>>Shingai, my definition of a growth hacking is data-driven product design that causes your customers to bring you more (new) customers<<
how is this different than normal product design and marketing? P&G has been doing this for more than 1/2 a century.
What is it about the product design of Tide laundry detergent or NyQuil that causes your customers to generate new customers for those P&G products? Answer: nothing. If a product works well, you might get word-of-mouth recommendations for those products, but that is not really a product design feature.
Compare that to FounderDating, where the design of this discussion forum is that by default, posts are broadcast the the poster's personal network via LinkedIn, Twitter, and Facebook. That is a product feature that primarily benefits FounderDating in drawing new people into the discussion and through that into FounderDating. It is a classic "growth hack".
Growth hacking is data driven in that hard data is used to determine whether and how well a growth hack works and to learn about the customers in the process. Pretty much all design is trial and error; the growth hacker has the data to determine what was an error and generate a plausible hypothesis as to what when wrong that can be further tested. P&G for the most part doesn't know if their ads on YouTube increased sales of NyQuil and if so, why. They have done focus groups and other kinds of testing, but they don't have a complete feedback loop that ties product design changes to desirable consumer actions; growth hackers do.
I, as a Growth Hacker, have great respect for what P&G (or especially Disney) can do in traditional sales and marketing. I am not even close to as good as they are in that. What I can do that they cannot is suggest and validate specific product changes (not pricing or package design or promotions) that lower customer acquisition costs while improving (or at least not harming) customer retention and satisfaction.
I can also analyze customer behavior data to identify potential customer issues and design targeted interventions to address them. I'm not as strongly committed to the idea that this is unique to growth hacking, but it is at the least a natural extension of it. Traditional marketers are moving in this direction as the availability of data allows. Target, for example, famously did it too well in targeting newly pregnant women.
So like many fields, the borders are fuzzy and there is overlap. To me, the bright line is that the growth hacker makes changes to the product itself and the result is that the customers themselves are the ones doing he marketing.