Throughout my career I have been in contact with many buyers from big retail organisations, while also being involved in making sales forecasts for B2C products such as video games. Buyers (a.k.a. Procurement Managers) and myself had to estimate sales potential based on historical data of comparable products, marketing budget available (which also depends on sales potential), feedback from early testers such as journalists and then apply our own "gut-feeling" to ultimately decide on a figure.
Now, from a Marketing point of view you can obviously generate extra sales via smart and targeted campaigns, but the relation between marketing spent and sales revenue is not a linear equation. In the case of entertainment products it is possible to spend huge amounts on communication and advertising and not even come close to your sales targets. In the music industry there is no set recipe for a "hit" (Stock, Aitken & Waterman however seem to have had the recipe for the late 80s) and in the video games sector new franchises/titles are notoriously difficult to evaluate. The specialised games press has always been considered a very important source of feedback, but still a great score is no guarantee for commercial success and vice versa.
Nothing new so far. But social media monitoring could help organisations with an early warning system that would give an indication as to the potential of products. If the opinions formulated in social media are mostly negative, organisations might be more prudent as to the forecasts and the marketing budgets linked to these products
New social media monitoring tools, such as Techrigy's SM2, now allow you to dig much deeper into the "overall feeling and sentiment" about a product's release, before the actual launch of the product. Some algorithms will even analyse the Positive versus Negative Opinions and represent these graphically (e.g. pie chart: 500 opinions of which 75% positive, 25% negative). Now, computers obviously don't understand sarcasm so a sample of these opinions should quickly be scanned manually to see if the figures are representative but this tiny aspect of social media monitoring could prove to be extremely useful to organisations when forecasting sales of entertainment products.
Such data could then be incorporated in a marketer or buyer's set of forecasting tools. Over a period of 12 months enough historical data could be collected in order to start using Positive/Negative Opinions as a standard decision weighing tool.