There’s a new tool that can help companies predict sales for the coming weeks, or decide whether to increase inventories or put items on sale in certain stores.
It’s Twitter.

Social-media sites such as Twitter have made it increasingly easy to find out what consumers think and want without the limitations and bias associated with older market-research tools such as surveys and focus groups. With Twitter, users broadcast what they are doing or thinking via “tweets,” short messages of 140 characters or less. People can “tweet” about anything at any time—from the long lines at the grocery store to a great sale at the mall to a new restaurant or movie—which allows for word-of-mouth to spread at astonishing speed. Anyone can follow a user’s messages, and tweets are easily searchable using keywords.

We believe executives can make accurate predictions about sales trends by analyzing tweets that mention their products or services, and we have created a model based on Twitter’s keyword-search function to help them do that.

What’s the Buzz?

Imagine a company is releasing a new product into the marketplace and has spent a lot of money on advertising to create a “buzz.” Our model would allow the company to track the buzz, determine whether the overall opinion is positive or negative and focus on specific areas of the country. The company could track the progression of tweets during and after the product’s launch to determine whether there are shifts in opinion, giving the company a chance to react quickly if there is a problem.

What’s more, if executives notice a sudden surge of tweets in New York City, signaling that people will go out and buy their product over the weekend, they may want to make sure stores in the area have enough stock. Inversely, if they notice that the buzz about the product is dying out, they may decide to put the product on sale, eliminate inventory and come