This paper will explore the use of data analytics tools and statistical modelling to forecast and anticipate customer behaviour in order to bring focus to business strategy more effectively and to tailor marketing and product offering for customers, to enhance their experience.
How can we improve customer experience?
Improving customer experience should always come from a place of answering the needs of the user. Providing easy solutions to their problems, be that with a product or service and delivering them in a way that puts the customer first, is the most simple formula to improve customer experience. One tool that can help businesses achieve this through an approach which puts the user first, is predictive analytics.
Improving the customer journey
Customer experience is a collection of all the small steps that build towards the customer finding the right product. The experience needs to meet or exceed their needs, within an environment that makes this process easy and straightforward, with a great service experience from the very beginning all the way right through to the end. In this sense, there is an opportunity to improve the customer experience at every touchpoint of the customer journey.
The customer experience for a brand begins the moment a customer sees their advertising or search result online, or the moment they see the front of the store on the high street. From that point on, the customer experience starts as they begin to investigate what the brand is offering and assess if it meets their needs, through to initial purchase, product experience and hopefully repeat purchase.
Many of these customer experience elements we now take for granted as ‘common place’ on our preferred online sites, because they make the experience so easy and convenient to use. It’s not until we then go somewhere that gets it wrong, that we remember how important it is and how easily it turns us off and can lead to us just abandoning the shopping mission in frustration.
Poorly designed websites that make the products hard to find or clunky checkout processes that don’t offer our preferred payment method or delivery options, is enough to make us switch to another site in search of a more convenient option. What’s more, as consumers our expectations on what is a ‘good customer experience’ is constantly on the rise, as companies know how powerful a tool it is to keep us coming back. This makes it increasingly important to get it right and as such, predictive customer analytics and the customer experience make great partners.
How to use predictive analytics to improve customer experience
There are a number of ways to improve the customer experience with predictive analytics. The process will usually start with a customer diagnostic analysis across the whole customer journey. This quickly gives an idea of what is working and what needs attention, meaning the focus can be to improve the right part of the journey first, in each case. Typically there are four broad areas where big improvements can be made:
1. Customer needs and demand forecasting
At the start of the journey understanding customer needs is critical. This is achieved through a series of data analytics and customer segmentation processes to dive deep into the behaviours and preferences of customers based on their historical purchase patterns. Predictive analytics can take these techniques a step further and add the predictive layer to forecast these needs. Forecasting customer needs with predictive analytics plays two key roles in the customer experience: it means the business can plan around what, when and where customers will be buying and make sure the stock is ordered and in the right place; and it acts as an early warning sign to spot changes in customer behaviour as a precursor to evolving changes in customer needs, meaning that the business can proactively respond to and address these changes.Using a combination of historical transaction data and retail finance data, companies will be able to help clients make significant improvements to their forecasting capabilities.
Case study - For one clothing retailer we were able to reduce their lost demand, (e.g. where they were not ordering sufficient product and therefore unable to sell), by £149 million using predictive analytics.
2. Personalised marketing
Reaching out to customers and prospective customers to remind them of the brand has become increasingly challenging, due to both the explosion in different media options and the sheer volume of competition for customers’ attention.Personalised marketing is about making sure to target each individual customer with a relevant message, at the right time, through the optimal channel. With millions of customers this is no mean feat and the skill comes from being able to package up the messaging with compelling, attention-grabbing content, the right call to action and, once you have the customer within reach, making sure the product, pricing and things like shipping, are all on the money. Predictive analytics allows companies to tailor these messages using data to create hyper personalised experiences, which can drive up conversion rates and return on investment for your marketing.
Case study - We were able to make significant gains in targeted marketing to a store card customer base, for one of our department store clients. Through predictive analytics and a structured test and learn program, we were able to achieve uplifts of 395% in response rates to some customer segments and almost $500,000 of additional sales.
3. Customer churn
It is a well acknowledged fact that it costs more to acquire a new customer than it does to retain an existing one and that finding ways to identify customers before they leave is a proven way to manage growth. Predictive analytics can be used to identify the likelihood of a customer leaving, by using their behavioural and transactional data to spot events or changes in behaviour that are known to be early indicators. These events may be something such as a poor customer service experience captured in a feedback survey or a change in shopping patterns, like visit frequency.The success of predictive analytics in preventing churn comes from the power of data analytics to identify the drivers of churn using the events and behaviours from previously churned customers as the learning data and then spotting these in the events and behaviours of current customers. To successfully prevent churn, the business needs to have the right actions and responses in place to recover the customer and ideally without using excessive resources, which results in margin erosion. Therefore the earlier the potential churn is spotted the longer the runway there is to do something about it and ideally the lower the overall recovery costs will be.
