Smart Segmentation: Maximizing Engagement with AI

Smart Segmentation: Maximizing Engagement with AI

The Double Whammy Staring at Marketers

There was a time when only brands with massive financial clout could afford to run major marketing campaigns, simply because Above the Line (ATL) marketing is resource intensive. All of that changed with the digital explosion; now your neighbourhood mom and pop store could launch a Facebook or Adword campaign for as low as $50.

This resulted in an exponential rise in competition in the digital space for mindshare and customer loyalty. Unfortunately, the digital influx also pummelled consumer attention span to a meagre 8 seconds (a sharp decline from 12 seconds in 2000). As a result, today’s marketers find themselves in an intensely competitive universe where thousands of brands are battling it out for the ever-shrinking attention span of the consumer.

The Rise of Digital Marketing

In the new digital universe, consumer attention and loyalty soon became an invaluable and coveted resource. And brands started pouring millions of marketing dollars to capture consumer mindshare. For a while, this worked, but the human psyche is incredibly good at filtering out irrelevant information and there started a growing distaste for ads. During the period from 2015 to 2019, the number of Americans using adblockers grew from 15% to 26%.  

Marketing 2: 0: Engaging the Connected & Empowered Consumer

Smart marketers soon realized that blasting consumers with a barrage of generic, irrelevant communication is hurting them twofold: it quickly drains precious marketing budget and also spreads the sentiment that the brand is apathetic to consumer needs. To gain mindshare in an attention-deficit consumer, marketers needed to make personalized communications

A key pillar within this reformed marketing philosophy is personalization because it sets the stage and gives a contextual framework for a relevant and value-driven engagement.

The Holy Grail for Marketers: Hyperpersonalization at Scale

In its simplest sense, personalization is a targeted approach to customer engagement that delivers tailored, useful and relevant communication to every customer.  For smaller businesses with small to medium customer sets, the personalization efforts are fairly simple and achievable with a good analyst team. But it gets increasingly complex and a herculean task at enterprise levels of data and consumer sets. In fact, less than 10% of top retailers say that are good at effective personalization.

When done well, personalization accelerates sales and business growth for brands while increasing overall customer satisfaction levels. Personalization improves conversion rates by a whopping 70% and it impacts ripples across the entire consumer lifecycle: from acquisition costs, engagement levels, average order values to repeat purchases.

The challenge to implementing personalization at scale basically boils down to two things: the pace of change in consumer behaviour and exponential rise in data volumes. Combine these two and you get rapidly changing data, constantly shifting customer segments and frequent changes in the type of insights needed from the data.

Smart Segmentation with AI: Powering the New Personalization Paradigm

Since smart segmentation serves as the primer for personalization, it’s vital to create relevant and accurate segments before embarking on your engagement strategy.

  • The Need for Smart Segmentation

While basic segmentations like demographics and location can be done manually, creating complex personas and purchase patterns segments can be tedious and time-consuming. Moreover, these generic segmentations cannot be used for building advanced purchase propensities modelling.

For successful smart segmentation, brands will need to layer basic segmentation attributes with ‘behavioural aspects’ of a customer. These include explicit behavioural aspects like purchase history, search history etc. and implicit behaviour aspects like dwell time on a specific product page, heatmaps and other storefront interactions. Moreover, these behaviour segments need to be dynamic and adaptive based on evolving customer needs.

This is where emerging technologies like AI and ML can help brands optimize their data and resources to achieve their business goals.

  • AI-Powered Adaptive Segmentation

The true power of Artificial Intelligence lies in its ability to find complex and disparate correlations which are almost impossible to uncover through manual intervention.

The need of the hour for brands is to value-driven experiences at every stage in the customer journey.  However, multiple data sources and non-linear customer journeys have made it difficult for brands to create a seamless customer experience. Adaptive segmentation is a great way to create centralized segmentation based on multiple factors like demographic data, behaviour metrics and time.

Adaptive segmentation is powered by an intelligent algorithm that constantly ‘learns’ more about the customer every time he/she engages with the brand – whether on social media, email, website or in store. Adaptive segmentation also enables brands to optimize segments based on higher conversion probability rather than simple demographics metrics.

