According to Struto, it costs six to seven times more to acquire a new customer than it does to retain an existing customer. With growing awareness of the value of customer retention, more companies are shifting their focus from new customer acquisition to strategies that foster customer loyalty and retention. One of the most effective ways to boost customer retention and increase loyalty is to create a positive, engaging customer experience, and that’s where predictive analytics comes in.
Companies use predictive analytics for detecting fraud, reducing risk, optimizing marketing campaigns and improving operational processes, such as managing inventory and setting prices. It enables companies to gain a competitive advantage, discover new product or service opportunities and gain a deeper understanding of their customers. By leveraging predictive analytics, companies can predict problems before they occur and take action to meet customer expectations to reduce churn.
CallMiner Eureka leverages artificial intelligence (AI) and machine learning (ML) to analyze all customer interactions across channels, including calls, emails, chat, texts, surveys and social media, to uncover actionable intelligence across the entire customer journey. Through automated scoring (performance and sentiment), rich categorization and other advanced analytics features, CallMiner Eureka enables contact centers to uncover the root causes of inefficient call handling, improve first call resolution, measure the effectiveness of sales pitches, correlate behaviors on contact, and more. Converting customer contacts into a format for analysis, CallMiner’s customer journey analytics solution provides a wealth of data for gaining predictive insights.
Download our white paper, How AI Improves the Customer Experience, to learn more about how companies are complementing their existing contact center technologies with AI, automation, and predictive analytics to improve the customer experience.
To learn more about the innovative ways that companies are leveraging predictive analytics to improve the customer experience today, we reached out to a panel of marketers, analytics pros and customer experience experts and asked them to answer this question:
“What’s the most creative use of predictive analytics that you’ve seen a company use to improve customer experience?”
Meet Our Panel of Marketers, Analytics Pros & Customer Experience Experts:
Read more about the most creative uses of predictive analytics companies are using to improve the customer experience.
Erico Franco is an Inbound Marketing Manager at Agencia de Marketing Digital.
“We have been using predictive analytics to score our leads who contact our sales department…”
Our salespeople always know which prospect to pay the most attention to, based on each prospect’s score. We set up the scoring system based on a detailed analysis of the historical behavior of our prospects by CRM and the probability of them closing a deal with our company based not only on contact data, such as company size, company sector, etc., but also based on the behavior of this contact with our company on touch points such as:
- Contacts who fill out a second contact form telling us more of their goals and challenges
- Contacts who are engaged in receiving a call from our team after knowing the price of the service
Predictive lead scores have proven to be very effective in determining which prospect our sales department should attend to, improving customer experience and sales at the same time.
Maksym Babych is the CEO at SpdLoad.
“The company used hyper-personalized memes in email marketing activities to keep customers who are going to stop the usage of a product…”
It looks funny, but in terms of marketing, there hides a deep analysis and creative approach:
- The company does a deep analysis of ICP and current customer portraits to identify the points when users stop the use of a product.
- Then, the team builds up a pipeline of emails for a cohort of users who show the signs of churn.
- These users were engaged in the email campaign to boost retention rate.
- The usage of memes in emails helps to keep the open rate at a high level. It allows delivering marketing messages in a simple, attractive, and viral way. And it really looks positive.
Thus, memes and a predictive analysis helped maintain the high engagement of users.
Mike Falahee is the owner of Marygrove Awning Co.
“I’ve seen a few companies use search frequency and engagement time on their websites and specific pages to recommend specific items for them…”
For example, if a user searched for a product and products related to it and stayed on those pages for a while – as it is assumed they are reading the product description or contemplating a purchase – they are recommended more products like it so they can consider if the one they are looking at is really what they want.
These recommendations help to create a personal experience for that customer. This approach can also be used to track the amount of staff that need to be available for calls, as high-volume traffic on the website usually means more calls.
Amrita Singh Jaswal
Amrita Singh Jaswal is a Digital Marketing Analyst at Signity Software Solutions.
“Sometimes, it’s almost impossible to manage the baffling amount of data…”
For this reason, businesses are not able to optimize and use customer data efficiently, and customer experience turns into torture for customers. For example, you can say a customer will definitely get irritated when you send them irrelevant suggestions, messages, notifications, etc. And, in some cases, businesses are only able to use 1% of customer data.
Here comes the problem solver: predictive analytics.
With the help of predictive analytics, companies can dig deeper into their customer data to provide personalized customer experiences.
