Data Science for Digital Commerce
0 likes359 views
Data science can provide benefits for digital commerce companies through personalized recommendations, dynamic pricing, and search engine optimization. At Info Edge India, which operates job, real estate, and matrimonial sites, data science is used for applications like real-time recommendation engines, lead scoring, price trend analysis, and semantic search. Data science techniques employed include machine learning, text mining, natural language processing, and using big data technologies and nosql databases.
Data Science for Digital Commerce
- 1. Data Science for Digital Commerce MANISH GUPTA, PhD SVP-Analytics Info Edge India Ltd. (Naukri, Jeevansathi,99acres,Shiksha) Email: manish.iitdelhi@gmail.com
- 2. Outline Why Data Science for Digital Commerce? Issues & Challenges Data Science Applications for Digital Commerce – Recommendations – Dynamic Pricing – SEO Data Science Work Overview @InfoEdge – Live Demo Data Science Techniques 2
- 3. Why Data Science & Predictive Analytics for Digital Commerce 73% higher sales for companies which use Predictive Analytics than those who have never done it? 65% of customers feel completely frustrated with the bad customer experience, targeting/offers 60% increase in business margins and a 1% improvement in labor productivity for retailers who started using Big Data (McKinsey) 45% of online shoppers are more likely to shop on a site that offers personalized recommendation (invest Consulting) 68% in upsell and cross sell through Real time personalization. Many more reasons… 3
- 4. Social Media Statistics As of July 2015, total worldwide population is 7.3 billion and The internet has 3.17 billion users There are 2.3 billion active social media users 91% of retail brands use 2 or more social media channels Internet users have an average of 5.54 social media accounts Social media users have risen by 176 million in the last year 1 million new active mobile social users are added every day. That’s 12 each second Facebook Messenger and Whatsapp handle 60 billion messages a day Social networks earned an estimated $8.3 billion from advertising in 2015 38% of organizations plan to spend more than 20% of their total advertising budgets on social media channels in 2015, up from 13% a year ago Source: https://www.brandwatch.com/2016/03/96-amazing-social-media-statistics-and-facts-for-2016/ 4
- 5. Digital Commerce/Internet Companies Search Based: Google, Bing Market Place: Amazon, Flipkart, Snapdeal Social: Facebook, Twitter, LinkedIn Life Events: Naukri, Jeevansathi, 99Acres, Shiksha Peer2Peer: Uber, Airbnb, LendingClub LifeStyle: Netflix, Zoomato, Promotions: Groupon, CouponDunia Many more…
- 6. Issues & Challenges Data Explosion Problem (Volume, Variety & Velocity) – Cloud-based data storage and low-cost, high-speed data processing has become increasingly cheap. – Naukri Example Volume: Millions of Profile, Searches, Jobs Views, Applies Variety: Structured, Semi-Structured, Unstructured Velocity: Daily Volume of Applies, Views, Searches etc… Scale & Performance – Millions of Requests – RealTime Recommendation Engines 5 Exabytes: All words ever spoken by human beings.
- 7. Data Science Applications in Digital Commerce Data Science in Digital Commerce Personalized Recommend ations Semantic Search Contextual Advertising SEO Customer Retention Dynamic Pricing 7 Customer Acquisition • Funnel Optimization • Lead Generation & Scoring • Telesales Prioritization • Cross Selling & Upselling Customer Service • Voice of Customer Analytics Diagram • Chat Bots Data Driven Insights • Price Trends • Similar Locality/Properties • Content Creation Fraud/Spam Detection
- 9. Recommendations – Why? How? – Why? Attempt to cross-sell or up-sell Provide customers with alternatives that might please them even more – Traditional approach No recommendations at all Products in the same category Manually managed cross-selling opportunities per product – Why are these approaches fundamentally flawed? They all start from the seller perspective, not the customer! “We know what you should be buying” Manual recommendations are too costly and time-consuming to maintain – even impossible with large catalogs
- 10. Recommendations – Online (RealTime) vs Offline Main focus on online, but why? Who knows best what products to recommend? Learn from your data, don’t take decisions based on a feeling. – Customer based recommendations Learn from your customers and their past. Understand Behavior, Segment Android vs iOS smartphones. – Time based recommendations Recommend or cross sell different products depending on – season? – holiday? – weather?
