Our pricing structure is the biggest point of debate among us, however we are testing different models across a few different marketing platforms and will stick

Our pricing structure is the biggest point of debate among us, however we are testing different models across a few different marketing platforms and will stick with what performs the best in the long run. As for sentiment analysis, we were performing a rudimentary analysis on all non-rated mentions (Twitter, Facebook, etc.), however the algorithm we were using was providing too many wrong sentiments, so we removed it for the time being. We are currently evaluating new algorithms and APIs, and may bring it back in if we can find one that can comfortably handle sarcasm (a favorite of the review sites) and mentions with no sentiment attached to it (the phrase "I am eating at <restaurant name>" is neither positive, nor negative).