Here are three examples of creative uses of predictive analytics that are changing the way brands interact with customers:
American Express Uses Predictive Analytics to Forecast and Stop Fraud
The finance sector is the most important industry. Every day, we hear the news about credit card fraud, bill discounting fraud, check kiting, demand draft fraud, fraudulent loan applications, fake bank inspectors who call the customers and ask for their details, etc.
These kinds of fraud definitely can ruin a customer’s experience. Plus, customers want to have a card and brand that they can trust.
American Express uses predictive analytics to predict potential fraud and identify the customers at the highest risk. The company can then take proactive action like warning customers about potential fraud attempts and what to look for by calling them or via emails.
By finding potential attacks before they occur and helping customers stay safe, American Express creates loyal customers and has one of the lowest fraud loss rates in the industry.
Netflix Uses Data for Personalized Recommendations
No list of companies using predictive analytics in a creative way would complete without Netflix. Everything Netflix does is based on data, from the shows it creates to the movies it promotes.
Netflix collects enormous amounts of data on each user that it puts into an RPA-powered algorithm to predict what they’ll want to watch next. Data like demographics, watch history, ratings, and preferences influence what the algorithm will predict, and it’s almost always accurate.
About 80% of what is watched on Netflix is due to the recommendations. Having such a strong system saves the company $1 billion a year in customer retention. I am a Netflix fan myself, and my genre is sci-fi. And, trust me, my whole watchlist and suggestions are full of sci-fi and thriller movies.
deltaDNA Converts Gaming Analytics by Boosting Player Engagement
deltaDNA is the games industry’s leading real-time analytics and marketing platform. The combination of market-leading deep data functionality, ultra-high performance, and a flexible, open environment allows users to maximize engagement and lifetime value through player segmentation, predictive modeling, and real-time targeted interventions.
deltaDNA’s solution is able to process 99.9% of messages within 200 milliseconds, enabling games to respond to player behavior in real-time.
As a result, player engagement increased by up to 350%, helping to boost retention and loyalty. Sophisticated predictive data analytics provides the foundation for improved market insights and competitive advantage.
Sidney Alexander is the CEO of Digitzd.
“For one of our clients, we built a chatbot that could learn how to deal with…”
Common queries by scanning the customer’s message for keyword combinations, and then deploy an appropriate response. It was designed so that customers didn’t know they were communicating with a robot. At the start, when the database of strings was small, the bot failed and the customer would be transferred to a human agent, but over time as the database grew, the chatbot grew to handle over 50% of queries.
Our biggest win was when we solved how to deal with misspellings or local slang. We would return a phrase like, “I am sorry, I don’t fully understand what you mean, can you rephrase that please?” When the customer rephased their query, the chatbot was able to identify it correctly over 80% of the time.
It was a huge time and money saver for the company, and the best part is the chatbot is still getting smarter and learning more as the database grows. It has allowed them to scale customer acquisition without scaling their customer support. Win-win.
Manny Hernandez is a CEO and the co-Founder of Wealth Growth Wisdom, LLC.
“Predictive analytics can be used to help organizations allocate their resources more efficiently and productively…”
Retailers can combine insight from their store footprints, logistics, and customer behavior to perfectly plan staffing levels in advance. Doing so will help ensure that customers have a smooth and better experience. Therefore, companies can become more efficient, streamline costs, and reduce resource waste. Also, customers can receive the timely and personalized experiences that they have come to expect. It’s a win-win situation.
Ajit Ghuman runs Product Marketing at Narvar.
“The Australian company’s NOC (Network Operations Center) tracks the IP addresses of surveys…”
As they are being submitted on a giant monitor that covers all Australian states and operating units. Surveys filled out with poor scores are tracked real-time, and their technicians travel to troubled areas even before the compliant calls come into the call center. It helps to ensure a positive customer experience and keeps call center costs low.
Oren Greenberg is an on-demand CMO and the founder of Kurve.
“The most creative use of predictive analytics to improve customer experience I have seen is to…”
Create personalized messages based on customer behavior that fits into specific segments or cohorts. Here are two examples:
- An eCommerce website detects a dip in purchase frequency based on historical data. It then offers relevant products to those customers – which it would do anyway – but in this instance, it would offer a discounted rate or promotion to try and retain them. So, it was tailored to the customer type based on behavior.
- A B2B SaaS company predicted behavior based on product use rather than purchase. It could predict when people were using the product less, and it would be proactive in outreaching to the customer to rekindle interest via the customer success team.
Both are very powerful uses for predictive analytics.
“Analytics can help you detect the precursors of change in customer behavior…”
Using predictive analytics is like looking ahead with a telescope instead of glancing into the rear-view mirror.
