Grain & Glue

How to Build Outdoor Furniture That Lasts

Rain is not the problem — repeated wetting and drying is. The glue, the screws, the joints and the one design rule that decides whether a bench survives five winters.

By Stephen V.Last updated How we pick

How this is funded: we earn a commission if you buy through our links, at no extra cost to you. It never changes which tool we recommend, and we’ll tell you plainly when to skip one. Full disclosure.

Outdoor furniture does not fail because it gets wet. It fails because it gets wet and then dries, over and over, for years. Every one of those cycles makes the wood swell and shrink, and every joint in the piece has to survive that movement while being held together by a glue line a few thousandths of an inch thick.

That is the whole design problem, and it reframes every decision. Three things change when a project moves outside: the glue (it needs a water-resistance rating, and a published one), the fasteners (uncoated steel rusts, and rust both weakens the screw and stains the wood permanently), and the geometry (every surface that can hold water will hold water, and that is where the rot starts). Species and finish matter too, but less than people assume, and in that order.

What to look for in a glue: the classes, and what they really mean

Wood glue water resistance is not marketing language, it is a tested specification, and the two levels you will see named are ANSI/HPVA Type I and Type II. The distinction is how brutal the test is.

Titebond’s Australian distributor describes the Type II procedure directly: glued blocks are cut into specimens, soaked for four hours, baked in a 120°F oven for nineteen hours, and the cycle is repeated three times, with no delamination permitted. Type I is described as the four-hour boiltest, called for “when substantial moisture is a concern”. Titebond states that Titebond III passes Type I and Titebond II passes Type II.

So the practical rule is simple:

  • In the weather — rain, snow, sprinklers: use a Type I glue.
  • Covered porch, screened room, somewhere it only gets damp: Type II is genuinely adequate and costs less.
  • Interior only: an ordinary PVA is fine, and wood glue explained covers why.

And the limit both makers state, which matters more than the class: these glues are not for continuous submersion or below the waterline, and not for structural or load-bearing use.Read that as an instruction about design, not about glue. The glue is never the only thing holding an outdoor chair together — the joint geometry and the fastener do that, and the glue seals and stiffens it.

One working note on Type I glue: Titebond states open assembly time of 8 to 10 minutes and total assembly time of 20 to 25 minutes at 70°F and 50% relative humidity, and an application temperature above 47°F. That is more working time than the interior version, which is useful on an outdoor piece with a lot of joints, and the temperature floor is the part people trip over — gluing up in an unheated garage in February is how a joint fails before it is ever rained on.

What to look for in a fastener

Outdoor furniture is fastened as well as glued, and the fastener does two jobs at once: it carries load, and it resists corrosion. Get either wrong and the piece fails in a way that is both structural and cosmetic.

Coating or stainless, never bare steel.An uncoated screw in damp wood rusts, loses cross-section, and bleeds a black stain into the wood around it — a halo that goes deeper than any sanding will reach, because it is a chemical reaction with the wood rather than a surface mark. Exterior screws carry a proprietary corrosion-resistant coating and a drive type that does not cam out; stainless costs more and is the right answer near salt air.

Structural screws where load is real. A bench leg, a table base or a seat rail is a load path, and a structural timber screw with a washer head and a published allowable load is a different product from a general-purpose wood screw. The pick below is the one our screw roundup already names for a workbench base, for the same reason.

And the pilot-hole arithmetic still applies.The US Forest Service’s Wood Handbook is specific about this: withdrawal resistance varies directly with the square of the specific gravity of the wood and with the depth of thread penetration, lead holes should be about 70% of the thread root diameter in softwoods and 90% in hardwoods, and — the figure that matters most outdoors — end-grain withdrawal averages only 75% of side-grain withdrawal even when splitting is avoided. Outdoor furniture is full of end-grain joints, so that 25% discount is being applied to most of the fasteners in the piece. The handbook is also blunt that screws should always be turned in and never started with a hammer, which tears the fibres and seriously reduces load capacity.

ToolWhy you need itPrice
Titebond III Ultimate Wood Glue, 16 oz (1414)
Titebond III Ultimate Wood Glue, 16 oz (1414)
The Type I glue: its maker states it passes the ANSI/HPVA Type I specification, with 8-10 minutes open time and application above 47F. Not for submersion or structural use.$9.58
Power Pro Premium Exterior Wood Screws, #8 x 2-1/2 in, 595 count, Bronze (48637)
Power Pro Premium Exterior Wood Screws, #8 x 2-1/2 in, 595 count, Bronze (48637)
Coated exterior screws for the general assembly — decks aside, this is the box for an outdoor bench, planter or porch table. Star drive, bit included.$46.99
Simpson Strong-Tie Strong-Drive SDWS Timber Screw, 0.220 in x 3 in, 50 count (SDWS22300DB-R50)
Simpson Strong-Tie Strong-Drive SDWS Timber Screw, 0.220 in x 3 in, 50 count (SDWS22300DB-R50)
The structural screw for a leg, a base or a seat rail — thick shank, washer head that cannot pull through, and a load rating published by the manufacturer.$32.86
Starter kit subtotal $89.43

Live Amazon prices as of October 5, 2026; they change, and we don’t display stale numbers — a tool with no live price shows “Check price” and is left out of the subtotal. #ad How we’re funded.

