5 ways data and text analytics improve customer retention
Customer retention strategies fuelled by data ultimately influence how your team will approach customers — it’s proven to drive profit. In fact, “executive teams that make extensive use of customer data analytics across all business decisions see a 126% profit improvement over companies that don’t” (McKinsey, 2014).
This is no news. Among 334 executives surveyed by Bain, more than two-thirds said that their companies are investing in data and analytics. And the expectations are high. 40% expect to see “significantly positive” returns, with another 8% predicting “transformational” results (Bain & Co, 2017).
While the intention is there, according to Forrester, “only 15% of senior leaders actually use customer data consistently to inform business decisions” (“The B2B Marketers Guide to Benchmarking Customer Maturity”, Forrester, 2017). So, companies do realize the need for data but expect some sort of magic to happen in order to implement?
“Influencing customer loyalty […] doesn’t require magic, it requires data – usually data that you already have but aren’t using to full advantage. Regardless of industry, most organizations today generate mountains of data. In fact, many customers tell me that they have so much data that their biggest problem is how to manage all the data they have”, says Mike Flannagan, vice president, and general manager of Cisco.
5 Ways Data and Text Analytics Improve Customer Retention
1. Develop a data roadmap and stick to it
As many as 30% of the executives in the aforementioned Bain & Co study said that they lack a clear strategy for embedding data and analytics in their companies. McKinsey’s findings show that taking an integrative approach, meaning seeing analytics as a strategic driver of growth instead of using it in a silo or only as a part of IT, ultimately leads to achieving the desired result (McKinsey, 2014).
Successful companies do two things differently: First, they make use of the data they have. Second, they implement the organizational changes once they understand what the data tells them. So, you have the data – make sure you actually use it and enforce any changes needed in the business to make it happen quickly.
A good approach is to develop a data roadmap and stick to it. Steps that you take within the organization can be to:
- Ensure corporate KPIs are automated, scalable and repeatable.
- Gather key stakeholders and define the top 3 business problems you want to solve.
- Categorize the issues into data vs. systems issues (often you’ll find that the issue is not with “data” at all, but with how people use it or manage it).
- Prioritization of tasks is required along with assessing the technical feasibility of your plan.
- To stay on track, reassess progress every 3 months.
- The human factor – ensure behavioral change
Another key factor is hiring senior executives who take a hands-on approach to customer analytics. Not only do they need to understand the importance of analytics but also have the skills to analyze it themselves, so use this as a benchmark when hiring.
Although 70% of companies have data strategies in place, many will fail to deliver what’s needed due to one factor alone: people. You may have the most advanced tools and excellent data scientists; however, all efforts fail without the correct behavioral changes needed internally to ultimately take action (Bain & Co 2017).
Employees may not be committed to using data analytics, internal teams may not be communicating with each other, or the data solutions adopted aren’t user-friendly. Behavioral change, continuous monitoring of results, along with a “one-team approach” is needed to ensure that advanced analytics within an organization can survive and prosper (Bain & Co, 2017). No surprises here, behavior change being the hardest part of any performance improvement plan and why as many as 38% change efforts fail (Bain & Co, 2016).
2. Only focus on high-quality leads
Customers are less likely to churn if they are similar to your primary target customers. If you have access to data about both your customers and a list of potential customers, this is a great opportunity to focus on only those who are less likely to churn.
How? By applying algorithms comparing the features and characteristics of your customers to those of your potential customers. Those that have similar characteristics (FTE size, annual spend, job title, type of industry) to your existing customers are probably those most likely to want your product, to find it valuable and therefore stick around. Your segmentation now becomes crucial. Each customer segment provides you with distinct features that help easily identify your next customers.
For example, tools like HubSpot provide this type of information in an integrated way, where you can see characteristics and patterns easily.
3. Use machine learning methods to create predictive models
Companies analyze data using different types of analytics, including predictive analytics, which is used to look at the relationships among different metrics.
