Penetrating Oil vs. Film Finish for Outdoor Furniture: Service Life, Cost, and How Each Fails

You finish the bench in April, and then you have to decide what goes on it. That's where the trouble starts, because the internet will hand you two confident opposite answers about the same can. Same class of product, "lasts three to five years" on one page and "fails in a season" on the next, both written by people who sound like they've done it.

They both have. The contradiction is real, it resolves, and it resolves in a way that matters more here than it does most places. Nearly every published service-life number for exterior wood finish was measured on vertical, saw-textured siding. Your bench is planed smooth and has a seat and a tabletop aimed straight up at a sun that does not negotiate. Two documented multipliers stack against you before you open the can, and nobody quotes them next to the number.

So skip the adjectives. Go to the data sheets, the maintenance intervals, and the failure mode.

Two Ways to Die

The USDA Forest Service's Forest Products Laboratory draws the line hard in the 2021 Wood Handbook (FPL-GTR-282, Chapter 16). A penetrating finish does not peel. It fades, the pigment erodes, the wood grays, mildew moves in. Refinishing is, in FPL's words, "a simple matter of cleaning the surface and applying another coat."

A film finish, meaning varnish, spar urethane, anything that builds a skin, fails by cracking, blistering and peeling, and the trigger is usually adhesion rather than wear. FPL says films are "prone to peeling if the bare wood has previously been exposed to sunlight," and prone to blister and peel "if there is persistent high moisture in the wood underneath the film."

Then the line that should end most of these arguments. Writing about decks, and by extension every horizontal surface on your furniture, FPL says penetrating finishes "often last just as long on decks, are far easier to maintain, and because they do not peel, their failures do not attract as much attention."

Read that twice. Oil doesn't win on service life. It wins on how it loses. It fails quietly and gray. Varnish fails loud, in flakes, with a weekend of sanding attached.

One more from the same chapter, and it explains half the online shouting: "penetrating" is not a regulated word. FPL warns that products labelled penetrating can still leave a film, and gives you the test. If the surface reflects light, if there's a sheen, you have a film, and a film can peel. FPL also notes a second coat can form a film even when the first didn't.

The Numbers Fight Each Other Out Loud

Three respectable sources, three answers.

FPL's 1999 edition (GTR-113, Table 15-4) listed clear varnish, three coats minimum, at "2 years or at breakdown," maintenance cost High. The 2021 edition's body text says the best clear exterior coatings "last about ten years when fully exposed," and its maintenance table dropped clear varnish entirely. Purdue Extension's NCR-135 says that in severe exposures clear film finishes "may last barely one year," and that the film will crack and peel regardless of how many coats you apply.

Two years, ten years, one year, same product class. FPL also revised itself upward across editions on nearly everything, semitransparent stain from 3–6 years to 4–8, paint from 7–10 to 10–20. The text doesn't say why, so I won't guess.

The disagreement isn't about facts. It's about what "lasts" means. Ten years is "hasn't peeled off yet." One year is "still looks like a finish." Decide which one you're buying.

The closest thing I found to a controlled multi-summer head-to-head is Practical Sailor's exposure rack: 18 products, four coats each regardless of the maker's own schedule, on a south-facing frame set at 45 degrees in Sarasota, Florida. At 24 months, not a single one-part varnish rated better than Fair for coating integrity or gloss retention. No varnish alternatives or teak oils were recommended at two years either. At 12 months they'd already pulled Star brite Tropical Teak Sealer as well past due for a recoat, and TeaQua for black mildew spots in the failure areas. It's a data point against both marketing claims at once.

The Multipliers Nobody Quotes

Smooth wood halves it. FPL 2021 on semitransparent stains: on smooth wood, "expect approximately half the service life compared with saw-textured surfaces, simply because more finish gets applied to textured wood." Their headline version is two coats that may last 10 years on rough or saw-textured wood versus 2 to 4 years for one coat on new smooth wood. Furniture is planed smooth. Start at the bottom of every published range.

Angle does the rest. Williams, Knaebe, Evans and Feist at FPL published erosion rates in Wood and Fiber Science for bare wood exposed near Madison, Wisconsin at 90 degrees, 45 degrees and flat. Douglas-fir earlywood eroded 42 µm a year vertical, 97 at 45 degrees, 81 flat. Their northern red oak numbers extrapolate to 3 mm a century vertical, 7.5 mm at 45 degrees, 11 mm horizontal. Water and dirt sit on the flat stuff, and total radiation is higher on tilted and horizontal specimens.

Keep the caveat honest: that's bare wood in Wisconsin, not coated wood in Central Texas. It gives direction and rough magnitude, nothing more. But the direction is the point. The tabletop and the seat are a harder job than the legs on the same piece of furniture.

That split shows up in industry's own paperwork. Cabot's Australian Timber Oil sheet (#19400 Series) says the finish "typically lasts 1–3 years on decks and 2–4 years on siding and fences." A manufacturer publishing its own horizontal-versus-vertical spread, with horizontal on the short end.

