Suff Syed
Suff Syed
Nov 29, 2018 · 11 min read
Image for post
Image for post

For the best reading experience, I’d highly recommend downloading the PDF here.

Data Science is a New Frontier for Retail Innovation

As techniques in Machine Learning begin to prove themselves, we’re beginning to see every major technology player make key strategic bets in the field of Artificial Intelligence.

Google is uniquely positioning itself as an “AI-first company” — organizing every product team to embody a data-centric approach.

Facebook is training every single developer in the field of Machine Learning enabling adoption at scale. Along with setting a 10-year roadmap with AI at its core.

Amazon has been rolling out product after product with a heavy focus on Artificial Intelligence, Machine Learning, Deep Learning, and Augmented Reality.

Microsoft is focusing on building its Azure cloud services, which is central to leveraging and democratizing Artificial Intelligence at its core.

Apple is infusing their entire lineup of products with Machine Learning — along with launching a range of processors that can take advantage of ML.

The benefits of Machine Learning — how it fundamentally works, what makes it an important area to invest in, and how it’s delivered at scale to impact customer experiences — will have a profound impact on business strategy by becoming a key differentiator for brands.

Retail Executives are beginning to take notice, recognizing that Machine Learning has passed the hype curve and is beginning to become a critical area of strategic focus.

Here, you’ll discover a list of comprehensive use cases across the End-to-End Customer Lifecycle that will help the retail executive gain a deeper understanding of how techniques in Applied-Machine Learning and Data Science can help tackle ongoing business challenges.

There is a plethora of opportunities across the customer journey that can be elevated through the use of Data Science — directly impacting key areas of business strategy, customer experience, and business outcomes.

This list focuses on areas that show tremendous promise, along with areas that in general are new ways to look at existing business challenges. It is meant to be used as guardrails to think about better, more meaningful outcomes — instead of being perceived as a comprehensive list of opportunity areas that mandate focus and investment.

Artificial Intelligence, Machine Learning, Deep Learning and Data Science all bring with them a unique advantage of building a modern data-centric organization. One that leverages data at its core to build meaningful customer experiences that drive both business and human value.

Image for post
Image for post
Image for post
Image for post

Emerging Technology Platforms

Virtual assistants and chatbots are the most commonly understood use cases of AI. We’ve seen them in movies, and with the introduction of popular platforms like Amazon’s Alexa and Google Home these technologies are becoming a lot more accessible and familiar to the end user.

The most frictionless way for people to interact with computers is through natural language. This is why Natural Language Understanding is seen as one of the opportunity areas to elevate customer experience. Advances in algorithms focused on analyzing understanding and deriving meaning from human language are opening up new and profound applications for the end consumer.

As we begin to get better with Natural Language Understanding, so does our ability to understand Voice. Instead of forcing your customers to be heads down in a screen and constantly push them to enter their input in the form of keystrokes, we can use voice and allow users to speak out commands and opinions through voice.

One of the most powerful applications of modern AI is its ability to see — to visualize a user’s environment and identify the things they interact with. This opens up new avenues of customer interaction by understanding the things they’re holding, touching, feeling, and using. With visual recognition, the Retail industry is no longer confined to a physical store or a digital touchpoint — it can truly be anywhere.

With almost every big player heavily invested in AR (Google Tango, Facebook CameraEffects, Apple AR Kit), it is very well positioned to be the next platform for innovation. Snapchat and Pokemon Go have proven AR can work at scale and turned into something meaningful. If there has ever been a time to invest in AR it is right now.

Virtual Reality allows you to transport consumers into entirely new worlds that are limitless. The ability to build and curate environments, which you as a brand can dictate, is something that can fundamentally change how consumers think about a brand. While VR is achieving scale, it brings with it great promise and a whole new way to engage with the consumers.

Expert Systems are automated systems powered by Artificial Intelligence to solve one or more complex problems. With intelligence at their core, an expert system can leverage the latest advances in Applied Machine Learning techniques to find solutions to existing problems in whole new ways — through the power of new computing standards and capabilities.

Advances in Robotic technology powered by AI are finding whole new ways to enter the Retail landscape. We’re beginning to see new applications ranging from Inventory Management Bots (Amazon), to Retail Service (Lowe’s Bot), to Delivery Bots (Starship Delivery Bot) enter the Retail environment and truly change how we go about designing consumer experiences.

Data Science & Artificial Intelligence being infused into the E2E Retail Customer Experience Journey

Stage 1 — Awareness & Inspiration

Look-A-Like Audience Modeling — Develop a model that discovers, engages, and builds new target segments based on a significant overlap of characteristics and patterns by cross referencing existing customers.