Design so water runs out, not in

This is the part that outlives every product choice on the page, and it costs nothing. Water does not damage wood by touching it; it damages wood by staying. Five rules:

  1. Nothing horizontal that can hold a puddle. Slope seat slats and table tops by a degree or two, or leave gaps between them so water falls through. A continuous flat tabletop outdoors is a shallow tray.
  2. Get end grain off the ground.A leg standing on wet concrete or soil wicks water upward through the end grain like a bundle of straws. Feet, a chamfer, a rail, a paver — anything that lets the end dry.
  3. Avoid joints that trap water against end grain. A rail butted into a post creates a horizontal seam holding water against a cut end, which is the single most common failure point in outdoor furniture. Either cap it, slope it, or design the joint so the seam faces down.
  4. Seal end grain, generously. End grain absorbs several times as much moisture as a face, so it takes several times as much finish. Two extra coats on every cut end is the cheapest durability upgrade available.
  5. Let the piece breathe. Gaps between slats are not a style choice, they are drainage and airflow. A tight-fitting outdoor bench stays damp in the middle for days after rain.

The wood-movement consequence is the other half of this. Outdoor furniture goes through a far wider moisture swing than anything indoors, so a solid top fastened rigidly across its width will crack itself apart faster outside than it ever would in a dining room. Wood movement explained covers the mechanism and the fastening methods that allow for it; outdoors, treat its warnings as doubled rather than optional.

Species: what to buy, and what not to bother paying for

The traditional outdoor furniture species are durable because of what is in them rather than how dense they are: cedar, redwood, cypress, white oak and teak all contain extractives that resist decay. Two notes that matter for a beginner buying at a home center:

  • Red oak is not white oak.They look similar on a rack and behave completely differently in weather — red oak’s open pores wick water along the grain, which is why it is an interior wood. Ask, and check.
  • Pressure-treated lumber is usually the wrong choice for furniture. It is sold wet, and it shrinks and twists noticeably as it dries. That is acceptable in a deck frame and a problem in a chair. If it is what you have, let it dry before you build with it — how to store lumber covers stacking it so it dries flat, and wood species for beginners covers the rest of the choices.

Finish: an honest answer, including what we do not stock

None of the five finishes in our picks is rated for exterior exposure. That is a real gap and it is stated here rather than papered over: our wood finishes roundupcovers interior products — wipe-on polyurethane, Danish oil, an oil-based topcoat, paste wax and a food-safe butcher block conditioner — and none of them belongs on a bench that lives in the rain. We would rather say so than recommend an interior finish for an exterior job.

What you actually have are three legitimate choices, and the worst outcome is hesitating between them:

  • Leave it bare. Cedar, white oak and teak weather to silver-grey and last for years untreated. This is how most garden benches live, it is free, and there is no maintenance to forget.
  • A penetrating exterior oil or stain. It soaks in rather than forming a film, so it fades rather than peeling, and it wants reapplying every one to two years. The right choice if you want the colour of fresh wood kept.
  • An exterior film finish — only with commitment. Spar varnish and exterior polyurethane look superb and are a maintenance contract: UV eventually cracks the film, water gets under it, and a peeling film traps moisture against the wood, which is worse than no finish at all. Excellent on a door you will maintain; a trap on a bench you will not.

Whichever you pick, the application principles do not change — thin coats, clean surface, extra attention to end grain. Those are in how to apply a wood finish, and the oily-rag disposal rule in shop fire safety applies to exterior oils exactly as much as interior ones.

The short version

  • Type I glue in the weather, Type II under cover, and never treat glue as the structure.
  • Coated or stainless fasteners, structural screws where load is real, pilot holes near the ends, and a 25% discount assumed on anything driven into end grain.
  • Design for drainage— slope the horizontals, lift the end grain off the ground, never trap water against a cut end.
  • A durable species if you can get one, but good design in ordinary wood beats bad design in teak.
  • Decide on the finish honestly: bare and grey, or oiled and maintained. Not a film you will not look after.

Build it with those five and a bench will still be sound when the hardware-store one it replaced has gone to the kerb.

Frequently asked questions

What is the best wood glue for outdoor furniture?

A glue that passes the ANSI/HPVA Type I water-resistance specification. Titebond III is the common one-part answer and its maker states it passes Type I, which is a four-hour boil test; Titebond II passes the lesser Type II soak test and is fine under a covered porch. Both makers state the same two limits: not for continuous submersion or below the waterline, and not for structural or load-bearing use — so the glue is never the only thing holding an outdoor chair together.

What is the difference between Type I and Type II water resistance?

They are different test severities. Per Titebond's Australian distributor, Type II is a soak test — glued blocks are cut into specimens, soaked four hours, then baked at 120F for nineteen hours, repeated three times, with no delamination allowed. Type I is described as the four-hour boil test and is called for when substantial moisture is a concern. Type I is the one to look for on furniture that lives in the weather; Type II is adequate for a covered or sheltered position.

