New Product Failure Rate

What the research shows

Castellion & Markham (2013) examined in the Journal of Product Innovation Management where the high figures come from. Their finding: the assumption that 80 percent or more of new products fail is a myth. Empirical studies since 1977 find a failure rate of 40 percent or less. The authors explain the persistence of the high figure by its emotional persuasiveness and by the self-interest of those who quote it.

How high the rate is in an individual case depends on industry, degree of novelty and definition. Some studies count only products that are withdrawn from the market, others also those that miss their sales targets. Most values are based on information provided by the companies themselves.

Why this matters for your decision

A realistic figure is more honest and still reason enough to act. If, say, two out of five launches miss their targets, every failure represents a budget for development, tooling, packaging, listing and marketing. Especially for expensive decisions that are hard to reverse, it pays to check demand beforehand.

The evidenced figure is also stronger for making the case internally. Anyone who argues with 80 or 90 percent risks colleagues questioning the source, and with it the whole argument. Even better than an industry figure is your own: how many of your company's recent launches reached their targets, and what evidence were they based on?

Example

A consumer goods manufacturer reviews its last ten new launches. Four clearly missed their sales target in the first year. For all four, a positive concept test was available before launch; no behavioural test took place. For the next launches, the team decides to test the final variants in a Painted Door Test before tooling is released.

How it differs

The failure rate describes an outcome after launch. Pre-launch tests, demand validation and the Behavioural Gate come in earlier and are intended to lower the rate. The failure rate of start-ups, often reported as the share of companies that fail, is a different measure from the failure rate of new products in established companies.

Limitations

A behavioural test before launch does not prevent every failure. It measures demand for an offer in one channel at one point in time. It does not reveal problems with quality, delivery capability, retail listing or repeat purchase. And not every failure is avoidable: some markets change between test and launch.

Before the investment, Horizon measures which variant real people choose, with a Painted Door Test in their familiar online environment. Nothing is sold.