Test small ideas too, because 1 of them can be worth millions
Updated: 17 hours ago
Teams cannot easily predict which ideas will be successful.
The study. Ron Kohavi, who led experiments at Microsoft, and Stefan Thomke of Harvard wrote about online experiments at Microsoft, Bing and other companies.
The finding. An engineer at Bing suggested a longer ad headline. Nobody tested the idea for months. When the team tested it, revenue increased by 12%. This increase had a value of more than $100 million a year in the USA. At Microsoft, about 1/3 of tested ideas helped. About 1/3 had no effect, and about 1/3 had a bad effect.
Why it matters. Teams choose ideas by opinion. But experience shows that most ideas do not give the expected result. Some small ideas give very good results. With cheap tests, the data can make the decision.
The takeaway. Make tests cheap and fast, so that you can test small ideas as well as large ones.
Want to read more?
Subscribe to plaintalkbusiness.com to keep reading this exclusive post.