Here are some great ways businesses have used predictive analytics to improve their customer experience:
- Hyper-personalized marketing for your customers: Predictive analytics can help businesses know what their customers want even before they visit their website. For example, sportswear-maker Helly Hansen achieved a 170% conversion rate by combining analytics with geo-targeting to predict what its website visitors were likely to buy based on weather forecasts in the visitor’s location.
- Use analytics to estimate customer behavior: Analytics can help you predict customer loyalty and customer churn. For example, AT&T Business has a customer experience machine learning system that parses millions of unique data elements, such as customer effort, cycle time, retry rates, etc. throughout customer project lifecycles. It then predicts whether a customer will stay a promoter or if they will start moving towards neutral or detractor status.
- Create a virtual concierge service for your customers: Use predictive analytics to deliver seamless instant gratification to your customers, tailored to their needs. For example, Spotify and Netflix change their suggestions for what you should consume next based on what you are watching or listening to in the moment.
Malte Scholz is a Product Manager and co-founder of Airfocus.
“One of the best uses of predictive analytics that we’ve had was when we had a very high bounce rate on one of our main landing pages…”
Analytics told us that there was lots of traffic coming in, but none of it seemed to convert for some reason.
We had a feeling that our offer was good but that the call to action and the “meat” of the offer were too far below the fold. We ran Hotjar to find out what the visitors were doing and discovered that most people scrolled just below the fold but never scrolled down far enough to get the most important information that would make them convert.
This combination of our own intuition, Google Analytics, and Hotjar’s heatmaps actually saved us many times, and it’s the reason why some of our offers have been so successful.
Denis J. Whelan
Denis J. Whelan is the Chief Executive Officer of Projector PSA.
“Delivering project-based services that exceed client expectations is complex, demanding, and high stakes…”
Project managers need to deliver on-time and on-budget, yet projects often change course mid-stream and unforeseen challenges go unnoticed. Next-generation services teams leverage predictive analytics to anticipate project fluctuations so they can take proactive measures and better serve their clients. Workstream forecasting and project budgeting are two examples of how predictive analytics can be leveraged to deliver better project experiences.
Most organizations have some way of forecasting their work, but visibility and accuracy typically declines further ahead into the future. Predictive analytics allows organizations to monitor their forecasts in tandem with their historical performance, analyze variance and variation, and refine predictions on a recurring basis. This helps leaders understand in advance if they have the teams and capabilities required to take on another project or if their pipeline is oversold. Better planning leads to better client outcomes and an overall better customer experience.
Service teams can also leverage predictive analytics to project a budget’s estimate-at-completion, in addition to actuals-to-date. By projecting labor and resourcing needs along with the foreseeable incurred costs, project managers can predict the outcome of a project. An on-time, under-budget project makes for happy clients.
To leverage predictive analytics in both of these situations, organizations need the ability to gather accurate, clean, and consistent data. Technology that is specifically built for service-based organizations, such as Professional Services Automation (PSA) software, provides this data as a single source of truth. PSA software not only acts as a platform for predictive analytics, but it also helps organizations exceed their clients’ expectations by giving team members confidence that they understand what it will take to make a client contract successful with fewer unexpected mishaps.
Kateryna Reshetilo is a Marketing and Business Development Manager at Greenice.
“One creative use of predictive analytics is an AI-powered prize recommendation engine to help you decide on the optimal award for your architecture project on a crowdsourcing marketplace…”
This engine is an integral part of Arcbazar.com, a crowdsourcing marketplace for architects and designers. On this platform, clients can create competitions for their projects and receive design offers from all over the world. As a result of such a competition, a client chooses the top three winners who will receive the award.
Before the recommendation engine, clients had to come up with the award amount themselves, and it was a problem. They didn’t know what award would be appropriate for their projects. On the one hand, it should attract talented designers, but on the other, it should not cause the client to spend a fortune.
That is why the platform owner decided to create an ML-based recommendation system. It learns from over 1,500 competitions conducted on the platform since launch. The system is looking for the relationship between the project award selected by the customer and the characteristics they defined. Each time a new project is completed, the database is updated and, thus, the recommender engine learns again and again to perfect itself.
David Morneau is a co-founder of inBeat.co.
“Here are just a few of the many possible creative uses of predictive analytics…”
- Make your marketing personalized through behavior-based customer insights and customer sentiment analysis.
- Forecast what messages or offers are most likely to help your customers. Create customized solutions.
- Predict life events and approach your customers with recent offers or products exactly when they need them.
- Increase early product adoption by analyzing your customers’ influence in their network.
- Plan staffing levels in advance. Streamline your operations to offer your customers a smooth experience. Mitigate challenges by predicting demand spikes and rush hours.