To create solid customer retention strategies, we can use predictive analytics to make predictions about the future, by looking at historical data, to learn what customers may like or dislike.
Often, you might be overwhelmed by the number of variables you have to manage and analyze all at once. Although you may have a highly skilled data analyst at hand, it’s still time-consuming and labor-intensive to manually and quickly sift through the sheer volume of data to find the optimal predictive model.
To create the best predictive models of retention, rely on the power of machine learning to quickly and accurately uncover the underlying reasons why customers are churning or why they’re loyal to your brand.
Machine learning uses math, statistics and probability to find connections among variables that help optimize important outcomes such as retention. These models are then applied to new customer data to make predictions.
Machine learning algorithms are iterative and learn on a continual basis. The more data they ingest, the better they get. Compared to human performance, they can deliver insights quickly thanks to the processing capability of today.
For example, you can use analytics to identify which up-sell or cross-sell products will be the most relevant based on your customer’s past purchase or browsing history.
Often, companies don’t have employees with high-level analytics (data science) skills. Third party providers can provide a solution that automates data integration and analysis.
4. Get data-driven insights with text analytics
To get deep, data-driven insights, don’t forget to analyze your free-text responses to your open-ended survey questions. If you don’t you may well miss them!
You can do this with text analytics solutions. With a text analytics tool that uses sentiment analysis, it’s easy to spot customer pain points.
And, if you collect lots of data, make sure you actually use it. One study found that only 15% of senior leaders actually use customer data consistently to inform business decisions (Harvard Business Review).
At Thematic, we have developed an AI algorithm that automates analyzing free-text feedback in surveys using machine learning and natural language processing, and in essence, simplified the way businesses are getting insight from their customer data.
5. Segment to focus on retaining the right customers
Using data analytics to segment people into different groups means you can identify how each segment engages with your brand and product. This then allows you to look at each subgroup and draw insights, followed by adopting different communication and servicing strategies to increase retention of your most wanted customers.
Analyze data such as your customer demographics, lifestyle, products purchased by each category and type of customer, the frequency of purchase and purchase value. In this way, you’ll discover which type of customers are driving the most revenue. Some cost too much to deliver revenue, so you’ll know if you want to focus your efforts on.
Understanding the difference between these types of customers, can in some cases make or break a business, especially if you’re just starting out. Knowing customer value is crucial to be able to make critical decisions. You can segment by historical value, lifetime value, value over the next year or the average customer value by segment. Using the right segmentation, you’ll then create highly targeted product recommendation offers. Segment your customers to offer relevant discounts for different channels (in-store, online, mobile). Mix it up a bit, every customer doesn’t have to receive the same offer.
Another useful way to use segmentation is to monitor the time-sensitivity and seasonality of your promotional codes. By monitoring sales data, you can see whether these codes are redeemed more often in the morning or afternoons or perhaps straight after a sales communication. The more you know about what a demographic responds to, the more you can focus on taking the right actions.
Top 3 Tips for Analysis
Gather multiple data points to be able to make relevant recommendations.
Be pragmatic and avoid making assumptions from solely one piece of data. Because someone living in California buys winter boots doesn’t mean they want to be bombarded with similar product suggestions. Maybe they bought them for their sister who lives in Chicago!
Leverage social proof where you can.
If your customers don’t respond to certain products, maybe all they need is a little reminder that others similar to them are using them and are happy with them. Pull in positive testimonials from surveys and social media comments to your marketing communications and website.
Remember: it’s the ability to swiftly translate insightful data into concrete action that counts.
It’s a fact: better data means better results. If you don’t have good data now, you can test your way to better data. Just by improving your internal data collection, you can often arrive at better data. In other cases, you might have to purchase better data. Good data is not static, it’s a continual process of observing, acting and learning.
Finally, the challenge of the vast data volume that large businesses have, is also the opportunity. Bringing together structured and unstructured historical data across organizational silos, and combining it with key data about ongoing customer interaction provides a compelling opportunity to influence customer experience in real time.
This article was published here first.