Wednesday, September 4, 2013

Forecasting with Data Analytics and Crowdsourcing
I plan to discuss an fast evolving-alternate view to forecasting provided in the article “Cleaning the Crystal Ball”, which proposes efficient ways of creating forecasting models. The article makes some convincing points about forecasting when it says that forecasting models competence is not determined by its accuracy , instead by how well uncertainty has been accounted for. In order for the organizations to succeed at making better forecasting models it suggests having multiple people in the forecasting team, using historical data and embracing an open culture.

Internet-based Organizations these days are a step ahead at forecasting where they are using data analytics and predictive algorithms for forecasting purposes. This coupled with crowdsourcing has helped certain companies keep their inventory low and have zero unsold products.

For example the clothes apparel company Threadless, provides a platform for designers to showcase their work, which is rated by an active community of users. Threadless runs weeklong competitions and selects the highest rated designs to be printed on different apparel. Those who rate the designs, buy the same products they rated for. Threadless gets a live pulse on the design trends and gets an exact idea of the no. of tshirts to manufacture thus keeping inventory costs low. In addition to this Threadless is also present on various social media platforms where it engages with its customers. It has currently partnered with a data mining and an analytics solutions provider (JMeter) to integrate customers in a centralized location. They have gone a step further in their marketing efforts, by tracking their social interactions through Google Analytics and tying it on the back end to JMeter. This has helped them track customer interaction channel by channel, Track how customer preferences have changed over time and what factors influence them. As a result of which they have seen longer customer retention rates and increased consumer engagement.

There have also been more examples where e-tailers have used data mining and analytics to forecast what the customers would want to buy and when. Amazon, Target, Walmart are some of these who have used these technologies to better engage with customers and “sell them what they want”. Amazon has developed a recommendation which suggests things you could be interested in buying depending on what you search for or what you have bought before.

Applications of forecasting have been extended to beyond just predicting sales to effectively targeting customers and increasing revenues using data analytics.

References:

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.