ProjectPriced
Home / Guides / Roofing

Repair or Replace Your Roof? A Homeowner's Decision Guide

By the ProjectPriced Team · Updated September 2026 · Written to our editorial standards

Most homeowners ask "how bad is it?" when the real question is "how much life is left?" Here's how a roofer actually makes the repair-or-replace call — and when spending money on a repair is the wrong move.

I've turned down more repair jobs than I've taken. Not because I don't want the work — because on a lot of roofs, a repair is the wrong answer, and the homeowner just doesn't know it yet. They're looking at one leak or a patch of missing shingles and thinking "that's a $600 problem." Sometimes it is. Sometimes it's a $600 down payment on a $9,000 problem you'll pay for anyway in eighteen months.

So let's do this properly. Repair versus replace isn't a gut call and it isn't about how the roof "looks." It's a decision you can reason through with three inputs: how old the roof is relative to its material's real lifespan, how widespread the damage actually is, and whether the repair cost is a sane fraction of what a new roof would run. Get those three straight and the answer usually picks itself.

Start with age, not the damage

The single most useful number you have is not the size of the leak — it's the age of the roof measured against what that material is actually supposed to give you. A missing shingle on a 6-year-old roof and a missing shingle on a 24-year-old roof are completely different conversations, even though they look identical from the driveway.

Here are the working lifespans I use. These aren't warranty numbers (warranties are a legal document, not a prediction) — they're realistic service life in a normal US climate. Hot southern sun, coastal salt, and hail country all pull the low end lower.

Roofing materialRealistic service lifeNotes from the field
3-tab asphalt shingle~15–20 yearsThe builder-grade default. Ages fastest; loses granules and gets brittle.
Architectural (dimensional) asphalt~25–30 yearsThicker, heavier, worth the small upcharge. Most common today.
Standing-seam / metal~40–70 yearsFails at fasteners and sealant long before the panels do.
Clay / concrete tile~50+ yearsTiles outlive the underlayment beneath them, which is the real clock.
Slate~60–150 yearsEffectively permanent; you repair, you don't replace.

Notice the pattern with the long-lived materials: the roofing surface isn't usually what dies. On tile and metal, it's the underlayment, flashing, and fasteners. That matters, because it means an old tile roof leaking at a valley is often a genuine repair, while a 22-year-old 3-tab asphalt roof leaking anywhere is usually telling you the whole system is done.

The rule of thumb I lean on: once a roof is past roughly 75–80% of its expected life, start leaning toward replacement — even for damage that would be an easy repair on a younger roof. Why 75–80%? Because the failure curve isn't linear. Asphalt doesn't degrade a little each year and then fall apart on schedule; it holds up reasonably well, then enters a phase where the mat gets brittle and problems start compounding — one leak, then another six feet away, then a third the next season. Paying a tradesperson to come out three times in two years costs more than one clean tear-off, and you've been living with leaks the whole time. If your architectural roof is 24 years old, you're at that ~80% mark. Plan, don't patch.

Then measure the damage: localized or widespread?

The second input is honest scope. There's a real, bright line between localized damage and widespread damage, and homeowners routinely misjudge it because they only see the part that's leaking.

Localized means a discrete, identifiable cause affecting a small area: a dozen shingles blown off in one wind event, a single cracked pipe boot, one length of failed step flashing against a wall, a branch that punched through. The rest of the roof is sound. This is repair territory — assuming the roof is young enough to be worth repairing.

Widespread means the deterioration is a property of the whole roof, not one spot: granule loss across multiple slopes, curling or cupping shingles you can see from the ground, leaks appearing in different, unrelated areas, or the same problem recurring after you've already fixed it once. That's not damage — that's aging. You can't patch your way out of aging.

Here's the tell most people miss: where are the leaks? One leak, one location, one cause — repairable. Multiple leaks in unrelated parts of the house over a season or two almost always means the underlayment or the shingle mat has given up system-wide, and every fix just chases water to the next weak point.

The cost math that settles it

Now put money on it. Get a real replacement number for your specific roof before you decide anything — square footage, pitch, number of layers to tear off, and your material all move the price. Ballpark full asphalt replacement in most US markets runs roughly $5,500 to $14,000+ for an average single-family home, and a straightforward repair runs anywhere from a couple hundred dollars to about $1,500. Run your own numbers with a roof cost calculator and cross-check them against typical roof replacement cost ranges and itemized roof repair cost figures for your region so you're negotiating from real data, not a contractor's first number.

