This indeed applies to all industries and all products. In short, how do you detect new uses of a declining product (cigars in this case) that could turn them into a good product and revive the industry in question? The tobacco industry has tried to sell outside US (especially Asia) to make up for declining sales in US. They came with e-cigarettes. But they never thought of a tobacco use in a context not associated with addiction.

Here's one: sell cigars that people (non-smokers) will smell rather than smoke. Based on a small experiment, on average women don't like the smell, men do. Smelling a cigar provides an hedonist experience very similar to finishing a great diner with a cognac or (if you are a smoker) a smoked cigar. It comes with several benefits: no smoking, replacing hard liquor (cognac) with an harmless product, and the ability to taste very expensive cigars since they last for a few weeks (as you never intend to smoke them - the cigar will keep its taste for a few weeks). And cigar producers or retailers can now target non-smokers and ex-smokers, a much bigger market. Pipe tobacco (with a sweet smell) could also be used in a similar way, that is, smokelessly.

So how does big data helps discover such new marketing opportunities? Maybe big data is not big enough yet and it's not a data issue. Or maybe if you analyze tweets and posts that talk about cigars, you will find opportunities like this one. Or maybe you can do a survey to identify potential new uses of your product by asking people. After all, this is something done all the time with pharmaceutical drugs: a drug initially designed to cure disease A, but later found very valuable to cure another disease B. 

Other links

• Data Science and Analytics Jobs
• Hadoop resources
• 17 short tutorials all data scientists should read (and practice)
• 10 types of data scientists
• 66 job interview questions for data scientists
• Our Wiley Book on Data Science
• Data Science Top Articles
• Our Data Science Weekly Newsletter
• Practical illustration of Map-Reduce (Hadoop-style), on real data
• What makes up data science?
• Data science webinars
• Data science competition