Co-Author: James Prashanth L A
In the decade bygone predictive analytics has driven the retail industry to garner higher revenue extraction per household. On the other hand, prescriptive analytics had a low-profile existence, though it had numerous advantages. The demand for prescriptive analytics is growing exponentially and is forecast to evolve into a $1.1 billion industry by 2019.
With this paradigm shift, it is barely startling to see the big names in retail closing stores. This leads us to unsurprisingly conclude that these brands are victims of their online retail counterparts. On the other hand, it is also apparent that online retail is here to stay.
Did you know?
An antithesis to this theory of flourishing online retail brands is the fact that retail brands is the fact that the world’s largest retailer Amazon is setting up physical stores at various locations
Both online and store retailers track customer purchase data, feedback, inventory availability, etc., but there is a wide gap in data processing and reacting to the market. Here, we would see how physical store retailers, if equipped with advanced prescriptive analytics and Big Data tools, could compete with the online retailers. An exodus to prescriptive analytics would unquestionably be the Holy Grail for store retailers.
B. Evolution of analytics
Analytical techniques can be broadly classified as in the image below:
It should be noted that as the value delivered by the analytic techniques increases, the complexity of the underlying models also increases.
All this has been possible with the change in technology landscape, which has led to,
- Faster data capture
- Power to process data
- Accuracy of the analytics
I) Descriptive analytics:
Descriptive analytics is used to understand trailing data or to summarize past events. These techniques usually involve simple statistical methods. The methodology can be reproduced to facilitate reporting or business intelligence techniques.
Did you know?
Descriptive analytics helps answer:
- "What happened?"
- "Which sectors did well?"
- "Which products maximize sales/profits?”
Common usage: Analyzing traffic patterns for the past one year between two locations. Usage could include identifying peak traffic hours, days, time calculations, etc.
Diagnostic analytics: Diagnostic analytics is used to identify factors that have contributed toward an outcome. These techniques utilize correlation analysis. Descriptive analytics techniques help in identifying outcomes, for which causal events need to be identified.
Did you know?
Diagnostic analytics helps answer:
- "Why did an event occur?"
- "What are the prime factors which have led an increase in sales?"
Common usage: Identifying the factors leading to peak traffic times such as office timings, holiday seasons, and game days.
Predictive analytics: Predictive analytics is used to identify outcomes that are likely to occur in the future based on past data. Descriptive and diagnostic techniques form the basis for predictive analysis and involve complex statistical methods such as Bayesian models and Hidden Markov models.
Did you know?
Predictive analytics helps answer:
- "Estimating revenue forecast for future"
- "Identifying the top selling brands/products for the next holiday season".
Common usage: The prominent users of predictive analytics are credit rating agencies in the US. For example, how probable is a customer likely to default. It should be noted that any future predictions cannot be made with 100% accuracy but these techniques help in identifying outcomes, which are most likely to occur. For example, estimating traffic between two locations in the next two hours.
Prescriptive analytics helps in recommending actions for achieving specific outcomes. These techniques help in predicting multiple futures based on a decision-maker's actions. With the help of identifying the outcome of each action, the decision-maker can then choose the optimum outcome and can execute the action. This is also considered the final frontier for data analytics.
Did you know?
Prescriptive analytics helps answer:
- "What should be done to achieve sales forecast, if a competitor drops pricing?"
- “What should be done to reach my destination on time if a traffic jam occurs?”
Common usage: Google Now recognizes the flight that you have booked, calculates the travel time from your current location to the airport, and recommends an approximate start time. It also warns of any changing traffic conditions.
Prescriptive analytics is highly effective since it emphasizes ‘do it now actions’ or ‘reactive actions’ without losing a valuable customer in the holiday season. With the help of proper in-store personnel training and real-time utilization of tools, the industry would get a boost.
C. What is the role of Big Data in prescriptive analytics?
The ‘3Vs of Big Data’ (volume, velocity, and variety) of Big data have enabled companies to process a massive amount of both, structured (internal) and unstructured (external) data in real time, in unprecedented manner & scale.
Some of the advantages of Big Data include:
I) Ability to process both structured and unstructured data: The ability to process unstructured data has helped companies harness the power of rich data in the form of social media, mobile data, tweets, and real-time feeds.
Did you know?
Big Data is only applicable to big players and does not work for SMEs
There is a popular myth that Big Data is only applicable to large and sizeable organizations to get ROIs. With the advent of ‘Big Data as a Service (BDaaS)’, there is reduction in upfront cost along with a decline in the ongoing cost of storage and management of large quantities of information. ‘BDaaS’ includes the supply of data and the supply of tools to analyze actual data. With this, the SMEs can have in-house storage of confidential proprietary information while making use of cloud services for nonconfidential data and data analysis.
II) Economical way to analyze huge amounts of data: The ability to economically process and store 2.5 petabytes (1 petabyte = 1,000,000 gigabytes) of information each hour has aided the processing of data in real time.
III) Emphasis on correlation over causality: Pre Big Data, the emphasis was on causal techniques that had a dependence on purified data. But with the ability to process huge volumes of data, the effect of outliers has been nullified. This has led to faster data-driven decision-making based on correlations.
IV) Faster feedback and response time: The ability to process and analyze data has also led to faster feedback. This reduced the response time and prevented companies from acting on stale data.
A mass departure to prescriptive analytics would indisputably be the Holy Grail for any retail organization. Irrespective of it operating in the online or store models, this exodus is inevitable.
The future is promising for any organization that adopts prescriptive analytics to reset their go-to market strategy. The ability of faster and objective decision-making increases the effectiveness of prescriptive analytics to a larger extent. The traditional retailers need to get their act right and come to terms with this reality, the sooner the better! Also, they need to realize the benefits of Big Data and be wary of the traps set by factors such as ‘existence of silos’ and ‘improper training of personnel’ which might lead to suboptimal results.
Finally, reflecting on the antithesis of Amazon’s physical retail stores mentioned earlier, we realize the crux for their success is the mastery of prescriptive analytics organization wide.