The threshold I use, and that most honest roofers use: if the repair cost approaches ~30–40% of the replacement cost, stop and seriously consider replacing. A $2,500 repair on a roof that costs $8,000 to replace is 31% — that's the zone where you're pouring a third of a new roof into an old one and getting no new lifespan for it. Spend the $2,500 and you still have the same old roof, now with one fresh patch on it. Spend the $8,000 and the clock resets for 25+ years. The repair only wins if the roof has enough life left that you'll actually get years of value before replacement.

That's the whole logic in one sentence: a repair buys you time; a replacement buys you a new lifespan. Only pay for time if you have a lot of life left to protect.

A decision checklist you can actually run

Walk your roof (or your inspector's report) through this in order. The first row that lands you on REPLACE usually settles it.

QuestionLean REPAIR if…Lean REPLACE if…
How old is the roof vs its material life?Under ~70% of expected lifePast ~75–80% of expected life
How much of the roof is affected?Localized, one area, one causeWidespread across multiple slopes
How many separate leaks?One, single identifiable sourceMultiple, unrelated locations
Repair cost vs replacement cost?Under ~15–20%Approaching ~30–40% or more
Have you repaired this roof before?No — first fixYes — repair-on-repair
Is the decking (the wood under the shingles) sound?Solid, no sag or rotSpongy, sagging, or water-stained

That last row is the one that overrides everything. If the decking is compromised, you're not having a shingle conversation anymore — you're having a structural one, and it's replacement plus deck repair.

What the damage signals actually mean

Reading the roof correctly is most of the job. Here's what each common signal is telling you:

The insurance angle — this changes the math

Before you pay out of pocket, figure out whether this is an insurance event, because it can flip the entire decision. The line insurers draw is simple and unforgiving: sudden, accidental damage may be covered; wear and tear is not.

A hailstorm, a windstorm that strips shingles, a tree that falls on the roof — those are sudden events, and a standard homeowners policy will often cover the repair or even a full replacement of the damaged slopes. A roof that's simply worn out from 23 years of sun? That's maintenance, and it's on you. Wind and hail are the most common homeowners claims in the country, so storm damage is worth pursuing.

Practical sequence: if you suspect storm damage, document it before you touch anything — photos and video of the roof, the attic, any interior water entry, plus the date and the storm. File the claim before repairing; the insurer sends an adjuster and may decide replacement is warranted rather than a patch. Don't let a contractor talk you into fixing it first "to stop the damage" beyond emergency tarping, or you may forfeit the claim. And know that an insurer will look hard for an excuse to call your damage age-related on an older roof — clean documentation of the storm is your defense.

When a cheap repair is a false economy

This is where I lose jobs, and I'm fine with it. Some repairs are just bad money:

The honest test: after this repair, will I get several genuine years of service before I replace? If yes, repair. If the honest answer is "maybe one, and I'm crossing my fingers," you already know.

FAQ

Can I just replace one slope or one section?

Sometimes — if that slope took isolated damage (say, the windward side lost shingles in a storm) and the rest of the roof genuinely has life left. The catch is matching: new shingles won't match weathered ones, and you'll see the line. It's most defensible when insurance is paying for storm damage to specific slopes.

How do I know how old my roof really is?

Check your closing documents or ask the previous owner; permit records at your city or county building department often show the last roofing permit. If it came with the house and nobody knows, a roofer or home inspector can estimate age from shingle condition and wear.

Is a more expensive shingle worth it over 3-tab?

Usually yes. The jump from 3-tab to architectural is a modest upcharge for roughly 5–10 more years of life plus much better wind resistance. Compare the real numbers for your roof on the asphalt shingles roof cost page before you default to the cheapest option — the cheapest shingle is rarely the cheapest roof over time.

Should I get more than one opinion?

Always, especially if the first contractor jumps straight to full replacement. Get two or three written assessments, make them tell you the age, the scope, and the repair-vs-replace cost math in writing. A roofer who can't explain why in those terms is guessing or selling.

Sources

More ProjectPriced guides