Case study - For one of our B2B financial services clients, we were able to predict potential churn up to six months out, which gave the business substantially more time to address the operational and service issues that were behind much of their churn.
4. Workforce scheduling
Predictive analytics can also provide the insights to help a business plan their workforce scheduling more optimally to make sure that they have the right level and type of staff online to meet customer demand and maintain productivity.Integrating the customer demand forecasting models with the store estate, call centre and warehousing data, can enable more accurate planning for staffing levels across the customer experience to meet demands and keep customers happy.
How predictive analysis will transform the CRM experience
Predictive analytics done right is a bit like cloning the best salesperson or customer service agent and planting them all over a business. It enables the business to effectively watch over every customer, spot the triggers that indicate a certain requirement, and react accordingly.
Simply put predictive analytics can help transform business operations from something that needs to apply ‘best-fit’ to meet as many needs as possible, to empowering them with the best of technology to offer a truly personalised service to each customer.
How can you improve the process of predictive analysis?
Our approach is to start with building the basic foundations of great customer insight to understand their needs and make sure the operational aspects for this are ready to go. This can be largely done using descriptive data analytics after which the business can start to optimise and become more sophisticated with the predictive models.
Starting with descriptive analytics.
Get to know customers by segmenting them into distinctive groups based on their behavioural differences.
Use these groups to track them through the entire customer journey – Where did they first hear about the brand?; What was their entry point (e.g. store or web)?; What did they look at?; What did they buy?; How did they pay and when and why did they return?
Model their profitability by working back through the journey, picking out the successful journeys that end in repeat purchases, a sale etc. and those that drop off.
Allocate the costs and profitability for each customer group, to identify the scale of financial return or otherwise at each stage of the journey.
Review these findings, with the goal of picking out the big successes and losses, and begin to develop hypotheses about how to possibly impact/change these. E.g. We could encourage more people to stay on the site and start browsing, if we could tailor the landing page that they see.
With this foundation of analytical insights, the business will be able to develop a set of areas to go after with a tangible prize at the end of it. It will enable them to identify the quick wins where there are obvious and easy changes to make that can be tested – such as trialling a few different landing pages based on the visible insights.
Only once the business has made a good start on the above, are they ready to start thinking about the role of Predictive Customer Analytics in the customer experience. At this stage, it should be clear which areas are going to show the greatest bang for their buck and thus be able to apply the power of predictive capabilities to further optimise the journey.
The time and effort spent on understanding the fundamentals of customer behaviour will mean the business has the basis from which to quickly build predictive models, get them working faster and reap the rewards earlier.
Where is the best place to use predictive analytics?
Identifying the best place to use predictive analytics will differ for every business and the answer will really relate to how their existing operations perform.
A traditional bricks and mortar business will likely derive most benefits from the customer demand and needs forecasting and workforce scheduling. Making sure the shelves are adequately stocked and the stores fully staffed when the customers are out shopping is a crucial requirement to get right. There is nothing worse for the consumer than finding empty shelves or insufficient product to complete their mission.
For an online retailer the bigger wins may well lie in basket optimisation ensuring product placement on the site and associated products that go together are well merchandised with attractive delivery offers or promotions for multi-buys.
For a subscription or membership-based business model, like a credit card or content platform, the biggest wins may well lie in managing churn. Having spent all the marketing budget recruiting new members and offering attractive entry rates the challenge is to keep them engaged with the product; be it regularly switching to the content platform or keeping the credit card front of wallet.
Predictive analytics comes into its own for these businesses because they give the business time to react. For example, the sooner a business can spot that a customer is using a product ever so slightly differently, which indicates the probability that they are distancing themselves from the product (or getting distracted by a competitor’s), the sooner they can respond.
The sooner a business can respond the less they will probably need to do/spend to get that customer ‘back on track’. Without predictive analytics it is probable that a business won’t spot that someone is churning before it is too late and then the danger is they start spending heavily to retain customers when it is too late.