Social Insights And ListeningManage social interactions around the globe by automatically analyzing and understanding audience sentiment, tone, nuances, and cultural attributes to inform business strategy.

Intelligent RecommendationsIncrease average order values by converting more anonymous visitors by recommending relevant products; tailor content and offers across channels based on visitor preferences.

Response ModelingMaximize overall performance of marketing campaigns and promotional events by finding a set of promising segments that are most likely to convert based on promotional information.

Visualize Audience RelationshipsBuild an audience dashboard that allows you to visualize and monitor customer needs based on trends and behavioral patterns — driving confidence in decision-making during execution.

Content Analysis And Management — Optimize content based on formats that resonate the most with audiences. Personalize messaging by combining content analysis during campaigns with developed inferences.

Stage 2 — Consideration & Evaluation

Recommendation SystemsBuild intelligent models to anticipate, predict, prescribe, rank, and refine tailored recommendations to users based on user tastes and preferences.

Customer Segmentation Modeling — Understand customers at a finer level by modeling key customer segments based on their digital footprints to engage with them through more meaningful experiences.

Product Attribute Graph — Rank products based on how they perform and resonate across key customer segments and use that ranking criteria to showcase the right products to the appropriate customers.

Category Management — Calculate the appropriate balance of products across varying physical shelf/window/aisle constraints in a way to maximize gross margins and substitute the right products.

Recommendation Systems — Build models that rank product and multi-user taste preferences across a range of critical parameters in order to return accurate results by considering context and need.

Needs-Based State Machines — Build predictive models that consider the different states a user may belong to by understanding their needs within each state and possible triggers that generate new needs.

Pricing And Profit Optimization — Model a pricing strategy by keeping in mind overall business goals and quarterly targets in mind. Assign individual price to each customer in order to maximize overall revenue.

Demand Modeling — Recognize key trends that are both internal (customer/business related) and external (market/ industry related) that may influence the increase/ decrease of demand for key products.

Predictive SearchEnable data-driven search recommendations that allow the user to perform rich search queries that hand hold them to the right products in the most customized way possible.

Catalog Ingestion — Automate SKU creation, curation, and management by recognizing patterns from available merchandising data, and use conflict resolution as a means to increase confidence in results.

Collaborative Filtering — Allow customers to use meaningful and relevant filtering criteria to arrive at products they’re looking for — learn from patterns and behaviors to inform better product placement.

Stage 3: Shopping Cart

Struggle Detection — Automatically detect struggle with little or no human intervention by analyzing critical user flows and tasks and learn through observations to increase user experience where required.

Merchandising Analytics — Recognize patterns in customer behavior across buying patterns to improve product associations and recommendations in the shopping cart.

Basket Analysis — Analyze how customers make purchases and the behaviors that inform the products they buy to create relevant product associations enabling a smarter purchase.

Fraud Detection — Take advantage of cutting edge security solutions in fraud detection that recognize behavioral signals of fraudulent transactions and employ counter measures to void them.

Dynamic Pricing / Price Discrimination — Make optimal pricing decisions, keeping customer and business value in mind, by setting optimal prices to increase revenues and engaging with price-sensitive customers.

Sales Value Decomposition — Identify how much proportion of sales is being driven by which driver factors that pertain to internal and external triggers. Use them to test hypothesis and validate through experiments.

Propensity Modeling — Conduct experiments and build models to predict the key shopping behaviors that determine likelihood of conversion versus likelihood of abandonment.

Stage 4 — First Purchase

Response Modeling — Understand how customers respond to using your product by employing a mindset of collecting
feedback and analyzing the responses to inform the design and experience of products.

Stage 5 — New Customer

Profile Insights — Build rich and informative customer profiles that allow a deeper understanding of customers, segmentation, preferences, and needs that inform product strategy and roadmaps.

Preference Modeling — Model multi-taste customer preferences based on collected data to understand what customers are looking for in products and services that can be used to inform strategic thinking.

Stage 6 — Second Purchase

Customer Churn Analysis — Build models to understand why customers might not return to the brand, and fill the gaps across the user journey in order to better engage with customers.

Propensity To Category Expansion — Increase likelihood of selling related or complementary products to customers by making data-driven recommendations that are highly personalized.

Discount Modeling — Build unique models of discounting and incentivizing customer segments to increase the likelihood of their next purchase by understanding motivations and triggers.

Call Back Analysis — Determine accurate, contextual, and highly personalized strategies to conduct reach-out programs with actionable triggers that can be measured.

