Weibull Analysis: How We Precisely Measure Product Lifespan

Why the average is not enough

When someone says "the average washing machine lasts 11 years", most of us take that as fact. But the average is deceptive.

If 10% of washing machines last 20 years and 30% break within 5 years, the average looks acceptable — even though a third of customers end up with a product on the edge of quality. The average hides the spread. And the spread is exactly what consumers care about.

That is why IKOR uses the Weibull distribution.

What the Weibull distribution is

The Weibull distribution is a statistical model that describes the probability of failure over time. It was originally proposed in 1951 by Swedish mathematician Waloddi Weibull for analysing material strength — today it is used by the aerospace industry, car manufacturers and insurance companies alike.

The model works with two key parameters:

• β (shape) — describes how the risk of failure changes over time
  • β < 1: products fail most often immediately after purchase (manufacturing defects)
  • β = 1: the risk of failure is constant over time (random)
  • β > 1: the risk of failure increases with age (wear, obsolescence)
• η (scale) — the characteristic life; the time by which approximately 63.2% of products will have failed

For most consumer appliances, β comes out between 1.5 and 3 — meaning products "fatigue" and fail with increasing frequency as they age.

How IKOR calculates the Durability Score

Input data:

1. Results from expert tests (consumer organisation reviews, Stiftung Warentest, test laboratories) — test scores as percentages
2. User ratings from e-commerce platforms — review count, average rating, proportion of 1-star reviews
3. Guaranteed spare parts availability (per Ecodesign regulations)
4. Service data from available fault databases

Calculation:

Ds = 0.40 × T_test + 0.30 × T_user + 0.20 × T_repair + 0.10 × T_parts

Where:

• T_test = normalised test score (0–100)
• T_user = adjusted user rating (filtered for extreme reviews)
• T_repair = repairability score (iFixit + parts availability)
• T_parts = spare parts availability after 5 and 10 years

The resulting Ds is a number from 0 to 100. Products above 70 points are designated IKOR Recommended.

Where the method has limits

Transparency is central to IKOR — which is why we also acknowledge the weaknesses:

• Test data is a snapshot in time. A product tested in 2022 may have undergone a design change since.
• User reviews are biased. Dissatisfied customers write more.
• Weibull requires a sufficient sample. For less widely sold products we have less data.
• The methodology evolves. The current version is 1.2 — we refine it with each new dataset.

The complete methodological document (in Czech and English) is available on the Methodology page.

What we are planning next

In 2026 we are launching Durability Score Phase 2 — incorporating direct measurements from repair centres and warranty claim data from selected retailers. If you would like to contribute data or become a research partner, write to us at info@durability.institute.