9 min read · Cost
Most window problems do not require a new window — and a contractor who leads every conversation with full replacement is selling, not diagnosing. The honest version is a decision tree: some failures are cheap mechanical repairs, some are glass-only swaps that save the frame, and some are the frame telling you its era is over. This guide walks the triage symptom by symptom, then the age-and-quantity math that decides whole houses. Two companion guides own their branches in depth — foggy glass and seal failure and the energy-savings reality check — so this page stays on the decision itself.
Start with the split: hardware, glass, or frame?
Every window complaint sorts into one of three layers, and the layer sets the economics. Hardware problems — a sash that will not stay up, a crank that grinds, a lock that will not catch, weatherstripping gone flat — are the window's consumables, and they are repairable almost by definition: sash balances, operators, locks, and seals are replaceable parts, stocked generically for common windows and by the manufacturer for name brands. A hardware repair costs a fraction of any replacement and restores full function; on an otherwise sound window it is the obvious call, and we will say so even when we were invited to quote a replacement. Glass problems — a cracked pane, a fogged insulated unit — live in the middle: the glass or the IGU can be replaced without touching the frame, which is the pivot point most homeowners do not know exists (next section). Frame problems — rot in the sash or jamb, warp, racking out of square — are the terminal layer: the frame is the window, and when it fails, repair economics collapse quickly. The first diagnostic question on any service call is simply which layer is talking — and a bidder who cannot tell you which layer your problem lives in has not diagnosed it.
Glass-only replacement is legitimate — here's when
The window industry's quietest secret: a broken pane does not obsolete a window. A single cracked lite in a structurally sound frame is a glass-shop job — the pane or the sealed IGU comes out, a new one goes in, and the frame, trim, siding, and interior finish never know it happened. The same applies to fogged dual-pane units: an IGU swap replaces the sealed glass package inside your existing sash, restoring clarity and insulation at a fraction of full-replacement cost. When glass-only is the right call: the frame is sound and reasonably young, the failure is one or a few units rather than a pattern, and the existing windows otherwise perform — a rock through a ten-year-old vinyl slider is a glass job, full stop. When it stops making sense: the failures are systemic (three fogged units this year, two last year — that is a product generation aging out, and per-unit swaps become a subscription), the frames are single-pane-era aluminum or builder vinyl whose energy performance was never good, or the swap cost approaches replacement money on a window you would upgrade anyway. Two honest cautions: a replacement IGU should match the original's coating and spec or the repaired window will read visibly different beside its neighbors, and openings in safety-glazing locations must be brought to current requirements when the glass is replaced — the swap does not dodge the code.
When the frame ends the debate
Frame failures close the repair conversation faster than anything else, and rot closes it fastest. Wood sash and frame rot starts where water sits — sill horns, bottom rails, the joint under a weep — and by the time it shows at the surface, the wet path behind it has usually been running for years; patch-and-epoxy work buys time on a historic window worth preserving, but on an ordinary builder window the repair labor is bespoke carpentry billed at bespoke rates against a mass-produced replacement, and the economics fade fast. Worse, rot is rarely just the window's problem — water that rotted a sill has had access to the rough opening and the wall below, which is exactly the hidden-damage story we open walls to find on re-side projects. Operational failures tell the same story from a different angle: a window painted shut through five repaint cycles, sashes that rack because the frame is out of square, a slider whose frame has bowed — these are geometry problems, and you cannot adjust geometry back. As a rule: replace hardware, replace glass, but when the frame itself is compromised — rot, warp, racking — replacement is not the upsell, it is the repair. The one standing exception is genuinely historic fabric, where restoration is its own discipline with its own justification.
The age-and-quantity math that decides whole houses
Single-window logic and whole-house logic are different calculations. One failed balance in a five-year-old production home is a warranty call or a cheap repair — nothing about it impugns the other nineteen windows. But when the windows are 30-plus years old and the failures have started arriving — a fogged unit here, a stiff sash there, a rotted sill on the weather side — each symptom is the product generation announcing itself, and repairing failures one at a time means paying retail mobilization on every visit while the next failure queues up. That is when planning a phased replacement beats whack-a-mole: prioritize the failing and weather-beaten elevations first, batch the work so mobilization is paid once, and get the pricing leverage of a multi-window order — the per-opening math in our Sacramento cost guide shows how strongly quantity moves it. The middle band — windows 15 to 25 years old with scattered failures — is where judgment lives: repair the outliers if the fleet is healthy; start planning if the failures cluster. And fold in the timing intelligence: if a re-side is on your horizon, windows and siding done together integrate flashing properly and split the mobilization — replacing windows the year after new siding wraps them is paying twice for the same wall.