- Analyze areas for service improvement by getting insights into the motivation and needs of your customers.
Alistair Dodds is the Marketing Director and Co-Founder of EIC Marketing.
“I really like the work Indatalabs did in the fitness app space…”
It’s incredibly useful, targeted, and optimized for the user experience. The app enables users to enjoy personalized training recommendations based on their fitness level. If a user is getting used to a certain level, then the system pushes the user on to a more challenging workout regime. If a user is still getting used to a given level, then the app will monitor and give the user time until they see improved performance.
This real-life application of taking the work of a personal trainer and putting it into app form is an incredibly powerful and well-applied means of improving the customer experience.
Tom Allen is the Founder of The AI Journal.
“One fascinating use of predictive analytics is in the government and public sector, as well as in banking and finance services, for fraud detection…”
Also known as fraud analytics, it’s becoming an increasingly well-paid and sought-after skill set.
You need someone who knows their stuff about how a fraud transaction happens as well as how to build the architecture and data to protect customers or business transactions on a predictive model.
Banks such as the Commonwealth Bank can use analytics to predict the likelihood of fraud activity for any given transaction before it’s authorized in as quickly as 40 milliseconds of the transaction initiation. The more times it happens, the better its predictions become through testing of the algorithms and patterns using machine learning.
We recently had David Gyori do an article for us showing 20 reasons how banks can use AI. This article also explains through these points why banks need it in order to better predict their consumer trends. You’re already seeing the gap between FinTech companies, such as Monzo and Starling, compared to traditional banks. FinTechs give themselves better access to consumers’ data, therefore predicting where a customer’s next spend will be and in what amounts.
Jeff is a Founder and CEO of Fig Loans.
“We started as a collaboration with the United Way, and that original product ultimately became the data source for…”
Our core technologies: risk models and customer service AI. I call it our ironman suit for customer service. Our service tech automates responses to inquiries so customers can get the answers they want, faster. First, our system identifies exactly who the customer is and their relationship with us. Then we use natural language processing to guess their intention. Last, our predictive models generate a response based on all the historical questions we’ve seen before. High probability predictions are sent automatically, while lower probabilities are reviewed by our team and the results fed into future predictions. As a result, our 1.5 service FTEs serve over 20,000 active customers with a same-day response time.
Will is the CEO and founder of Assistive Listening HQ.
“We started by collecting data regarding what users do once they are on a particular product page…”
After adding the product to the cart, 78% of them went to the search bar to search for a related product. Using this data, we made a list of the top five searched queries and displayed them under related products on the product page itself. This not only saved our customers time but also exposed them to other products that are searched by similar customers.
It is basically using predictive modeling to figure out the future actions of users who put a specific product in the cart, in this case, CHG 408 – Charging Bay (8-BAY). An interesting side effect was that we not only improved the user experience but also improved our conversion rates!
Mateusz Dziekoński is a Big Data Specialist at Zety.
“Live segmentation based on demographic and behavioral data is the future of customer experience for apps and websites…”
Big data solutions are able to classify users precisely in the area of their purchase intent, technological awareness, or social group based on historical data. Such information is a game-changer for personalization opportunities.
Better-suited content and even layout may lead to higher engagement, which as a result improves customer experience significantly.
Nobody wants to be treated like everyone.
Deepu Prakash is the SVP – Process, and Technology Innovation at Fingent.
“Almost every industry segment benefits from predictive analytics…”
Although it might not be the best option for every industry. For example, prescriptive analytics could work better for healthcare in providing actionable insights.
Nevertheless, one of the most creative and successful uses of predictive analytics to improve customer experience, which in turn led to tremendous business growth, has to be that of mBank. (Reference: SAPnews, mBank: Picked Up Your Reward Today?)
Data is really important for banks, as they do not have frequent face to face interactions with customers. mBank used customer data from card transactions to deploy a predictive analytics model that would help recommend relevant products and services, not just their own, but also those of partner organizations.
These recommendations involved anything from consumer electronic products to restaurant deals. This helped mBank stand apart from the competition by building up a network and by providing an enriching experience to its customers rather than just pure banking.
They also used predictive analytics to get dramatic results and transformative response rates for their loans and financial products by gaining insights from tracking customer needs.
The result was:
- 400% higher hit rate for non-mortgage loans
- 250% higher hit rate for savings products
- 200% increased hit rate for insurance products
Predictive analytics takes time, with the average time between deployment and gaining insights being two to three months. But once deployed, the predictions will only get more accurate over time.