4 Step Process to Smart Segmentation

AI-based segmentation typically comprises of 4 stages :

  • Pre-processing: The initial validation and cleaning of data sets. A ‘gold standard’ training set will need to be identified at this stage which will serve as a primer for future use cases.
  • Modelling: This is where you identify the variables that make up your segmentation. These variables are then stacked based on the order of importance and applied to the ‘gold standard’ training set to help the machine understand the common properties for your segment. The popular segmentation attributes are :
        • Traffic source (website, PPC, email campaigns, social media etc)
        • New or returning customer
        • Platform (mobile, desktop)
        • Location
        • Average Order Value
        • Demographic, Likes and preferences
        • Search Behaviour
        • Past content interactions (product page, blog post)
  • Evaluation: At this stage, a confusion matrix is used to identify previously incorrect contacts and also evaluate the accuracy of the model. If there are unbalanced data sets across the segments, a statistical coefficient is applied to account for class imbalance.
  • Transformation: Output data is achieved and your customers are now smartly segmented in accordance with the ‘gold standard’ training set.

Benefits of AI-Powered Segmentation

  • Eliminate human bias and stereotyping while segmenting
  • Create an almost infinite number of segments and sub-segments
  • Real-time updation of segments based on current behaviours and market trends
  • Uncover hidden patterns and trends (that goes against prevailing assumptions)
  • Highly scalability
  • Higher engagement rates and ROI

Despite the overwhelming success rate of AI-powered segmentation in driving personalization, its adoption rate has been low with 55 percent of marketers admitting to not implementing it due to siloed data sources and in some cases, lack of sufficient data itself.  To stay ahead of the digital revolution and be consumer-ready, it’s imperative for brands to create centralized, data pools using omnichannel analytics platforms and deliver a more personalized customer experience using AI-powered marketing solutions.

eCommerce in the Era of Artificial Intelligence

eCommerce in the Era of Artificial Intelligence

Sci-fi geeks previously mocked for their far-fetched ideas about robots taking over the world can today smugly say, ‘hah! Told you so. We are no longer at the cusp of an era of artificial intelligence; we are living in it. Our daily lives are peppered with gadgets that use voice recognition, search predictions, and facial recognition. From systems that forecast the weather to those that predict stock prices, to self-driving cars—we are breaking new frontiers so rapidly that we rarely stop to think, about the constant evolution of these machines as we use them.

Artificial intelligence, simply put, is the ability for a machine to learn. It does this by continually analyzing patterns from the data it collects. The greater the collection of data, the more accurate the machine can become at making predictions.  Gartner, Inc, projects that over 25% of customer interactions will be managed with virtual customer assistants by 2020. Machine-learning technologies can go beyond the scope of humans by learning and understanding every individual customer, even in an audience of millions.

The evolution of AI

Artificial Intelligence truly began in the unleashed imagination of fiction writers. Although the rich pages of Isaac Asimov may come to mind, the technology started impacting the real-world only in the 1950s when Alan Turing contemplated the question, “Can machines think?” The Turing test was once considered to be the benchmark for Artificial Intelligence—to pass the test, a machine must deceive a questioner into believing it is a human.

In recent decades, AI research has seen unprecedented acceleration. From the year 2008 to 2012, it was growing at roughly 5 percent annually; but since then, it has boomed with a growth of more than 12 percent annually. Today, Europe still leads in AI research, but within the next four years, China is predicted to take over as the global pioneer in artificial intelligence.

A New Era

At the same time, trends have been changing in the world of business and retail. In 2018, e-commerce sales accounted for a whopping 11.9% of all retail sales worldwide. The number of global digital buyers is expected to rise to over 2.14 billion by 2021.

Since every marketing strategy boils down to understanding what consumers want, this puts AI at the forefront of the new era of commerce. Artificial Intelligence collects and organizes heaps of data that manual processing would take years to do (if at all). Based on this analysed data, it can make predictions that can save an organization vast amounts of time and money.

According to Juniper Research, AI chatbots could save businesses up to £6 billion per year across industries. But more importantly, it can give an organization greater precision while marketing, by knowing their customers better.

Artificial Intelligence in eCommerce

Here are some key ways that AI has made an entrance into the realm of e-commerce:

1)    Chatbots

It might soon be the end of the line for contact forms, email and phone calls. More importantly, it’s a step in the right direction when it comes to delivering stellar customer experience. According to a report, 51% of customers never come back to a business after a bad experience. By kickstarting the ‘conversational commerce’ trend, Chatbots give e-commerce websites the ability to provide 24/7 customer support. They simulate conversations with customers and can execute tasks, automate order processing, and can also provide accurate answers to customers about product details, quantities and shipping terms.

According to a 2017 report by PwC, 34% of executives say the time they freed up using chatbots allowed them to focus on deep thinking and creating.

Chatbots increase response time and answer more than 80% of regular questions; freeing agents to carry out other tasks. Brands like Lyft, Alibaba, and Spotify are using chatbots to carry out various operations, such as booking a cab, finding the right music, and placing orders.