Permits, energy code, and the repairs that dodge both
The code line between repair and replacement is worth real money. Like-for-like repairs — hardware, weatherstripping, a glass or IGU swap in the existing sash — generally need no permit and trigger no energy-code obligations. Replacement is different: swapping the window unit is an alteration under Title 24, which means a permit, current U-factor and SHGC targets, NFRC labels verified at final, plus today's safety-glazing and egress rules for the opening. That asymmetry cuts both ways honestly: a glass-only repair keeps a modest project modest — legitimately — while an unpermitted replacement 'to keep it simple' creates a disclosure problem that surfaces at sale. What it should never do is distort the diagnosis: replace-worthy windows do not become repair-worthy because repair skips paperwork. On the energy side, keep the claims honest too — replacement windows are a comfort and function upgrade with modest bill savings in most of California (the honest math is here), so 'the windows will pay for themselves' should never be the reason you replace early. ENERGY STAR's residential windows program and the Efficient Windows Collaborative are the neutral references for what new glazing actually delivers.
When repair is the right answer — and we'll say so
The honesty differentiator in this trade is the willingness to leave money on the table, so here it is in writing — the cases where repair wins and a replacement pitch should make you suspicious: a hardware failure on a window under 15 years old; a single cracked pane in a sound frame; one fogged IGU in an otherwise healthy set — swap the unit, matched to spec; flat weatherstripping and drafty meeting rails — new seals cost almost nothing against a window; a warranty-eligible failure on a name-brand window, where the manufacturer owes you the fix; and a genuinely historic window whose value is the fabric itself, where restoration beats any replacement. What repair cannot do is also worth saying plainly: it cannot make a 1985 aluminum single-pane window into an efficient one, cannot straighten a racked frame, and cannot outrun a fleet that has started failing on schedule. Our own test on a service call is simple — would we make this repair on our own house? If yes, that is the recommendation, and the replacement conversation waits for the year it is actually true. When that year comes, the decision stack is already written: install method, frame material, and glass spec, in that order. Either way, verify whoever does the work at the CSLB — repair jobs attract unlicensed operators precisely because they fly under the permit radar.
Window triage by symptom — repair or replace
| Symptom | Usual verdict | Why |
|---|---|---|
| Sash won't stay up / crank grinds / lock fails | Repair | Balances, operators, and locks are replaceable parts |
| Flat weatherstripping, drafts at the seals | Repair | New seals cost almost nothing against a window |
| Single cracked pane, sound frame | Glass-only replacement | The frame never has to know |
| One fogged IGU in a healthy set | IGU swap | Restores clarity and insulation without a new window |
| Repeated seal failures across the house | Plan replacement | A product generation aging out — swaps become a subscription |
| Rot in sash, sill, or frame | Replace | Bespoke frame carpentry loses; check the wall behind it |
| Painted shut, racked, out of square | Replace | Geometry can't be adjusted back |
| Every window 30+ years old, failures arriving | Phased replacement | Batch the work; pay mobilization once |
Key takeaways
- Sort every window problem into hardware, glass, or frame — hardware is repairable by definition, glass is swappable without touching the frame, and frame failure ends the debate
- Glass-only and IGU replacement are legitimate: a cracked pane or one fogged unit in a sound frame is a glass job, not a window sale — but match the coating spec and honor safety-glazing rules
- Rot, warp, and out-of-square are terminal: bespoke frame carpentry loses to mass-produced replacement, and rot usually signals water damage beyond the window
- The whole-house heuristic: one failure in a young set = repair; scattered failures in 30-year-old windows = plan a phased replacement and stop paying per-visit retail
- Repairs generally dodge permits and Title 24; replacement triggers both — a real cost difference that should inform timing but never distort the diagnosis
FAQ
Quick Answers
Yes — a cracked pane or a fogged insulated glass unit can be replaced inside your existing sash, at a fraction of full-replacement cost. It's the right call when the frame is sound and the failure is isolated. Match the new unit's low-E spec to its neighbors, and note that openings in safety-glazing locations must meet current code when the glass is swapped.
When the frame itself has failed — rot in the sash or sill, warp, or a frame racked out of square — or when failures have become a pattern across an aging set. Frame repair is bespoke carpentry against a mass-produced replacement, and systemic failures mean each repair just queues the next one.
Like-for-like repairs — hardware, weatherstripping, glass or IGU swaps in the existing sash — generally don't require one. Replacing the window unit does: it's an alteration under Title 24, with energy targets, NFRC label verification, and current safety-glazing and egress rules applied to the opening.
If the set is young and one window failed, repair it. If the windows are 30-plus years old and failing on schedule, batch the work: phased replacement pays mobilization once, earns multi-window pricing, and lets you prioritize the weather-beaten elevations. And if a re-side is coming, do windows and siding together — the flashing integration is cleaner and cheaper.
Sources
Authoritative references
- ENERGY STAR — Residential Windows, Doors & Skylights
- Efficient Windows Collaborative (NFRC) — window glazing: single, double, and triple pane
- National Fenestration Rating Council (NFRC) — window performance ratings
- Contractors State License Board (CSLB) — verify a California contractor
External links to government, code, and manufacturer sources. Sierra Siding is not affiliated with these organizations; references are provided for verification.