Nobody shills for third-party ink here — the math either supports it or it doesn’t, and the answer is genuinely “sometimes.” Compatible and remanufactured cartridges sell at 30–60% discounts because they skip the R&D, the marketing and the margins — and because a small fraction of them arrive dead, clog early, or yield less than rated. Whether that’s a good deal isn’t a vibes question; it’s a one-line expected-value calculation.

The formula that prices the risk

The naive comparison — sticker vs sticker — overstates the third-party saving because it ignores failures. The honest version amortizes the failures into the yield:

effective CPP = price ÷ (rated yield × success rate)

If 5% of your third-party cartridges fail outright or die young, each one effectively yields 95% of its rating — you paid for 100% and got 95%. Run both cartridges through this one line:

• OEM: $24 ÷ (300 × 0.99) ≈ 8.1¢/page (OEM failure rate is ~1% — low, not zero)
• Decent third-party: $12 ÷ (300 × 0.93) ≈ 4.3¢/page
• Sketchy no-name: $9 ÷ (300 × 0.80) ≈ 3.8¢/page — cheaper still, until you price the hassle

The saving survives realistic failure rates, which is why third-party exists at scale. What the formula doesn’t price: the re-order round-trip for a dead unit, the print job that failed at 10 p.m., and the small residue of printer-damage risk (a leaking cartridge is a real, if rare, event). Those belong in the decision even if they don’t fit the formula.

What “failure” actually looks like

Not all failure modes cost the same. In rough order of frequency:

1. Chip/rejection errors — printer refuses to recognize the cartridge. Cost: the whole cartridge (you can sometimes re-seat or clean the chip, often not). This is the dominant failure mode and the one firmware updates occasionally make worse — a maker can’t brick your third-party cartridge, but a firmware push can tighten its chip check.
2. Early depletion / lower real yield — cartridge reports empty early or yields visibly fewer pages. Cost: a yield haircut (the ×success-rate term covers the average; a chronically low yield is a quieter version of the same tax).
3. Print-quality drift — streaks, banding, off colors. Cost: wasted paper plus the diagnostic dance to prove it’s the cartridge.
4. Leaks/clogs — ink in the wrong places, including inside the printer. Rarest, most expensive — the scenario where warranty arguments and actual damage both live.

When the math says buy

Third-party is usually the right call when all four hold:

• The saving survives a realistic failure rate. At 30–60% off, any failure rate under ~25% still wins on pure CPP — the bar is low; the question is whether your supplier clears it.
• The print jobs are fungible. Documents, drafts, invoices, labels — jobs where a mid-print failure is an annoyance, not a disaster.
• Volume makes the saving material. 4¢/page saved × 200 pages/month ≈ $96/year — worth the occasional dead cartridge. At 20 pages/month it’s $10/year — worth less than the time spent thinking about it.
• The supplier has a real warranty. Established remanufacturers replace dead units with a tracking number and no argument. “Ships from nowhere, 40 reviews” sellers replace nothing.

When to stay OEM

• Photo work and color-critical output — third-party ink can shift color and fade faster; when the page is the product, the ink isn’t where to save.
• Very low volume — the annual saving is pocket change; reliability and never-thinking-about-it win.
• A printer still in its warranty honeymoon — not because third-party voids the warranty (in the US it doesn’t, per Magnuson-Moss), but because you’d rather not litigate what caused a failure with a support agent holding a script.
• One-shot critical cartridges — the cartridge you need for tomorrow’s 200-page print run is not the slot machine to play.

The honest middle path

The play that captures most of the saving with least of the risk: OEM for the cartridge that hurts when it fails, third-party for the workhorse that doesn’t. For most inkjet owners that’s OEM black (or whatever channel is mission-critical) and third-party color; for laser owners it’s third-party toner with an OEM spare on the shelf. Then measure: note each cartridge’s actual page count and any DOA units, because your own failure-rate history beats every estimate in this article — and it’s the number the formula above wants anyway. Plug both cartridges into the calculator with your measured yields and the answer usually makes itself.

Frequently asked questions

Will third-party cartridges void my printer warranty?

In the US, no — the Magnuson-Moss Warranty Act bars makers from voiding a warranty merely for using third-party consumables (they'd have to show the cartridge caused the failure they're denying). Damage a bad cartridge actually causes isn't covered, which is the real residual risk — rare but not zero.

What failure rate should I assume?

Honest range: ~1–3% for established remanufacturers with their own QA, ~5–15% for the cheapest no-name compatibles. Your own history is the best number — if three cartridges from a supplier have all worked, their empirical failure rate for you is low; one early death in five tells a different story.

Is 'remanufactured' the same as 'compatible'?

No. Remanufactured = a used OEM shell, cleaned, refilled and (ideally) fitted with a new chip. Compatible = a newly-built cartridge made to fit. Both are third-party; reman tends to win on chip reliability, compatibles on price. Either can be excellent or awful depending on who's behind the label.

Why does my printer warn about 'non-genuine' cartridges?

Firmware checks the chip and complains — sometimes a dismissable nag, sometimes a block until you acknowledge. Warnings you can click through are part of the tax; firmware that refuses to print is a real cost (wasted cartridges). Check reports for your specific model before buying in bulk.

When should I definitely stick with OEM?

Photo printing where color accuracy pays the bills, mission-critical single cartridges (the one time a 100%-dead-on-arrival unit costs more than the saving), and printers still under a warranty claim period where you'd rather not argue. The math below quantifies everything else.