2) Image search

With AI, consumers can now search for products based on images they’ve come across. See an outfit you love? Simply take a picture and get matched to similar items on ecommerce websites.

Pinterest is already leveraging this technology by allowing its users to select any item from any photograph online and then throws up similar items through an image recognition software.

3) Cybersecurity

Since ecommerce sites process multiple financial transactions throughout the day, they are frequent targets for cyber attacks, fraud, and unauthorized access from hackers. Cybersecurity is a priority among top ecommerce stores that realise that cyber attacks can cost millions, and worse, damage their reputation as a trustworthy brand.

AI systems can detect irregular patterns, like spam and fraud, because of constantly processing data, and sound the alarm in real time when it detects suspicious activity. For example, the cybersecurity company Prolexic monitors malicious cyber threats globally and analyzes DDoS attacks using their proprietary intelligent system.

4) CRM

Customer Relationship Management (CRM) have traditionally relied on employees to collect huge amounts of data in order to better serve clients. Today, several CRM software are capable of leveraging artificial intelligence to help in identifying trends, and predict which customers are most likely to buy products with impressive accuracy.  This frees up sales teams to focus on their revenue goals.

5) Recommendation Engines

There are those who are familiar with recommendation engines, and then there are those who are shocked when their phone throws up ads for exactly what they needed. Is it a conspiracy?

eCommerce platforms powered by Machine learning algorithms are capable of analyzing customer behaviour from past searches, ‘like’ history, frequently bought products, and can recommend products for the user. Amazon, Facebook and Instagram are examples everyone is familiar with, frequently giving us recommended products and ads based on our history.

AI has the potential to spike sales and productivity like never before. Today, more than 75 percent of companies are using AI to power their decision-making processes and accelerate business development (as reported by Capgemini, 2017).

Perhaps the world is changing at a pace where we need to hold onto our seats, but Artificial Intelligence is more than a buzzword. It is a field of research that is unlikely to disappear anytime soon and businesses will do well to jump on the bandwagon at the earliest.

Inventory Management Basics for Marketplace Sellers

Inventory Management Basics for Marketplace Sellers

Being an ecommerce marketplace seller is not an easy job. You are sandwiched between marketplace regulations and ever-increasing customer expectations for superior buying and fulfilment experience. Thankfully, many of the roadblocks and issues plaguing ecommerce marketplace inventory management can be easily resolved by establishing a solid inventory strategy from day one.

Understanding Marketplace Inventory Management

Inventory Management is the lifeline for marketplace sellers since the success of the business hinges on delivering the right order to the right customer at the right time.

Here are some of the basics of inventory management you should be aware of before starting your journey as an ecommerce marketplace seller :

  • Organize Your SKUs – Stock Keeping Units (SKUs) are essentially unique identifiers assigned to each of your product to streamline the ordering and logistics process. SKUs can be customized as per your wish but they are typically categorized based on product, category, popularity etc. The first step in establishing an effective inventory management system is to have a simple and well-structured SKU system in place.
  • Focus on Product Types – In order to ensure a seamless workflow and minimize time wastage, you should have a thorough knowledge about the product type you are selling on marketplaces. Here are the major product types that ecommerce marketplace sellers usually deal with :

– Item : An individual product or an item that doesn’t require any special storage or shipping considerations

– Case pack : A group of usually similar items clubbed together. They might require special attention during storage

– Assembly : These are items that require assembling multiple parts scattered across your warehouse before shipping. Based on your sales forecast, you will need to organize them based on individual parts or end products

– Family : Similar products that have variations on colour and size. These will also need to be grouped based on your sales and demand forecast

  • Get Serious About Demand Forecasting – Use historical data and other demand forecasting techniques to prevent negative sales impact like stock outs. Also, pay close attention to your sales analytics to understand the rise and fall in demand for your product. Make sure to take into account seasonal variations and other customer behaviours that will impact the sales of your product.
  • Have a Backup Plan – Even the most experienced ecommerce sellers will not be able to predict product demand 100% accurately every time. At some point or the other, a random occurrence or event will disrupt even the most the most watertight strategy. Smart sellers should thus always have a backup plan in terms of surplus inventory of bestselling products or a reserve supplier.
  • Ensure Data Accuracy – Ultimately, inventory management is a case of numbers and data and even minute discrepancies can have a severe impact on your business. Some of the common but easily avoidable mistakes include adding multiple SKUs into a single slot andcounting case packs as a single item instead of the individual pieces

Get these inventory management basics right and you will set forth your ecommerce marketplace business on a solid footing.

