Making Data Science Actionable for Marketers
Written by Tom Coppock
The previous posts in our series on churn analysis looked at a high-level approach to analysing churn in a subscription business model, before getting into more detail exploring the steps to take in conducting churn analysis.
By this stage in your project to boost loyalty & retain customers most of the hard work is done – you now know what the leading indicators of churn are in your business & have predicted which customers are at risk of leaving. With the math project complete, the most important step remains: what to do with this information to keep your customers.
Making insights actionable is key to achieving ROI from your data science investment. Here we look at 3 important strategies to apply these insights to improve business results:
- – 360° Data Visualisation to Inform Decision Making & Guide Priorities.
- – Micro Segmentation to Drive Cross Channel Retention Marketing Campaigns.
- – Automated Churn Risk Scoring & Triggered Calls to Action.
360° Customer Data Visualisation to Inform Decision Making & Guide Priorities
The ability to analyse & visualise data has evolved massively over the last few years, but too often marketers rely on analytics & dashboards that are baked into channel specific marketing automation, web analytics or CRM tools. While useful, each system will only provide a narrow view of the data they have captured & not a holistic view of both the digital & real-world customer experience.
According to Salesforce the median number of marketing data sources used by a company is projected to jump from 10 in 2017 to 15 in 2019. Each new system is a potential silo of data which compounds the challenge facing marketers trying to get a true insight into what their customers are doing.
To identify trends, prioritise investment & react in a timely manner, marketers need a consolidated view combining insights from marketing & customer engagement systems with business metrics, real-world interactions & data science analysis.
Tools like Looker, Power BI & Tableau can all be used to create marketing dashboards visualising data from multiple source systems, saving marketers the legwork of performing analysis in each individual system.
Creating dashboards to deliver insights into specific scenarios such as churn takes this a step further. Rather than viewing all the metrics for a department, this approach provides stakeholders across the business with actionable insights that can be used to address specific challenges such as churn.
Identifying churners needs data from a range of sources (sales/billing data, service/product usage data, client engagement data, survey/csat data etc.). This combined data gives a single customer view that can be analysed for churn risk, ROI/ROAS & customer lifetime value. Visualising the data together allows you to prioritise next steps such as focusing on retaining the highest value customers at risk of churning while showing which marketing segments or personas are the ones at most risk.
The following example uses Power BI to create a dashboard to show trends & metrics relating to gym membership churn.
Figure 1 – Overview of Churn metrics, filterable by segment, location etc.
Figure 2 – Exploring lifetime value & churn risk by marketing segment & location (exportable to Salesforce Marketing Cloud or other Marketing Automation platform for retention marketing)
Micro Segmentation to Drive Cross Channel Retention Marketing Campaigns
Visualising the results of your churn analysis gives you the priorities for next steps & an understanding of the segments at risk, alongside the likely impact on the business. From here you can export the segments back to your marketing automation, CRM, data management platform or social/ads platform for use in campaigns or to change business processes to create a better experience.
In our example we have built the ability to export the segments into the marketing automation tool right into the dashboard, but the same results can be achieved if you are doing a stand-alone analysis or if you have automated the churn segmentation & scoring as part of your data warehouse workflow.
Building your segments from a consolidated data set in this way offers a lot of advantages over creating segments in a native tool like marketing automation or data management platforms. It allows you to account for all demographic, purchase history & real-world behaviour – going beyond the click & cookie data used by marketing tools – to create clean, accurate segments that are then published to all engagement channels in parallel.
In our experience, the best results come from a micro-segmentation strategy. If you start with one of your core segments (or target personas), maybe “Fit & Focused” in our gym scenario, then cut by churn risk, timeframe for the churn risk, even down to location & channel preference.
Use insights about what is driving churn in this group to re-engage them with a goal of reversing the behaviour creating the churn risk. This could be through marketing campaigns – if we want them to visit the gym more, perhaps a promotion for an after work out smoothie – or may require collaboration with teams outside marketing.
If repeated failure to book a preferred workout class is a leading indicator of churn, hold back space in the class for the churn risk segment. If a failed payment is the leading indicator, let them have the next month free rather than auto cancelling.
Here the greater your capability to personalise the customer experience the better. Making offers as specific as possible, for example calling out their local gym or favourite class in the message, will increase the chance of them engaging with the email, SMS or tweet offering priority access to the Zumba workout on Saturday morning. Once launched A/B test the campaign results to ensure your message is as effective as possible.
Automated Churn Risk Scoring & Triggered Calls to Action
Another approach is to use a more dynamic experience. Instead of running specific retention campaigns, create a series of (multi-channel) “always on” nurture programs that are triggered by different levels of churn risk scores for different segments.
These nurture programs are set up to display “calls to action” in your website, mobile app, email or social campaigns. This requires the churn analysis to be automated, with continuous or scheduled updates of scores in your marketing or CRM profiles, for example ranking the predicted 3-month churn risk of a given individual on a scale of 0 – 10.
When the risk increases above a configured threshold the marketing automation rules will then activate the churn reduction campaign, offers or content.
While this approach requires more investment up front in automation for both marketing & machine learning analysis of the churn risk factors, in the medium term the overall effort is reduced as create once, “evergreen” campaigns replace one-time churn reduction offers or campaigns.
Getting Started…
This concludes on our series on churn & retention marketing. We hope this content has helped inform your thinking about how data science & marketing can work together to drive business outcomes. If you’d like to learn more, request a free consultation using the form at the end of this page