that can be deferred through a downturn but not forever, and deferred replacement is stored demand that releases when confidence returns. The installed base also drives aftermarket parts and service revenue, typically higher margin, more recurring, and far less cyclical than original equipment sales. Aftermarket mix is one of the strongest quality-of-earnings arguments in the sector, and a large installed base earns a different multiple than a pure equipment maker.

Used-equipment pricing. Secondary market prices are a clean read on real-time demand and on substitution, because a cheap used machine competes directly with a new one. When used prices collapse, new orders follow.

Channel inventory, meaning sell-in versus sell-through. Sell-in is what the manufacturer ships to dealers and distributors. Sell-through is what those dealers sell to end users. The difference sits in the channel.

Customer capital expenditure budgets and, where relevant, permitting or project pipelines. For long-cycle equipment the customer's own budget cycle is the constraint, and budgets are set annually and revised reluctantly.

Destocking is the trap worth understanding cold. When the channel decides it holds too much inventory, dealers stop ordering until they work the excess down. End demand might fall 5 percent while reported orders fall 20 percent, because the channel absorbs the correction on top of the underlying decline. The reverse happens on the way up: when dealers restock, reported orders rise faster than end demand. Reported results therefore overshoot the real cycle in both directions. Interviewers like this one because they can hand you a frightening order decline and see whether you ask the follow-up. Ask what sell-through did, and whether weeks of supply in the channel are rising or falling.

How to actually model a cyclical industrial

The mistake almost every candidate makes is forecasting revenue as one growth rate. Build it from its two components instead.

Split the revenue forecast. Backlog-driven revenue is beginning backlog times the twelve-month conversion rate, adjusted for expected cancellations. Book-and-turn revenue is the short-lead-time business booked and shipped inside the same period, forecast off end-market indicators rather than backlog. Then forecast new orders separately, roll them into backlog, and let the next year's revenue draw on the new ending balance. That forces an explicit view on orders, the actual driver, rather than hiding it inside a blended growth rate, and it makes the model self-checking: if revenue grows while backlog goes negative, the assumptions are broken.

Then layer margin using incremental and decremental margins rather than a flat margin percentage. Incremental margin is the share of each additional revenue dollar that reaches EBITDA. Decremental margin is the share of each lost revenue dollar that comes out of it. Hypothetically: if revenue falls $100 million and the decremental margin is 30 percent, EBITDA falls $30 million.

High fixed-cost manufacturers have steep decrementals because plants, tooling, engineering staff, and skilled labor do not scale down with volume in the short run. Absorption works against you: fixed overhead spread across fewer units raises unit cost before pricing even gives way. More variable cost structures, outsourced manufacturing, or large aftermarket bases produce shallower decrementals. Asking what a business's decremental margin is, and why, is a strong question in any industrials interview.

Do not straight-line a peak year into terminal value. This is the most common valuation error in the sector. A cyclical company at a cycle high has margins that are not sustainable, and capitalizing them into perpetuity manufactures value that does not exist. The terminal year of a cyclical DCF has to be mid-cycle: normalized volumes, margins, and capital expenditure, with the forecast period long enough to walk the business from wherever it sits today back to that level. The same logic applies to multiples, where trailing EBITDA at a trough makes a company look expensive on a multiple that is cheap on normalized earnings, and the reverse at a peak. How that shows up in comparable company analysis and mid-cycle multiples is covered in how industrials companies are valued.

What to watch at each stage of the cycle

Written generically, with no claim about any particular moment in time:

The asymmetry worth naming is that capacity arrives late, near the peak, because approval and construction lead times run behind the demand that justified them. That is a structural reason industrials cycles overshoot, and it feeds directly into how much leverage a cyclical earnings stream can support, since a lender sizing debt against peak earnings is underwriting a number the capacity build is about to erode. Sponsors handle this by underwriting to trough or mid-cycle cash flow rather than trailing results, and leveraged finance terms covers the covenant vocabulary that gets used to hold that line.

The traps, named

Quoting backlog without asking what is in it. Funded or unfunded, cancellable or not, twelve-month conversion, embedded margin, escalation. If you cite a backlog number in an interview, attach at least one of those qualifiers.

Treating book-to-bill above 1.0 as automatically good. One large lumpy order can push a quarter above 1.0 in a business whose underlying order rate is deteriorating, and in long-cycle equipment a single multi-year award can exceed a quarter of normal bookings. Ask whether the ratio is broad-based or concentrated, and whether the trailing twelve-month figure tells the same story as the quarter.

Confusing an order with a signed contract. Booking policies differ, letters of intent are not contracts, and awards subject to financing, permitting, or appropriation can evaporate. An announced award, a booked order, and a funded contract are three different things.

Assuming record backlog means the company cannot miss. Backlog is a revenue reservoir, not a guarantee of earnings. Supply chain constraints can block conversion, cost inflation can destroy the margin embedded at award, customers can cancel or push out delivery, and the book can convert at prices set in a worse environment. Record backlog with a book-to-bill below 1.0 and a bad price mix describes a company at a peak, not a company that is safe.

Practice question

A company you cover just reported record backlog, but book-to-bill was 0.9. How do you interpret that, and what would you want to know next?

Those are two different things about two different time horizons. Backlog is a stock built from orders booked in earlier periods, so it supports revenue for however long it converts. Book-to-bill at 0.9 says the company shipped more than it booked this period, so the backlog is draining. If that persists, revenue rolls over roughly one backlog-duration out, even though reported revenue right now looks strong. In a long-cycle business orders lead and revenue lags, so I would not read record backlog as evidence that demand is healthy today.

The first thing I would want is the twelve-month conversion rate, because backlog duration determines how much runway that record number buys. Then whether the 0.9 was broad-based or distorted by a large order in the prior period, looking at the trailing twelve-month ratio rather than one quarter. After that, quality: how much of the backlog is funded, how much is cancellable and on what terms, what margin was embedded at award, and whether the contracts escalate. Finally, channel inventory, because if dealers are destocking, reported orders are falling faster than end demand and the 0.9 overstates how bad the picture is.

What the interviewer is listening for: Whether you know that book-to-bill is the leading indicator and revenue is the lagging one, rather than treating record backlog as a clean positive. They also want to see you go straight to backlog quality (funding, cancellability, duration, embedded margin) instead of stopping at the headline figure, and to hear you separate a genuine demand deterioration from a lumpy comparison or a channel inventory correction.