One more, free: FPL reports that western redcedar preweathered for one week before painting started peeling after 13 years, boards left out 16 weeks needed refinishing within 3 years, and unexposed panels were still sound at 20. A week of bare wood in July is not nothing. FPL notes the reverse for oils, which last longer on sunburned wood than on fresh.

Read the Data Sheet, Not the Can

Epifanes publishes the honest cost of a film. Their Clear Varnish sheet specifies a seven-coat system: first coat thinned 50%, second 25%, third 15%, then four more at 0–5%, with 24 hours and a sanding between each. Then this: "In warmer climates with high UV-levels, a minimum of 6 or more build-up coats is required for sufficient protection." For Austin that's three sealer coats plus six build coats. Nine coats, nine days minimum, on one chair. SMS Distributors, passing along the maker's instruction, puts annual maintenance at one to two more coats.

Minwax's 350 VOC Helmsman Spar Urethane sheet lists outdoor furniture as an intended use, calls for three coats outside, then says it's "not recommended for large exterior surfaces where maintenance would be difficult, such as decks and siding." Their product page goes further than marketing usually does: "No clear exterior finish will permanently protect wood against direct sun, snow, ice and rain. Be sure to check the finish annually." Worth knowing the same PDF disagrees with itself, one table giving 400–500 sq ft per gallon at 1.0 mil dry per coat, another giving 400–500 per quart at 2–3 mil. Not a gotcha. Just a reminder that data sheets are written by people.

Now the oil side, and notice what's absent. PPG's ProLuxe Cetol SRD RE sheet says one coat only, do not thin, wood at 18% moisture content or less, and for maintenance, "apply a maintenance coat when visibly required." No interval in years. Penofin's Blue Label sheet lists furniture, claims non-film-forming and 90% UV protection, wants a 20 to 30 minute dwell and then the excess wiped off with clean lint-free cloths, and prescribes cleaning twice yearly. No service life in years anywhere on it. Rubio Monocoat's Hybrid Wood Protector: 0 g/L VOC, not film-forming, QUV tested to EN 927-6, no recoat interval published.

Osmo published one and then took it back. The 06/15 UV-Protection-Oil Extra sheet said recoat "after approx. 3-4 years." The current 02/25 sheet drops that figure and says inspect at least annually, clean and recoat once thinly if necessary. The old number still circulates on retailer pages. It is not Osmo's current guidance. Osmo also wants two thin coats for its claimed UV protection factor 12, tested to EN 927-6, and says one coat makes the life of the coating significantly shorter.

The Teak Standoff

Two companies that both sell wood products, flatly opposed. Kingsley-Bate's care guide, as published by the retailer AuthenTeak, says "DO NOT USE OIL on teak furniture that is used outdoors. Oiling outdoor teak may cause mildew or irregular coloring," and that doing it improperly isn't covered under warranty. Their position is that periodic cleaning is the only maintenance needed, that outdoor teak takes about nine months to go silver-gray, and that the checking you'll see along the way is cosmetic and doesn't affect structure. Country Casual Teak goes the other route, selling a water-based sealer on a two-coat initial application plus a single recoat at yearly intervals.

Against both, Cabot and Penofin list outdoor furniture as an intended substrate, and Penofin advertises superior mildew protection.

That's not anyone lying. It's finish makers assuming the finish gets maintained and furniture makers assuming it won't, writing warranties around the assumption. Be honest about which customer you are.

What It Costs and What I'd Do

Prices I checked on 26 July 2026, one retailer each, one day. The Stain Depot lists Cetol SRD at $67.95 a gallon. DecksDirect lists Penofin Blue Label at $54.99, discounted to $46.74. SMS Distributors sells Epifanes Clear Gloss at $44 a litre. Rockler wants $24.99 for 100 ml of Rubio Hybrid Wood Protector, stated to cover 25–30 sq ft.

My arithmetic below, not quoted figures, using only those prices and the sheets' own theoretical coverage with no waste allowance:

• Cetol SRD, one coat on smooth hardwood: $0.17–0.27 per sq ft
• Penofin Blue Label at list: $0.18–0.31 per sq ft
• Rubio Hybrid, one coat: $0.83–1.00 per sq ft
• Epifanes full seven-coat system: about $2.05 per sq ft, or about $2.64 at the nine-plus coats they specify for high-UV climates

Three summers on the same chair: Cetol once plus two maintenance coats works out around $0.51–0.81 a square foot. The Epifanes system plus two annual maintenance rounds runs about $2.63–3.23. Four to six times the material cost, and material is the cheap half. The real gap is nine 24-hour cure cycles with sanding between each against a wipe-on coat, a twenty-minute dwell and a rag.

Film finish is the right tool for brightwork and for anything living under a roof, where you're willing to baby it on schedule. For a table sitting in the open here, penetrating oil, applied thin, wiped back, and maintained before it looks like it needs it.

Minwax says check annually. Osmo says inspect at least once a year. Cabot says evaluate each spring. Practical Sailor's real-world advice for one-part varnish is an annual maintenance coat "whether they look like they need it or not." Nobody's data sheet says wait until it fails. Put it on the calendar in March, not in your peripheral vision in September.