Big Data in e-Commerce.How to Use the Power of Data in E-Commerce?
Tom Karwatka
Monitoring E-Commerce Today
• NC (new customer);
• RC (retained customer);
• ROI (return on investment);
• CLV (customer lifetime value);
• ROI CLV;
• RR (return rate);
• CR (conversion rate);
• CPO (cost per order);
• CPNC (cost per new customer);
• CPRC (cost per retainedcustomer);
Today, the majority of the e-commerce world monitors the following indexes:
Sources of Data in E-Commerce
• E-commerceOrders
Products
Baskets
Visits
Users
Marketing campaigns
Referring links
Keywords
Catalogues browsing
• Social dataFB
• Cookies / reMarketing / MA
• Google Analytics
• … and many others
The Choice of Data Source in Traditional Retail Is Even Greater
Source: http://www.slideshare.net/MarketResearchReports/big-data-1
Already in 2012 the Walmart transaction database was estimated to have 2.5 petabyte of customer data.
Questions the Analytics Can Answer
• What are the best sellers in a category?
• Is the most watched product at the same time the best selling one?
• Which products sell best among the users who have already bought an item in the product category?
• How often does a given user group (eg., new users) return to your shop?
• …
The problem is, however, that answering these questions does not lead directlyto a bigger profit.
Companies often get discouraged as the answers are difficult to apply in real life.
The Actionable Data
• Collaborative filtering
• Using the information on users' actions to automatically findthe correlations between:Elements on a websiteA keyword and the link chosen
• RecommendationsProducts
Offers
• ClassificationUsers who continue shopping
Applying the Big Data solutions makes it possible to analyse data in real time. This allows us to use the data not for reports only, but to translate them into action –usually personalized and in real time.
• RegressionIndicating trends or the lack of trendsPredicting stocksAnticipating a product's futurepopularityAnticipating the future popularityof promotionsAssessing the effect of marketing activities on sales or the numberof users
• Categorization and segmentation
Customers
Products
Example: Actionable Data
If, thanks to Big Data, we can find the correlation between the socialmedia and our system data, then taking into account that:
40% users purchased a product after liking or sharing it on social media
71% users of social media buy mainly based on recommendations
We can prepare shopping recommendations for specific customers, based on their social media behavior.
Example: T-Mobile
Source: http://www.slideshare.net/Dell/big-data-use-cases-36019892?related=1
• Billings, social media data
• Selecting clients for migration to
premium models
• Detecting clients with high Lifetime
Customer Value
Example: CREDEM Banca
• Predicting what products and
services will a customer like
• Increasing an average revenue on a
customer by 22%
• Marketing costs reducted by 9%
Source: http://www.slideshare.net/Dell/big-data-use-cases-36019892?related=1
Example: STARBUCKS
• Collecting the data about the
customers' orders
• Personalizing adverts
• Personalizing vouchers
• Selecting the customers losing their
interest in the offer
• Recovering lost customers
Source: http://www.slideshare.net/Dell/big-data-use-cases-36019892?related=1
Example: NORDSTROM
• Aggregating data from www pages,
social media, transactions, loyalty
program.
• Choosing a message based on the
customer's preferred
communication channel.
Source: http://www.slideshare.net/Dell/big-data-use-cases-36019892?related=1
Example: EasySize
Analyzing orders and returns – using the findings to decide whichsizes in different brands would fit a given person.
Source: http://easysize.me
Example: EasySize
Results: decrease in returns by 35-40%
easysize.me
Source: http://easysize.me
Example: Promotional Activity of Brands
The Kizzu app is available on iPhone and Android. Over 10.000 users enjoy the app. It gives the information on current promotions in the users’ shopping malls.
• Using a consumer mobile app, we collected the information on the special offers in shopping malls that customers findattractive.
• The data let us answer the questions:
Which brands have the highestpromotional activity?
Which special offers are the most effective?
Example: Promotional Activity of Brands
The free-of charge magazine for the customers of Deichmann is published twice a year - in spring and fall. It shows the latest fashion trends - very popular online
• Among the most popular special offers, we found also some less popular, nichebrands.Internet / Mobile gives them opportunity to competeagainst strong brands for the customers' attention.They attract customers, offering big discounts.
• Among the most popular special offersthere are frequently content basedpromotion activities (a promotionalnewsletter or a magazine).
• Activities targeting the most loyalcustomers are also popular.
• The number of promotional activitiesdoes not depend on the status of a brand. Our TOP 50 includes also some of the brandspositioned as premium ones. Their customersapparently expect a frequent interaction with the brand.
Future: Big Data & Design
• Continuing to use Big Data together with the automation of the layout creation- Responsive-web design- Font-end frameworks
• Creating user-customizedlayouts
• Case study: https://www.behance.net/gallery/22089487/Tchibo-Content-Automation-Platform
Source: https://www.behance.net/gallery/22089487/Tchibo-Content-Automation-Platform
Future: Big Data & Machine Learning
http://www.ibm.com/smarterplanet/us/en/ibmwatson/developer-cloud-enterprise.html
Three days in and we’re already acting like it’s been here forever. (…) Alexa can maintain two lists for you: To-do and Shopping List. Adding things is as simple as ”Add butter to shopping list” and „addng gutters to to-do list.” (…) Once you’ve added things to your list, you access them through the app.
One great thing is that everyone in your household who installs the app shares everything. So when I was at the store, my wife texted me that she’d put some things on the Echo shopping list. Sure enough, I opened my app and there it was. I could check off the things I got and they disappeared.
http://www.engadget.com/products/amazon/echo/reviews/14cw/
•IBM Watson - Developer Cloud EnterpriseMedical diagnostics support
Legal consultations
•GoogleGoogle Now – the first apps for eBay
DeepMind
•Siri, Cortana, Amazon EchoAmazon Echo already makes it possible to createshopping lists, among others
Future: Big Data & Machine Learning
• The assistant will deduct the products we areabout to need from a number of data, and willorder them autonomously.
• As far as the mass products go, the competitionwill become more and more difficult.
• The promotion of FMCG as we know it will stop being recognizable by the customers.
• The companies controlling e-assistants willbecome the biggest shopping portals.
• Basic competitive advantage will grow in importance – the product's availability, competitive price, and swift logistics.
• Internet will become just another layer of technology – little interesting for an averageuser.
Source: https://itunes.apple.com/us/app/fetch-personal-buying-assistant/id867636554
Future: Big Data & Machine Learning
• Right now, it is worth to develop new mechanisms for data exchange and offer creation automation.
• It is also worth to expand your own client databases, so as to keep in direct touch with your customers as long as possible.
• Owned Media!
Source: https://itunes.apple.com/us/app/fetch-personal-buying-assistant/id867636554
Thank You for the Attention
• Are you interested in Big Data?
• Let's talk!
20
Tom Karwatkahttp://[email protected]