Descriptive, predictive and prescriptive analytics data are the three pillars of the stable eCommerce business development. You might not use these particular definitions, but you certainly already use at least two types of this data to improve your website efficiency.
In a nutshell, descriptive analytics is designed to analyse historical data, predictive analytics — to forecast future performance, and prescriptive — to develop a strategy for the predicted scenario.
Let’s take a closer look at each type of analytics and define for what specific purposes these are employed.
Analytics data as a way to improve customer experience
The global goal of analytics is to:
- identify and improve business weaknesses
- identify and enhance its strengths
- identify new effective solutions and find ideas on how to implement them.
In eCommerce, analytics tools now can measure all aspects of business: from operational processes to customer behaviour on a website. But especially when it comes to customer experience, data analysis is the most exciting science today. First of all, the reason lies in a customer-centric culture which retail companies rank as the most crucial factor of their success (higher than management), according to the Harvard Business Review.
Most large companies create a separate department focused only on the digital experience. For instance, such a department at ASOS has five teams: Product management, User Experience, Insights & Analytics, Customer Strategy & Insights and data science. Their deep learning algorithms for recommendations or identifying Customer Lifetime Value data really impress.
Despite the GDPR law, customer data is gathered by almost every eCommerce website. 40% of companies use it to customise or personalise experiences, 37% — to predict or anticipate consumer needs, and 20% of business operate it for creating an omnichannel experience. All these relate to prescriptive analytics which always includes descriptive and predictive analytics.
What is descriptive analytics?
This is easy to define the term ‘Descriptive analytics’ just because it comes from the word ‘Describe’. Basically, this is the statistics of your performance over a specific period in the past. In eCommerce, this may be all indicators in your Google Analytics account, such as conversion rate, churn rate, CPC within a specific ad campaign, average order value or the number of repeat sales — whatever. Or it may be your CRM data — revenue or total sales in May 2018, and so on.
Descriptive analytics allows you to monitor which of your implementations work better and generate more revenue, and which demonstrate poor results and drive your business downwards. Thus, this type of analytics includes two main stages:
- Data aggregation;
- Data mining.
Some of the purposes for which you can use descriptive analytics:
- Examine your actual audience;
- Get insights on consumer behavioural patterns;
- Understand the overall demand for your products and analyse demand within a specific category/segment/time etc.
- Estimate the effectiveness of marketing campaigns;
- Check demand for products via search queries popularity;
- Evaluate actual delivery costs and time.
- Compare indicators between different periods and so on.
Modern analytics software is mostly designed for descriptive analytics. With the help of measuring tools, you can receive reports on almost every customer action not only on your website but even in a brick-and-mortar store. For instance, you can track a heatmap with most viewed areas both on a product page and physical shop shelf. However, fashion is changing, and many tools now try to enhance their services with predictive analytics features.
The success of descriptive analytics significantly depends on your KPI governance. Carefully set and arranged goals are a solid basis for further efficient predictive and prescriptive analytics.
What is predictive analytics?
All flagship eCommerce companies highlight predictive tactics as a must for decision-making processes, pricing, shipping, marketing, and personalisation. As for a definition, predictive analytics is an analysis of the current and historical website and marketing performance, consumer behaviour, and purchasing patterns to forecast trends in sales and preclude risks.
If descriptive analytics requires skills to ‘reading’ figures and charts, predictive analytics calls for in-depth knowledge in interpreting these figures into an answer to the question “What is going to happen?”.
The historical data you managed to collect and process allows to:
- Determine the best price in the market;
- Improve website UX;
- Personalise promotions;
- Predict which products will under demand for each season;
- Forecast how many managers should support customers on Black Friday;
- Identify related products for best-sellers;
- Find ideas for A/B testing;
- Optimise stock;
- Enhance your actions on each stage of the sales funnel and so on.
The recent research from Dresner Advisory Services shows that only 23% of businesses employ predictive analytics when 26% of companies don’t even plan to use it.
