The huge amount of data that is being generated by the internet enabled devices and machine has given rise to predictive analytics, the practice of building analytical models that interpret this data in order to predict the about “what can happen in future?”But this is not as simple as it looks like. Many challenges come across while doing predictive business analytics. There are lot of problems that comes in between while doing predictive analytics.1) Data: Data is the main source on which the whole analytics depends of there is anything which is missing in the data or if the data is not capable to answer the relevant question then your prediction can go very wrong and will lead to losses.2) Trending Market: In the predictive analytics what we work on is the past data which shows us the flow of the market and what kind of fashion is going in the market. But market’s fashion is something very unpredictable. It can change at any point of time be it a natural disaster or be it a man made disaster.3) Technology: If we talk about the technology part.For maintaining the data organisation needs better integrated machines with good hardware and software support the only one can maintain the data properly otherwise there will be glitches in the data which may lead to inaccurate predictions.4) Trust Factor: After analyzing the data we come to a conclusion which is called prediction. Of course, prediction is made by all companies, but it was going on Gut feelings. So, at times it is difficult to make the higher authorities believe about the results we have gotten. For example no one was prepared for demonetization, Increase in GST or Cyber hack attack. These are some challenges which are faced while doing predictive analysis according to me.And now if I talk about how we can overcome these challenges. They are as follows:1) Data: In order to work with the data for having better results one should take care of few things which are:* Verify all the variables which will be used in model.*Check for missing values, identify them, and assess their impact on the overall analysis.*Keep an eye on the outliers and decide whether to include them in analysis or not.*Choosing the relevant dataset that is representative of the whole population.*Watch out for any duplicates in the data. 2) Trending Market: It is difficult to deal with the changing market, but few things can be done to have better predictions.* Goals set by the predictions should not be long term. Setting up short term goals can help or prevent the company from losses. * One model is not enough for the life of the company there should be regular changes made according to the changing trend of the market..3) Technology: Getting updated with technology is something which can be done but yes it needs lot of investments to be called in.So if the company is adapting predictive analytics then they should get updated with there machines aswell.Then only the team can work on data properly.4) Trust: Gaining the trust is difficult too but can be done by working on the past data and predicting what would have been done at that time. Telling the things happened in past can gain some trust of the higher authorities which can lead to more predictive decision making rather than on gut feelings.