Capillary has partnered with Amazon to help marketplace sellers to maximize profits and revenue from their ecommerce business through easy and efficient inventory management. You can benefit from this partnership with a free catalogue service for 1 month to launch your business on Amazon. Now reach crores of customers. Sign-up here.

#Ready Insights: Get to Know the Connected Customer

#Ready Insights: Get to Know the Connected Customer

Exponential advances in technology, the proliferation of digital devices and constant connectivity has given rise to a new legion of consumers who live in the Easyverse™. These Connected Customers expect fast, easy and personalized experiences anywhere, anytime.

The imperative for brands to understand the ‘Connected Customer’

APAC’s digital economy is expected to top $1.2 trillion by 2021. To cash in on this multi-billion dollar opportunity, brands will need to get a deep understanding of the Connected Customer.

Furthermore, the term ‘experience’ has become more important than ever before for brands. In an era of minimal product differentiation and diminishing brand loyalty, customer experience and service will spell the difference between growth and a painful downward spiral into irrelevance.

But, do retailers and brands have what it takes to meet the expectations of the ‘Connected Customer’?

#Ready

To answer this, we invited business leaders and brand experts Maya Hari (VP, Twitter, APAC) and Johan Vracken (MD, Nielsen, Singapore) to share exclusive insights about priorities, purchase preferences, and expectations of the new age, digital-savvy customer.

Understanding the Connected Customer

Johan Vracken

According to Nielsen, ecommerce in APAC is growing at a rapid pace with more than 98% of consumers making an online purchase against a global average of 95%.

Here are some more key takeaways and insights on the Connected Customers in APAC according to Johan Vracken (MD, Nielsen, Singapore) :

  • On average they spend 6.5 hours online every day and expect brands to create a delightful experience at every turn
  • SEA has 350 million online customers
  • Online purchase of beauty & personal, packaged grocery, pet food is expected to surge in the next few years
  • Connected consumers invest in experiences, spend heavily on travel, events, gaming, and consumer electronics
  • Consumers in APAC choose brands based on loyalty programs, subscriptions, fast website and multiple delivery options
  • APAC customers have an online purchase evolution that starts from small value products like skin care and eventually moves to premium and bulkier products

Within the Connected Customer universe, Millenials constitute a vital demography due to their numbers (50 per cent of the APAC population will fall in this age bracket by 2020) and their impact on global economy. However, engaging this consumer  segment has been a marketing nightmare for brands across the globe.

Maya Hari

Maya Hari (VP, Twitter, APAC) shared her insights on how brands can understand and connect with this mysterious demography in a better way

  • Patience is in short supply with millennials – the average attention span is around 8 seconds
  • 65% of millenials look for deals only at last minute
  • One size fits all approach won’t work with the millenials – be it reward programs or brand communications
  • Asian millennials are tech savvy, driven by purpose and obsessed with travel and fitness
  • Compared to their western counterparts, Asian millenials focus more on quality rather than fancy packages
  • 57% of millenials expect brands to stand for a social issue and they like brands that associate with movement
  • Asian millennials are no longer in a position to afford housing easily, so 63% are not keen on leaving families

To attract, engage and retain the connected consumer, businesses need to look beyond conventional approaches and rely on digital solutions and cutting-edge technology to address changing consumer needs effectively. However, adopting technology alone is insufficient, the most significant change needed is in the mindset of business leaders.

To learn how to lead a digital-first organization, read the next part in our #Ready Insights series – The New Gen Leader

The Future of Ecommerce & How to Be Ready For It

The Future of Ecommerce & How to Be Ready For It

For most of human history, people have been good at predicting future technologies and innovations for say, 100 years ahead.

Today, predicting things even just 5 years down the line is usually an attempt at futility.

Thousands of existing technologies and industries are blending and merging in millions of ways in a singularity that looks all set to explode into a big bang of endless possibilities. However, even amidst rapidly changing variables, there are constants that have stood the test of time.

In a simple sense, technological advances merely served as a substrate for achieving what we humans loved so dearly from times of yore: expend minimal energy and save time. In fact, if you trace any invention or innovation down the rabbit hole to its moment of inception, it most likely would have been sparked by a single factor: convenience.

If you’re a retailer who is caught between agile, well-funded competitors and rapidly changing consumer behaviour, fret not as this single entity might be the perfect leading light beacon for a safe and successful voyage into the future.

So, without further ado, here are the top trends that will rule the future of ecommerce.