Biland Sadek is a Regional Commercial Director for MEA & Duty Free at Philip Morris International (PMI).
“DiDi has a huge information advantage, as they generate terabytes of data through their transactions…”
They are able to create predictive models by matching the data they collect on every aspect of millions of rides with end-of-ride ratings from customers. Their models predict:
- What sorts of experiences typically produce promoters among its customers
- Which experiences produce detractors
As a result, they don’t need to gather NPS (Net Promoter Scores) from their riders, as their models generate a rating score for almost every ride. Research showed that their predicted scores (80%+) matched with what customers say in traditional NPS feedback.
This gives them two advantages:
- They provide almost instantaneous modeled feedback to their drivers.
- They instantly identify situations where there’s some need for relationship or service recovery, triggering an intervention.
For example, if their algorithms identify a ride that took longer than it should have to reach the destination, the company can issue an apology or a credit before the rider even exits the vehicle. If things went especially well, then their app can prompt the rider with ways to tell friends about the ride-sharing service’s benefits.
This approach offers a glimpse of how predictive analytics can help figure out whether customers are promoters or detractors, and then close the loop and enable direct action.
Though predictive analysis is important, I argue that equally important are prescriptive analytics (guided actions on specific customer opportunities that flow from predictive insights). When applied to sales or marketing, prescriptive analytics can help companies improve their RoI, optimize their conversion rates, or maximize their profit margins.
Pushpraj Kumar is a Business Analyst for iFour Technolab Pvt. Ltd.
“Predictive analytics provides insights into what your customer might do in the future…”
Predictive analytics observes customer behavior and visualizes what happens in real-time, analyzes the customer interaction, creates unified audience segments, and delivers the targeted content automatically, which is generally based on customer profiles. There are many companies that use predictive analytics by which the user can change their pricing
models based on their age or gender. By using predictive analytics, you can reach consumers at the right time.
Predicting customer needs is one of the most important uses for predictive analytics. By using the purchase data of a customer, we can easily predict his/her next purchase needs very efficiently. Predictive analytics help tailor a customer’s experience as it happens. Real-time product feedback observes the customer’s actions, such as watching a certain show or skipping certain songs, and impacts the next recommendations they will receive.
George Elfond is the CEO of Rallyware.
“The role of AI and predictive analytics in customer service management is becoming essential…”
And it is playing a major part in boosting customer experience. While the outdated LMS systems train frontline employees retroactively, predictive analytics allow for proactive training. The old-school corporate training is slowing down learning progress because it doesn’t validate new learning absorption, it is not personalized for each employee’s skill gaps and current performance, and it’s not delivered at the right time when it’s needed.
One of the biggest benefits of predictive analytics is that it allows companies to deliver the right training at the right time to their frontline employees so that their efforts have the greatest impact on the customer experience. Predictive analytics are powered by different types of data, such as POS data, Yelp reviews of a specific location, sales performance of a manager, guest reviews of a staff member, their learning progress on the latest corporate instructions, and hundreds of other data points. This allows creating that unique Netflix level of personalized learning experience that translates into the ultimate high level of customer service. Our data shows that by using predictive analytics to upskill the frontline staff, equipping them not only with the corporate policies knowledge but also with performance-driven new skills, results in a whopping 63% increase in customer satisfaction, on average.
Brian Atkiss is the Director of Advanced Analytics at Anexinet.
“Predictive analytics have been used in a number of ways related to customer experience…”
A great example is from a large insurance company that was collecting and analyzing Net Promoter Score surveys, call center data, and social media data and joining that data with other policy and transactional data. They used text analytics to classify the unstructured text from the surveys, call center, and social media data and also identify sentiment for each customer interaction. Then they were able to use certain data points to identify the transactional and policy data (demographics, policy type, policy changes, cancellations, etc.) associated with the customer interactions and build predictive models to identify customers that were satisfied or dissatisfied. They could then create workflows associated with each to either increase loyalty among the satisfied customers or attempt to increase satisfaction among the dissatisfied.
Shiv Gupta is the CEO of Incrementors Web Solutions.
“The most basic, but perhaps the most important, type of analytics is predicting customer needs…”
One use for predictive analytics is to determine historical and recent data to forecast future events, trends, and behaviors. Predictive SEO is the exact opposite of reactive SEO. Here, instead of reacting to trends, predictions are made through detailed analysis to predict the next keywords and search trends to generate content or tweak content according to the keywords. Furthermore, predictive capabilities enable brands to be proactive, empowering them to work on expectations and needs, and successfully serve their customers.
What creative uses of predictive analytics has your company employed to improve customer experience?