Building customer loyalty through data maturity
An opinion piece by the senior data scientist at Ascent.
May 2, 2022
An opinion piece by the senior data scientist at Ascent.
In this digital age, customer loyalty has never been more important – or more competitive – and brands that truly connect with shoppers’ habits and preferences on social, environmental and lifestyle levels are winning.
Two key factors have contributed to the growing importance of customer loyalty. First, the pandemic prompted a shift in consumer behavior, accelerating the digitalization of customer experience and levelling out the loyalty playing field − customers are gravitating towards brands that understand their needs and share their social values.
Second, technology advancements have given brands of all sizes fresh new ways to engage with customers. Consider blockchain and web3: As crypto adoption becomes increasingly mainstream, we’re seeing market leaders such as Starbucks explore new token-based loyalty programs.
Most organizations are somewhere on this journey. Many would agree that data is the most important driver for customer acquisition and engagement. More recently, by and large the data science/machine learning/artificial intelligence community has been shifting its focus from being model-centric to data-centric.
Getting specific with metrics
Let’s take a hypothetical example of an e-commerce retailer we’ll call TBS that specializes in organic, keto and vegan products to illustrate when and how an organization approaches data maturity in preparation for customer loyalty success.
With a strategic goal of increasing customer lifetime value by 20% in the next two years, the retailer invested in a data infrastructure and utilizes data analytics, data science and machine learning to achieve the growth goal.
However, customer lifetime value alone - the historical spend of a customer over time - is too broad a metric to be effective and doesn’t reflect the more granular detail of day-to-day business-customer engagements.
Instead, consider loyalty projects only where the metrics of interest are straightforward and actionable. If data is of insufficient quality or incomplete, it should be deprioritized – or better still, the goal can be reshaped, or a metric chosen for which ‘good’ data is readily available.
Managing large amounts of data
Big data is a side effect of TBS’s success, with data sources including historical transactions, third-party suppliers, social media, and Google Analytics. To align customer experience, TBS built an omnichannel database using Microsoft Azure, with the data engineering pipeline and platform as the core component.
Managing complex datasets requires understanding how new data aligns with existing reports, dashboards and expert domain knowledge. If the gaps are unacceptable, it’s best to identify the root causes before utilizing the data or to carry out a test run of advanced analytics projects on a slice of customer data of decent quality.
When not enough data exists
When TBS explored the customer journey on its mobile app, it became clear that some customer information within a login session was not being captured in its database.
In this case, an alternative to proprietary data may be used − for instance open-source data or third-party data. A mature data solution can also work, for example Geolocation data from the Office for National Statistics (ONS) or Customer profiling by Experian.
Data augmentation is another way to increase data volumes and dimensions for classical customer data. Where data is insufficient in representing a desired but absent customer segment, it can be simulated, and the actual dataset be reverted to for downstream projects.
When data is poor quality
If data is poor, the derived customer profiles can misrepresent the buyer community, which can lead to ineffective or incorrect incentives in customer interactions.
TBS planned to launch some pop-up stores and needed data relating to the relationship between customer locations and customer purchase behavior to decide on shop locations and what products to supply.
It had only 72% of customers’ addresses in its database, as some were missing due to being guest transactions going through a third-party website. Secondly, the postal codes were incorrect or invalid when they were joined with geographic and socioeconomic information.
As a result, apart from fixing the issue in historical data, their UI team replaced the free text box with a postal code lookup.
Effective resource allocation
To be sure, the more engagements with customers, the more data a business collects. And it’s the ability to leverage this data to delight customers that creates a competitive edge.
Answers to these questions can help to create that edge: How much have we invested in the capability to surface this data? How do we allocate the headcount of a data team? Do we invest in customer insight software or an application that automatically spits out the expected output once data is plugged in? Do we hire more data talent and integrate it into customer loyalty functions?
The benefit of tooling is that people with limited data science knowledge and skills in an organization can generate seemingly useful insights. However, this shouldn’t replace efforts on experiments and tests in creating the most effective rewards. Purchasing an ‘automatic’ solution requires careful evaluation.
Investing in people with shared goals and vision can improve team dynamics to ensure successful data analytics projects that realize value.
Data privacy considerations
It’s worth being mindful that customers should be informed explicitly about how their data is used and for what purpose. A frequent cause for companies being fined for a GDPR violation is non-compliance with laws on data processing of customer information such as age and gender.
We recommend that customers are informed explicitly about how their data is used and for what purpose. We suggest that you review your privacy policies carefully and regularly before embarking on any customer loyalty project to ensure projects run smoothly.
So, am I ready?
There’s no instant solution that would result in repeat customers and 5-star reviews. If an organization prepares for the long term, values the skills and capability required to make loyalty projects a success and understands the value of its data, that’s a big step towards maturity.
It’s a journey that’s best started as early as possible. The second best time is now if your company already is somewhat established. Start small and prove value quickly. With even the smallest of incremental improvements and accelerating your investment organically, you will see that your efforts will soon be rewarded.
You May Also Like