The blurring of digital and the real world

AR

The question of whether we are living in a computer simulation is a long-standing one amongst astrophysicists and philosophers. Well, the question will seem more pertinent in the future thanks to AR companies like Next/Now, INDE and Groove Jones which are working feverishly to bring to life screenless digital projections and holograms on to the real world; rendering reality almost indistinguishable from the digital.

These technologies will have a massive impact and use-cases in ecommerce and retail landscape by easing purchase frictions by empowering consumers to negate the guess-work. A case in point is the ability to project life-like furniture in your living room to check if it matches the decor.

Intelligent digital assistants will become widespread

Pepper Robot

While the current set of AI assistants like Siri, Alexa and Cortana are impressive, the next generation of digital assistants is expected to be capable of almost human-like cognition. We are talking ‘HAL 9000’ from 2001 Space Odyssey and Samantha from ‘Her’ levels of cognitive and reasoning capabilities.  Will they become the conduit for AI to go into overlord mode and enslave humanity? Elon Musk seems to think so but we’re not very sure. But they definitely will have huge ramifications in the ecommerce and retail space.

Macy ’s and a few other brands have already started piloting AI-based assistants in its stores. While its current capabilities are limited to guiding customers to the right section and offering product information, the day won’t be far when these assistants are capable of offering personalized fashion tips based on the occasion, local weather conditions, facial and body contour, and latest trends.

The shift from 2D view to 3D view of customer

3D View of Customoer

As consumers start engaging with newer technologies like wearables, connected cars and IoT-enabled appliances, the amount of customer data available to brands will skyrocket in the next 10 years. This will accelerate the movement from the current 2D view of the customer to a 360-degree accurate view. When coupled with AI and ML innovations, these datasets will be able to predict consumer behaviour to almost 100% accuracy.  The winners will leverage this data pool to build a persona for engaging with their customers in a contextually relevant and personalized way at every step in the shopping journey.

The pretzel-shaped shopping journey

Pretzel Customer Journey

The traditional linear purchase journey is already on its leg and in the next 5 years, brands can expect it to be totally phased out. Thanks to the proliferation of wearables and connected devices like cars, appliances and smart TVs , the consumers of the future will engage with brands across multiple platforms and devices in a pretzel-like twisted, crisscrossing loop. For retailers, the key here is to maintain a connected, unbroken, omnichannel journey and be ready for channel disconnects and reconnects anywhere, anytime.

Predictive analytics and curated shopping experiences

Predictive Shopping

The barrage of data and the rise of complex AI systems will allow brands to develop accurate predictive models for recommendations and R&D for new development.  This will give create the next step in curated shopping and subscription services to create personalized experiences to fit the exact needs of an individual customer. Predictive analytics is expected to have widespread retail use-cases like loyalty programs, customer engagement, in-store experience and fulfilment. For example, retail brands can devise a predictive model that will generate a customized discount code for customers who are at the risk of switching to a competitor.

Fulfilment will become a major differentiator

With the rising demand for same-day delivery, outdated supply-chain processes will no longer viable. The on-demand economy will force technology giants, logistic firms and retailers to combine forces to devise innovative delivery modes. Expect drones, driverless/autonomous vehicles, wearable and mobile technology and robots to be the most disruptive technologies. Alibaba, Amazon, UPS and Walmart are already experimenting with fully automated warehouses manned by an army of robots and drone delivery.  In fact, drone traffic has reached significant levels in Japan that Rakuten Air Map has launched an unmanned traffic management platform in the country. Brands are also likely to engage futuristic transportation modes like hyperloops and autonomous travel pods to further optimize their logistics network.

Hyperpersonalization will be the new norm

We are already witnessing personalization in many forms, but it doesn’t hold a candle to what is about to come. Consumers can expect a seamless, dynamic and real-time shifting world of content created around them by intelligent algorithms that know them better than they know themselves. Shoppers will have instant access to delivery dates, lowest shipping cost, and the best discount codes for multiple retailers within seconds. An agile ecommerce platform powered by AI and ML technologies will be critical in determining the success of your hypersonalization efforts. The winner may come down to brands that know the consumer the best, their product and delivery preferences, the likelihood of return shipments, and the packaging itself.

Omnichannel will evolve into the Omniverse

Omniverse

Omnichannel commerce is expected to become even more connected and seamless in a way that the individual blocks merge and finally fade into each other to create a unified real-time universe. This concurrent experience will involve consumers engaging with brands across digital and offline channels all at the same time. For instance, a customer walking into a retail store will be simultaneously comparing prices through his wearable smartglass.