What percentage of restaurants fail after one year?

I love statistics that are quoted by almost everyone as fact, even though none of us know the basis behind them. Often they’re entirely wrong, like the percentage of US citizens who have passports. Another favourite is that “90% of restaurants go out of business in the first year”. A conversation where this came up prompted me to poke around.

A widely-quoted press release from 2005 is about an Ohio State University study which found that 26% of the restaurants in Columbus, Ohio failed in their first year, and 57-61% in their first three years. (Read the full study here.)

A study from 2003 of 4,006 restaurants in Dallas, Texas over a five year period found that around 23% of them failed in their first year. There’s a link to more on that study down the page.

Another link is to an undated page on the same site which says that “several years ago, researchers at Cornell University and Michigan State University conducted a study of restaurants in three local markets over a 10-year period.” They found that 27% of startups failed “after the first year” (I assume they mean “within the first year”).

But then this article also mentions “national statistics culled by Cornell and Michigan State universities” which “reveal that more than 50 percent of restaurants fail in the first year”. Other sites mentioning what I assume is the same study date it to 1991, and say the 50% figures is the failure rate after three years.

This is just the result of a quick Google, but it looks like “90% of restaurants fail in the first year” is an urban myth, and the figure is more likely to be around 25-30%. That said, as we’ve seen, restaurants continue to fail as the years go by and the 90% failure rate figure will obviously be reached if you extend a study over enough years.

Of course, if we could compare this figure to that for other industries it would be more meaningful. This article suggests 15% of businesses fail in the first year but doesn’t provide much background information and it’s bound to vary hugely from sector to sector. Any other suggestions welcome…

Comments