Predictive analytics is a way of forecasting future probabilities based on past data.
With big data now available covering many industries, businesses are increasingly able to make predictions for the direction of their business based on a combination of business intelligence and Big Data.
Predictive analytics is a discipline that can encompass machine learning and artificial intelligence to make predictions about future trends.
It’s a key tool in the arsenal of international business, helping global organizations set prices, predict demand and identify opportunities. But, as well as the benefits, businesses need to understand the limits of predictive analytics and be prepared to adapt their business models based on the signals provided. Responding to the demands of utilizing predictive analytics effectively is a major challenge for modern organizations.
Although industry is getting much better at data analysis, predictive analytics is still only a way to guide business decisions via educated guesswork. Although your business can make excellent decisions within the available data set, there will always be factors outside your control.
It’s also the case that predictive analytics can only indicate trends based on available historic data. Data is unlikely to be able to predict the emergence of disruptive forces that change the entire industry landscape. An example might be the advent of new technology that could not easily have been predicted based on past data.
Predictive analytics for international organizations
Working with predictive analytics can bring advantages to businesses operating internationally, provided they are prepared to segment their audience based on past and future likely behaviour.
For larger global organizations that are generating and using big data, predictive analytics is a way to make sense of the confusion and serve their customers most effectively.
Predictive analytics is especially useful for larger organizations operating across a number of different territories.
Once an organization reaches a certain size and complexity, it no longer becomes possible to make effective operational decisions without employing some predictive modelling techniques. Machine analytics can be used to improve efficiency in large organizations to anticipate future demand, manage inventory and ensure sufficient resources.
Predictive analytics is helpful because it removes the tendency for humans to make assumptions about the future based on their own experiences and bias. That’s important for global companies where the key decision makers are not from the environment that the decisions will affect.
Tesco was one of the first big grocery retailers to embrace Big Data, and this helped it expand its market share significantly in the first decade of using this new approach.
Tesco used the data it obtained from customer loyalty schemes (‘Tesco Clubcard’) to segment its customer base and launch new product lines based on likely demand. It also used highly targeted mailings based on customer consumption patterns. By the end of 1999, over 145,000 variations on the standard mailing were being sent out. Tesco also used predictive analytics to reduce waste in its stock and hence improve the bottom line.
Although Tesco leveraged data analytics to triumph in its home market during the 1990’s, the company was less successful as it expanded into the US market.
It launched the Fresh & Easy convenience store chain in the States just as the economic crash occurred. This was unfortunate timing and the company eventually ended up writing off $3bn on the failed venture. Data analytics was less helpful to Tesco when it came to constructing an entirely new venture in an unknown market. There were several problems with the Fresh & Easy offering but these included US customer resistance to self-checkout tills: this wasn’t as easy to predict using data when Tesco had no previous experience in this market.
Tesco was also unable to predict the rise of rock bottom food retailers such as Lidl and Aldi, which also stole a march on their territory. These discount grocers also managed to offer a simpler shopping experience than Tesco.
Customers didn’t need to remember their Clubcard or paper vouchers in order to access cheaper groceries in these stores. Although being an early adopter of data analytics helped Tesco gained ground at first, it couldn’t save the company from ill-advised business decisions and external forces.
Reducing staff turnover
One other area of working life that is benefitting from predictive analytics is the hiring process. Deloitte helped a global organization reduce turnover and predict employee attrition in one specific part of its business: China.
A model was built that could identify the likelihood of each employee leaving the business, based on a large number of factors. This meant the company had access to a retention score out of 100 for every individual in every team. This was far more significant that just being able to reduce turnover: it helped the company predict resource availability in key teams to support their ongoing strategy in the region.
There was a strong element of human input into this strategy. As well as employing statistical modelling and visualization software, the researchers also conducted interviews with employees. Once the data was modelled, decision makers could then use the retention score as a basis for decisions such as retention strategies for Chinese employees predicted as likely to leave in the near future. This approach didn’t just use quantitative data but also used the qualitative angle of getting to understand the individuals behind the data more.
One of the key challenges of predictive analytics is perhaps not the one you might predict.
The main barrier to the adoption of predictive analytics in businesses is actually a lack of human expertise.
Data scientists are already in short supply and commanding high salaries, and the shortage is only going to get worse in future. It’s a difficult field of knowledge requiring specialist insight, and the skills shortage is a significant limiting factor to the industry.
When it comes to using predictive analytics successfully, companies can often make mistakes such as failing to ask the right questions of their data, or not having a specific goal in mind for the analytics programme.
It’s also very common for companies to fail to implement the findings of their predictive analytics programme simply because they find it hard to move away from old ways of decision making. Predictive analytics isn’t the final answer for businesses but it is a key tool that, if used wisely, can afford significant insights to support the business.