What screws should I use for outdoor furniture?

Screws with a corrosion-resistant coating rated for exterior use, or stainless steel. An uncoated steel screw in wet wood rusts, and the rust both weakens the fastener and stains a black halo into the wood around it that no sanding removes. For joints carrying real load — a bench leg, a table base — use a structural timber screw with a published allowable load rather than a general-purpose screw.

Does outdoor furniture need pressure-treated wood?

Not usually, and often it is the wrong choice. Pressure-treated lumber is sold wet, so it shrinks and twists significantly as it dries, which is tolerable on a deck frame and unhelpful on a chair. Naturally durable species — cedar, redwood, white oak, cypress, teak — are the traditional furniture answer. Whatever you choose, the design matters more than the species: furniture that sheds water outlasts furniture that holds it, in any wood.

Why does outdoor furniture fail at the joints?

Because the joints are where end grain and trapped water meet. End grain absorbs water like a bundle of straws, so a leg standing in a puddle or a rail butted into a post wicks moisture into the joint and holds it there. Add repeated wetting and drying — which makes the two pieces swell and shrink against each other — and the glue line is loaded in a way no indoor joint ever experiences. Seal end grain, and design so water runs out rather than collecting.

Do you need to finish outdoor furniture?

You do not have to. Left bare, cedar and white oak weather to grey and last for years — that is how most garden benches live. A finish is a maintenance commitment: film finishes crack and peel under UV and then trap water underneath, which is worse than no finish, while penetrating oils and stains do not peel but need reapplying every year or two. Choose bare-and-grey or committed-and-maintained, and avoid the middle.

Sources

  • Titebond — Titebond III Ultimate Wood Glue product page — "The first one-part, water cleanup wood glue ever offered that is proven waterproof" — passes the ANSI/HPVA Type I water-resistance specification. Application temperature above 47°F; open assembly time 8-10 minutes and total assembly time 20-25 minutes at 70°F/50% RH; FDA approved for indirect food contact (cutting boards); dried film light brown; bond strength 4,000 psi on hard maple at room temperature. Same submersion and structural limits as Titebond II. Item 1414 is the 16 oz bottle (accessed September 18, 2026)
  • Titebond — Titebond II Premium Wood Glue product page — Cross-linking polyvinyl acetate. Application temperature above 55°F; open assembly time 3-5 minutes and total assembly time 10-15 minutes at 70°F/50% RH; passes ANSI Type II water resistance; "U.S. FDA approved for indirect food contact (cutting boards)"; dried film translucent yellow; shelf life 24 months. Not for continuous submersion, below the waterline, or structural or load-bearing use. Item 5004 is the 16 oz bottle (accessed September 18, 2026)
  • Titebond (Woodbond Adhesives, the Australian distributor) — ANSI/HPVA Type II water resistance: wood glues — "Type II testing is a soak test, involving cutting the 150 x 150mm glued block assemblies into 50 x 125mm specimens, soaking them for 4 hours, then baking in a 120°F oven for 19 hours. This is repeated for a total of three cycles, and the bonds must not de-laminate to pass the Type II specification." Type I is described as the "USA 4 hour boil test", called for when substantial moisture is a concern (accessed September 18, 2026)
  • The Wood Database — Wood and moisture — How boards take on and give up moisture, and why lumber must acclimate before it is built with (accessed August 19, 2026)
  • The Wood Database — Dimensional shrinkage — How wood moves with moisture — the cause of most beginner joint failures (accessed July 24, 2026)
  • USDA Forest Service, Forest Products Laboratory — Wood Handbook (FPL-GTR-190), Chapter 8: Fastenings — The "Wood Screws" section, pages 8-10 to 8-12. "The root diameter for most sizes of screws averages about two-thirds the shank diameter." Withdrawal "varies directly with the square of the specific gravity of the wood" and, within limits, "directly with the depth of penetration of the threaded portion and the diameter of the screw" — so "the longer lengths of standard screws are therefore superfluous in dense hardwoods." The withdrawal equation "is applicable when screw lead holes have a diameter of about 70% of the root diameter of the threads in softwoods, and about 90% in hardwoods." End-grain withdrawal, "when splitting is avoided," averages 75% of side-grain. For lateral load the screw must penetrate the member holding the point by at least seven times the shank diameter, and the side member should be about half that depth. "Screws should always be turned in. They should never be started or driven with a hammer because this practice tears the wood fibers and injures the screw threads, seriously reducing the load carrying capacity of the screw." Soap or a similar lubricant "is recommended to facilitate insertion, especially in dense woods, and it will have little effect on ultimate withdrawal resistance." Table 8-9 gives shank diameters by gauge: #6 0.138 in, #8 0.164 in, #10 0.190 in, #12 0.216 in. In particleboard, withdrawal "is not significantly different for lead holes of 50% to 90% of the root diameter," and "care must be taken when tightening screws in particleboard to avoid stripping the threads" (accessed October 2, 2026)
  • USDA Forest Products Laboratory — Wood Handbook (FPL-GTR-190) — The U.S. government reference on wood as an engineering material (accessed July 24, 2026)

Keep reading