What is prescriptive analytics?
Prescriptive analytics, which has become a buzzword in the marketing world, is the automation of your statistical finding in order to simplify your operational decisions and improve the future seamless shopping experience.
Here the algorithms come. They make possible such eCommerce tricks as:
- Recommend visitors the most suitable product on your website, which interested other visitors with similar behavioural patterns;
- Show different prices to visitors with a high and low average cheque;
- Control stock and notify you when something is running out;
- Determine what a user would buy next.
In other words, the third phase of business analytics allows coming up with concrete solutions for existing issues, forecasted during real-time and historical data analysis.
You Need Analytics To Automate Time-Consuming Processes
The future of ecommerce analytics connects with AI technologies. According to the Transparency Market Research, predictive analytics software has touched $6.5 B globally in 2019.
One of the top-ranking multifunctional software which performs at the junction of predictive and prescriptive analytics is the Google Cloud ML (Machine Learning) Engine. It offers retailers these five solutions:
- Visual product search makes it possible for online stores to integrate Google length-type capabilities into their mobile apps. IKEA. For example, it allowed users to take a photo of a household item to find it or a similar one in the online catalogue.
- Recommendations AI allows retailers to improve user experience with the product recommender system, which offers personalised products based on preferences and tastes of a particular customer.
- Contact Center AI is designed to build modern care experience with speech recognition and search technology from Google
- AutoML Table helps predict customer demand.
- Real-Time Inventory Management and Analytics allow tracking products availability across shelves, aisles, and stockroom.
Another tool for predictive analysis is the top-ranking international software Science and Microsoft R Open. Additional techniques could be put to work, such as advanced changes detection, core-and-chip technology, etc. The analytics service is used for customer sentiment analysis, spam detection and routing customer requests.
Profitect prescriptive analytics platform use algorithms that process 7 types of data:
- Inventory movement;
- Activities on every point of sale;
- Delivery and receiving;
- Logistics and warehouse;
- Planning and buying;
- Marketing performance;
- Circular commerce.
The tool creates its own scenarios what action to take in different situations. For example, the system notifies a responsible person about an SKU that is run out of stock. The scenarios can be modified for your needs. The company promises that their prescriptive analytics software can help retailers generate by 300% better ROI.
How Retailers Use Predictive And Prescriptive Analytics
One of the most popular case studies of using analytics algorithms is the Amazon’s patent ‘Anticipatory Shipping’ model. It processes data on prior customer purchases, their order frequency, cart contents, and search history to ensure a relevant product will be shipped to the nearest consumer hub. This software improves the delivery time and optimises shipping costs, helping the marketplace increase sales and customer experience.
ASOS presented another example of using data-based pricing software. The program tracks prices on competitor’s websites and other market data to inform the company managers on what products to stock, what price to choose, as well as when and how much they should be discounted. The fashion retailer says that with this tool, they managed to increase their sales by 33% for a year.
The three-level qualitative approach in analytics always shows excellent results in eCommerce business development.
Descriptive analytics takes most of the time — 60%-75% of the entire process. This phase requires thoughtful decisions for what data you will collect, how and where this information will be used, and what benefits it can bring for improving customer experience on your website. Mostly used for reports, historical data is a solid basis for predictive and prescriptive analytics because of insights on customer shopping patterns and your operational productivity.
Predictive analytics takes 20%-30% of the process. According to findings, it allows forecasting and model future events. This data is used for machine learning to make a prediction for the average spends within your key audience, CPC costs, price fluctuation, product demand, and so on.
Finally, prescriptive analytics, which takes 5-20% of the process, is designed to find automated solutions for forecasted issues. BI algorithms which are actively developing and improving now allow eCommerce owners to avoid human mistakes, delegating numerous activities. Robots now can control huge businesses and ensure personalised approaches to every customer. Without prescriptive analytics, we wouldn’t have Amazon or the Alibaba group, who promulgate the idea of consumer-centric culture, being trendsetters in this market.