Industrials are, at their core, a bet on someone else's capital-spending decision. Machinery, equipment, engineering, and capital-goods makers sell into the capital expenditure (capex) budgets of other companies and governments — and those budgets swing hard with the business cycle, far more than consumer spending does. Two linked toolsets dominate fundamental analysis of the sector: the capex cycle (where in the expansion/contraction sequence the customer base sits, read through macro proxies like ISM/PMI and capacity utilization) and backlog/order metrics (book-to-bill, the backlog roll-forward, and disclosed remaining performance obligations), which give an early, company-specific read on demand inflections before they reach the income statement. The central tension: backlog and orders are the best leading visibility available, but they are forecasts, not contracts of certainty — and the income statement looks most reassuring exactly when the cycle is about to turn.

How it's calculated / formed

Book-to-bill ratio = new orders (bookings) ÷ amount billed or shipped (billings/shipments), over a period (month or quarter). Above 1.0 = orders outpace deliveries, backlog is building, future revenue growth implied. Below 1.0 = backlog depleting, future revenue contraction implied. (In the original semiconductor-equipment usage the denominator is shipments/billings; some industrials report it against recognized revenue — close but not identical, so check the company's own definition.) The metric originated among semiconductor-equipment makers in the late 1970s–early 1980s to navigate cyclical, long-lead-time demand; SEMI formalized a standardized monthly North American SME report, though it ceased publishing the bookings-based monthly book-to-bill after December 2016 (Wikipedia; SEMI; Semiconductor Digest).

Backlog roll-forward (waterfall): > Beginning backlog + new orders − revenue recognized − cancellations = ending backlog

The backlog conversion rate is how fast backlog turns into revenue. Per industrials-IB modeling guides, production programs commonly convert ~30–40% annually, while milestone-based development programs convert slower and less predictably; it should be modeled by program type, not company-wide.

Remaining Performance Obligations (RPO) is the ASC 606 standardized cousin of backlog (606-10-50-13). RPO = transaction price allocated to unsatisfied obligations = deferred revenue + un-invoiced contracted backlog. cRPO is the portion expected within 12 months. ASC 606 counts only enforceable consideration — it excludes cancel-for-convenience portions, usage minimums without commitment, and pipeline/renewals — so disclosed RPO is generally more conservative than a company's marketing "backlog" figure.

Capex split: Maintenance capex (sustains the asset base, roughly tracks depreciation, commonly cited at ~3–5% of revenue) versus growth capex (discretionary capacity expansion). Capacity-utilization zones drive growth-capex timing: below ~75% defers it, ~75–82% is normal, above ~85% triggers new capacity decisions (industrials-IB capex guide).

How it's used in practice

1. Order metrics as an early-warning system. For long-cycle industrials, the backlog is the revenue model. Analysts watch book-to-bill and sequential order growth as the earliest financial signal of a cyclical turn — orders inflect before revenue, which inflects before margins. 2. Cycle positioning via macro proxies. ISM Manufacturing PMI (>50 expansion / <50 contraction) is the most-watched industrial leading indicator and is commonly cited as leading industrial earnings by roughly two quarters. Analysts pair it with durable-goods orders, the Fed industrial-production index, capacity utilization, and freight volumes to locate the cycle stage. 3. Sub-sector cycle mapping. Capital goods/machinery (Caterpillar, Deere, Parker Hannifin, ITW) are textbook mid-cycle names that peak in expansions; building products are more early-cycle (housing-led); defense/aftermarket services are more late-cycle/defensive. 4. Through-cycle (normalized) valuation. Because trailing multiples invert across the cycle, analysts value cyclicals on mid-cycle earnings, not trough or peak. 5. Backlog-quality scrubbing. Funded vs. unfunded, fixed-price vs. cost-plus, firm orders vs. options — the composition matters as much as the headline number.

Adoption, debate & evidence

These are mainstream institutional tools, not contested folklore — book-to-bill, backlog disclosure, and PMI-based cycle reading are standard in sell-side and buy-side industrials coverage. The honest caveats are about reliability, not legitimacy:

• Trailing multiples mislead. Cyclical industrials look cheapest at peaks and most expensive at troughs. One industrials guide notes Caterpillar's trailing EV/EBITDA ranged roughly 7.7x to 26x over ~13 years (median ~13x), with low readings at earnings peaks — the classic value trap. (Single-source; treat the exact figures as illustrative, but the inverse relationship is well established.)
• Backlog is not guaranteed revenue. Orders can be deferred, modified, or cancelled. Cancellation assumptions in practice run ~1–3%/yr for defense (rarely triggered) to ~3–8% for commercial equipment. During COVID-19, customers deferred deliveries without cancelling — backlog held up on paper while near-term revenue evaporated.
• PMI is a survey, not a guarantee. It signals direction with lead time but is revised and noisy; the "leads earnings by ~two quarters" figure is a commonly cited rule of thumb, not a precise law.

Strengths & limitations

Works best for long-cycle, large-ticket businesses (aerospace, mining/construction equipment, power generation, defense, infrastructure E&C) where backlog spans multiple quarters and orders genuinely lead revenue. In those names, a deteriorating book-to-bill is one of the few hard, forward signals available before the cycle visibly rolls over.

Fails / misleads when: (a) backlog quality is ignored — unfunded, option, or cancel-for-convenience backlog padded into the headline; (b) book-to-bill is read as a point estimate rather than a trend (a single sub-1.0 quarter after a record book-to-bill is normal mean-reversion, not a recession); (c) it's applied to short-cycle, book-and-ship businesses (consumables, distribution) where backlog is structurally thin and meaningless.

The single most common misuse: treating large backlog as a safety cushion at a cycle peak. The reassurance is highest precisely when forward orders (book-to-bill < 1.0) have already turned down — the income statement is the last place the cycle shows up, not the first.

Sources

• Wikipedia — Book-to-bill ratio (definition, above/below 1.0 interpretation)
• SEMI — About the Billings Report (SME book-to-bill methodology, discontinuation post-2016)
• Semiconductor Digest — historical SEMI book-to-bill commentary
• IB Interview Questions, Industrials IB guides — Backlog & Book-to-Bill Modeling; How Cyclicality Flows Through Industrial Financials; Maintenance vs. Growth Capex & Capacity Utilization; Through-Cycle Multiples & Peak-Trough Analysis (roll-forward formula, conversion/cancellation ranges, utilization thresholds, Caterpillar multiple range — single-source, treated as illustrative)
• Ordway Labs / The SaaS CFO — Remaining Performance Obligations & ASC 606 (606-10-50-13, RPO = deferred revenue + backlog, cRPO, enforceable-consideration exclusions)
• ISM — Report on Business / ISM Manufacturing PMI (>50 / <50, leading-indicator role)

Flags / disputes: the Caterpillar EV/EBITDA range and specific conversion/cancellation percentages come from a single practitioner-education source — directionally sound and consistent with the well-known inverse-multiple phenomenon, but treat the exact figures as illustrative rather than canonical. The "PMI leads earnings by ~two quarters" is a rule of thumb, not a measured constant.