GBVIT_Discount Model In Ecommerce

Submitted By:

Anant Katyayni (G039)Krishna Kashid (H025)Rishabh Rastogi (H047)Kanwar Pal Singh (H059)

A Tale of Discounted Price Wars in E-Commerce

The deep-discount model followed by e-tailers is UNSUSTAINABLE

Are Deep-Discount models Sustainable?

http://articles.economictimes.indiatimes.com/2015-07-13/news/64370623_1_kumar-mangalam-birla-coal-payment-bank

E-commerce business model only works if you have growth. Month-on-month growth has now stopped. The day you reduce discounts, your sales drop by 40-50 per cent. This model has to survive on its own in some form. It will be reinvented. It will emerge and find its own space.

Predatory Pricing

The Dangers of Discounting

Effect of Discounts on Sales Target How much more you need to sell?

Discount

Source: Business Victoria, State Government of Victoria, Australia

Addressing real problem: Why don’t people buy?

Source: http://www.slantmarketing.com

1. Free Shipping: 61% consumers want free shipping

2. Special offers : e.g. Free Gift wrapping3. Bundled Pricing: Show customer the benefit of

buying complimentary products/services (Upselling) e.g.

[Shampoo + Conditioner] [Manicure + Pedicure]

4. Add VALUE to the purchases: Computer hardware supplier - free an

installation support. Hairdresser – a free facial massage or

blow wave with haircut.5. Quantity discounts / Bulk Discounts:

Offer a % discount when customers buy in bulk quantity Increases the SIZE and VALUE of customer orders . e.g. Buy 5, get-one-free

Alternatives to Discounting

71%Women search apparel online, but buy offline

Major Reason“Fitting Issues”

Current Scenario Business Insights20%

Average Return Rate in Apparels

Choose Body Type Hair wear Type Choose Model

How to ADD VALUE?

18Different

angles

Recommendation 1: Add VALUE by Differentiation

Source: IndianExpress.comeconsultancy.com

3D Trial

Room

Recommendation 2: Dynamic Pricing (Analytics)

Big Data competitor pricingmarket trendssales volume etc

Improves profit margins by 25% on average

Analytics and Insights

Analytics

Neural Network

Decision Tree

NearestNeighbour

Cluster Analysis

Decision Tree

• Strategic Data based decisions - Gerber Products Inc.

• Used to predict CLV• Value based

segmentation• Customer

Classification

Nearest Neighbour

• Identify violations in business transaction

• Unconventional or sporadic behaviour of consumer

• To study the impact of campaigns on top line

Cluster Analysis

• Grouping of certain type of customers for consumer insights

• Behaviour of teenager male and female can be studied

Neural Network

• Human brain neurons• Learning algorithm

used to increase accuracy over the time

• Robust system be built over a time

Exposure funnel: Suspects -> ProspectsAdoption funnel: Prospects -> CustomersRetention funnel: Customers -> Advocates

Fixed Segmentation vs Dynamic Segmentation

• Broad division based on customers needs/wants

• Limited number of concurrently active offers/messages

• Over reliance on discount marketing• Largely single channel• Low investment on technology

Fixed Segmentati

on

• Supporting next best offer for each customer

• Unconstrained set of offers/inspirational content

• Intelligent, highly targeted multi-channel communications

• Marketer friendly point and click interface

Dynamic Segmentati

on

Average order size Revenue per customer Customer profitability Cross brans customer

synergies Customer Tenure

Time to market new programs or offers

Sample Twitter Sentiment AnalysisVisualisation1

Visualisation2

Visualisation3

Visualisation4

For sentiment analysis, we have used text mining and statistical models to gauge twitter sentiment using R-programming for Sunslik. The R file of code is attached below.

Sample Twitter sentiment analysis:

Visualisation1 – Quantitative graph to measure polarity of the sentiment. For e.g.. +3 denotes strong positive sentiment, while -3 denotes strong negative sentiment and 0 signifies neutral commentVisualisation2 – This visualisation categorizes tweets on the basis of emotions such as joy, anger, sadness etc.Visualisation3 – Qualitative depiction of all the tweets in the form of positive, negative and neutral commentsVisualisation4 – WordCloud plot shows recent words used by people mentioning #Sunsilk

Twitter.R

Twitter Sentiment Analysis

Big Idea: Agent on hire (Portal & Mobile Application)

Hiring: Agents will be hired in collaboration with NGOs and Students, petrol pumps and kirana stores to act as delivery boys

Registration: A portal/app to register local and-hoc agents who can act as ideal manpower to furnish orders from e- commerce companies

Verification: Due background check to be done before getting the agents on board. If case of any misconduct, the source network of the agent will be blacklisted

Delivery Broadcast: Orders for delivery will be published on App which will be broadcasted to all agents; Agents confirming first would be responsible for delivering

Package Collection: Agent would collect the package from sub warehouse/locker located at a maximum distance of 50kms from city and deliver it

Cash Deposit: In case of CoD orders, the cash has to be deposited in the nationalised bank in the city within 1 day

Payment: Agent will be paid per delivery (6-8 packages will constitute one delivery) and commission on high value items

How It Works

Our Offering

Creating crowd sourced aggregator model for last mile deliveryOn demand pickup and delivery of packages by ad hoc agents

Value proposition

Facilitating CODIncrease in area under coverageEfficient DeliveryRecognition for partners

Target market

Unpenetrated areas of tier II and new tier III cities where maintaining a last mile delivery person is not feasible High mobile penetration

Marketing channelTraditional print media to execute a pull strategyArea Manager responsible for bringing business and people on board

Go To Market Strategy GAP Addressed

Optimizing the cost of last mile delivery by saving up on the cost of maintenance of delivery personnel in locations with lesser order traffic, by outsourcing the job to ad-hoc vendors

Game-Plan We will start with 20 tier 3 cities all over India

for Pilot run with following prerequisites: Sub Warehouse/ Locker in 50km radius

with good public transport connectivity Decent mobile/internet penetration Nationalized Bank availability

Using the local print media to publicize the launch and inviting applications

Creating localized campaigns focussing on engaging vendors and their people.

Launching the app in regional languages for better reach

Promoting the positive goodwill generated for both the parties after this association

In future Tie ups with local transport provider and cab aggregators

- Agent equipped with smart phone- Can accept / reject the order

based on the convenience- Advance Analytics to ensure the

agents are belonging in vicinity of destination addresses

- Real time GPS enabled tracking of the shipment

- Geography wise scheduling of deliveries

- Efficient planning of deliveries- Security of the agent will be

ensured by giving only one high value item at a tie

Predictive Shipping: They know it even before you think itofficially known as- method and system for anticipatory package shippingPatented by Amazon in Dec 2013

a method for shipping a package of one or more items “to the destination geographical area without completely specifying the delivery address at time of shipment,”

Predicting customers’ orders could Increase Sales and potentially Reduce Costs in: Shipping Inventory and Supply Chain

Forecasting model uses data: time on site duration of views links clicked/ hovered over shopping cart activity wish lists

Additionally the algorithm also sprinkles in real-world information: customer telephone inquiries responses to marketing

materials

Advantages in existing expedited delivery model

Recommendation: Ask Why, not How much?