Stage 7 — One and Done

Customer Lifetime Value Modeling — Combine unique targeted models in varying granularity across different categories to determine the measurement in dollars associated with customers to identify their lifetime value.

Propensity To Change Shopping Habits — Identify likelihood of life-changing events (e.g., relocation, wedding, etc.) triggering new patterns of customer behavior and how that might open up new categories to purchase from.

Share Of Wallet Modeling — Increase percentage of customer spending within categories that resonate with them by understanding the context in which they engage with the brand and make purchase decisions.

Risk Modeling — Build a process to identify customers early-on in the funnel who are most likely to disengage with the brand and most likely to not return to make another purchase.

Stage 8 — Repeat Customer

Uplift Modeling — Understand key patterns of repeat customers and how their behaviors may be different to regular shoppers in order to serve them differently, enabling a substantial sales boost.

Stage 9 — Loyal Customer

Advocate Definition — Building an acute understanding of what it means to be a loyal customer of the brand and what are the set of parameters that a customer needs to fulfill in order to sustain it.

Gift Card/Loyalty Card Analysis — Go from Personas to People with a highly personalized loyalty program that keeps individual preferences in mind and engages in a contextualized relationship.

Stage 10 — Lost Customer

Churn Modeling — Quickly identify customers with high propensity to churn, and target them with retention campaigns that may bring them back through a discount, gift, or other monetary benefit.

Foundational Platform Elements

Minimize distribution costs, warehouse assignment costs, route management, scheduling, product grouping, etc. by understanding the most optimal ways to manage your Supply Chain by leveraging Applied Machine Learning techniques.

Build accurate pricing and profit models based on a range of parameters and pricing elasticity models. Predict likelihood of customer spend in line with business goals and profit margins.

Automate order management, scheduling, delivery, and pick-up by understanding the crucial factors and risks involved to the overall business. Optimize models based on a dynamic data and cadence of smooth operations.

Predict optimal store or fulfillment center inventory based on historic and likelihood patterns of what customers may and may not order. Align with sales data to host the right products in the right geographies and store locations.

Predict likelihood of sales and conversion across different properties by evaluating trends and sales pattern data. Analyze how you can do better and constantly implement new strategies to fuel new ways of selling.

Detect which product combinations make sense for each customer segments and see how you can take advantage of a deeper understanding of the relationship-mapping of products to consumers.

Predict abnormal behavior early-on in systems that may potentially lead to downtimes/crashes. Understand historical patterns/triggers to actively manage peak system performance — allowing you to constantly maintain efficient systems.

Automate the detection of anomalies within systems for applications like Fraud Prevention, Fake Reviews, Coupon Transactions, etc. Build a deeper understanding of anomalies and employ counter measures to actively tackle them.

Calculate optimal warehouse assignment routes by understanding unique parameters that influence the matching of inventory to warehouses to either store or customer delivery.

Build a Demand Prediction model that analyzes product properties, sales events, shopping trends, and inventory status to model accurate demand and distribution.

Map the shortest, most optimal routes across various handoff stages for all logistical operations across the end-to-end customer experience in order to save costs or prioritize investment in crucial areas that have a direct business impact.

Automate sales event planning for flash sales, seasonal sales, clearance sales, coupons, etc. by optimizing revenue management systems that predict and model sales events based on customer data and propensity to buy.

Build a cognitive tagging mechanism for individual products across all consumer segments that defines a multi-dimensional attribute model to build product intelligence, which can be leveraged across multiple use cases and applications.

Democratize data across the organization by building a robust, scalable data platform allowing seamless access for key departments to execute data and build a data-centric strategy and execution plan.

Data Driven Investor

from confusion to clarity not insanity

Sign up for DDIntel

By Data Driven Investor

In each issue we share the best stories from the Data-Driven Investor's expert community. Take a look

By signing up, you will create a Medium account if you don’t already have one. Review our Privacy Policy for more information about our privacy practices.

Check your inbox
Medium sent you an email at to complete your subscription.

Suff Syed

Written by

Suff Syed

Maker. Polymath. Creator.

Data Driven Investor

from confusion to clarity not insanity

Suff Syed

Written by

Suff Syed

Maker. Polymath. Creator.

Data Driven Investor

from confusion to clarity not insanity

Medium is an open platform where 170 million readers come to find insightful and dynamic thinking. Here, expert and undiscovered voices alike dive into the heart of any topic and bring new ideas to the surface. Learn more

Follow the writers, publications, and topics that matter to you, and you’ll see them on your homepage and in your inbox. Explore

If you have a story to tell, knowledge to share, or a perspective to offer — welcome home. It’s easy and free to post your thinking on any topic. Write on Medium

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store