- 11. Recommendations – what does Amazon do? Cross-selling as realized with other (similar?) customers Starts from customer point of view! Recommendations based on perceived customer journeys Re-use the product comparisons that previous customers did! DATA DRIVEN!
- 12. Personalizations – Loyal (online) customer vs new customers. – Browsing habits and patterns. – Spending patterns. – Personalized discounts and/or content?
- 13. Personalizations Customer should be central – Provide a truly personalized shopping experience – Like high-end physical shops with personal approach to VIP customers Gather data about your customer – Surfing history – what products where looked at? How long? … – What products were bought? When? – Brand preference? – Product-segment preference? (budget, high-end, best-buy?) – Abandoned shopping carts Take action based on information mined from this data – Triggered e-mails, personal recommendations, …
- 15. Dynamic prices – End of life products? – Relevancy of products. – (Local) competition. – Customer!
- 16. Dynamic Prices – some ideas Auto-combination special offers based on cross-selling info Monitor stock & manage promotions accordingly – Example: stock of calendars in December (value decreases over time…) – Example: Customer history: needs incentive to buy? Why not give a small discount if bought together? Testing will show if and for which products and customers this increases revenue!
- 17. Dynamic Prices – some ideas Pricing vs competition scraping competition websites Analysis of tenders vs deals – What type of deals do we typically win, and which not? = Data mining on CRM data! – How can we optimize our chances to make a deal? Which tenders should we invest in? What offer should we make? Remark: in B2C scenarios, can be difficult / unwanted to use dynamic prices. Mind the legal impact!
- 18. SEO 18
- 19. How Search Engine Works 1. Gather Content – Crawler or spider moves recursively downloading content 2. Builds sophisticate index 3. Individual web searches run against index – Results are retrieved and ordered PageRank & Relevance
- 20. Search Engines
- 21. Google Search Placement Placement: importance and relevance PageRank (importance) – Counts links – Weights links Query matching (relevance) – sophisticated text-matching techniques – examines all aspects of the page's content (and the content of the pages linking to it)
- 22. SEO Optimization Categories Keywords – Keyword selection and keyword-rich text Crawler – A crawler-friendly site navigation scheme Links – Link popularity
- 23. Keyword Recommendations Meta tags: use but don’t stuff – <meta name="description" content="Free Web tutorials on HTML, CSS, XML, and XHTML"> Alt tags: use for graphics – <IMG src="star.gif" alt=“star logo"> Content is king – Write good content with relevant and important keywords in mind. Geo Targeting – Add geocentric terms to target local areas Domain Names – Use keywords as part of domain name
- 24. Crawler-friendly Engine spiders are primitive beings – choose simplicity over complexity Goal – All your web pages seen by crawlers – Google: enter in searchbox “allinurl:utexas.edu”
- 25. Link Development Inbound Links Impact PageRank PageRank (Popularity, importance) Number and quality of links pointing to a website Measure of usefulness of site Link Development Tradeoffs Advantage – it is dynamic, cumulative, and difficult to imitate Disadvantage – takes time (vs. advertising)
- 26. Link Development Approaches Quality content to start with Cultivate quality link (not quantity) Begin with web directories Harness online publicity Use Blogs and forums wisely Investigate competitors – Understand their strategy – Online publicity, blogs and forums – See inbound links ("link:domain.com" in Google, "linkdomain:domain.com" in Yahoo)
- 27. Data Science Work Overview @InfoEdge Naukri Real Time Recom. Engine Job Alert Mailers Recruiter’s Relevance Candidate Search to Recruiter Suggestor & SEO Semantic Personalized JobSearch Candidate Services • Lead-Scoring Engine • E-Learning • Recruiter Connection • Parser Naukri Gulf • Job Alerts • Real Time Recommendations • Recruiter Relevance JeevanSaathi • Profile Scoring basis Payment Propensity • Recommendation Engine • Spam Detection 99Acres • Price Trends • Similar Localities
- 28. Intelligence Powered by Analytics – Naukri Applications
- 29. Data Science Techniques Machine Learning Text Mining Natural Language Processing Semantic Technologies Information Retrieval Information Extraction BigData Technology like spark, mahoot No-SQL DBs like MongoDB Scripting Language- Python, Shell, R Lucene, Solr & Elastic Search
- 30. Credits to those authors who shared slides on internet and I own all the errors on this deck.