Why CLV should be an Organization’s North Star Metric
There’s one metric that executives and management should constantly be monitoring, measuring, and trying to grow. No, it’s not total sales, cost of acquisition, or net new customers. It’s Customer Lifetime Value. Now there’s nothing wrong with the other metrics I mentioned, but the truth is, those metrics are all focused on the past. And yes, while it’s essential to take stock of the past and measure historical performance, it’s equally important to look ahead. Now the problem with forward-looking metrics is, well — they’re forward-looking — they’re projections, estimates, forecasts. CLV is not a tangible metric or set in stone like Total Sales, and as a result, people have a hard time grasping it. I don’t blame them.
Which statement(s) makes you feel more comfortable?
“Our total sales was $5M last month.”
“With an 80% accuracy rate, our total sales will be between $3M and $5M next month.”
“There’s a 60% chance of rain tomorrow.”
Right?
People generally don’t deal well with uncertainties or probabilistic events.
But if the stock markets have taught us anything, the stock market is always forward-looking. And businesses should do the same — People want to know how much you’ve made and how much you’re expected to make in the future so they can right-size their investments accordingly.
This article will examine CLV from the business lens and illustrate how to apply it to increase profitability and sales lift.
Note: This article is based on the mathematical models that Peter Fader and Bruce Hardie built. The lifetimes package built by Cameron Davidson-Pilon was used to frame the business applications below. We acknowledge all the work and contributions they’ve made to advance CLV science. Our aim is to translate the technical aspects into a business-friendly article and illustrate the benefits businesses can achieve by focusing on optimizing for CLV.
3 Business Ideas Driven from CLV
- Use CLV to Track Business Health and Identify How Your Best Customers Are Acquired
Customer Lifetime Value can provide you with a deeper understanding of your customer base and how much each customer is worth to your business.
For this article, we generated insights from the UK Online Retail Commerce Data Set (Full GitHub Code and Jupyter Notebook here). We can see that customers acquired in 2009 Q4 are the most valuable for the business. We also see that the 2010 Q3 cohort, despite being smaller (n = 583) than previous cohorts, has a greater lifetime value than other cohorts both before and after (e.g., 2010 Q2, 2010 Q4, 2011 Q2). What we don’t know yet is how these customers were acquired. Customer acquisition is important when optimizing for profitability in our business. We may be acquiring customers with high lifetime values, but through unprofitable channels.
We also notice that the 2011 Q3 cohort has a much larger lifetime value than the six other previous cohorts. To understand why we performed a quick analysis of orders which revealed that this cohort’s CLV was driven by an increase in average order value. Again, knowing how these customers were acquired (e.g., social media, customer referral, mass marketing) would greatly help create a leaner and more profitable customer base. We don’t need all customers, only the best ones! Perhaps customers acquired via influencer campaigns spend less than customers acquired via social media ads.
However, in a hypothetical scenario where you DO have customer acquisition information, the next step would be to run different marketing experiments to see which tactics increase customer value. For example, we can partition out the 2011 Q2 cohort into test and control groups where one group will receive a marketing campaign/promotional offer to get them to increase their average order value. Welcome to A/B testing. The same principle can be applied to other cohorts while tweaking different content elements.
Bring in CLV into your CAC to get a better picture of your business’s health!
Quick Note on Customer Valuation
From a customer valuation perspective, calculating customer equity is as simple as totalling each customer’s lifetime value. This number is essentially how much the entire customer base is worth to the company and can be used in many financial models when trying to accurately value a firm such as a non-subscription-based company like Lyft or Wayfair. Customer Equity is a concept popularized by Peter Fader and Daniel McCarthy through the literature published here. In our retail example, our customer base is worth about ~£8.1M.
2. Use CLV to Perform Customer Segmentation
Most marketing strategies involve a customer journey map, persona, or segmentation. While I’m a fan of putting myself in my customer’s shoes, these mappings are often very generalized or even fictitious. I’ve even heard of companies that create their personas based on biased and unvalidated information. Basically an echo chamber.
So what happens when Customer A doesn’t fit Jimmy’s “average” persona, the 30-year old millennial who over-indexes on avocado purchases? Well, the fact is, every human being is unique and have their purchasing cycles and consuming habits. This is known as Customer Heterogeneity, the idea that every customer is unique and should be treated as such. The other problem with customer personas is that they don’t tie into any internal metrics that can be tracked over time. Using CLV as the primary metric, we can segment our customers into different Recency and Frequency groups, indicating how valuable they are to us and their likelihood to churn. Depending on how often you update your models, you can track your customer’s segment status over time along with their CLV.
Now what do you do with these newly created segments?
3. Use CLV to Create a Customized Strategy for Each Segment
Using the example segments above, we can create customized strategies to increase the value of each segment.
For example:
- Our Top Customers segment won’t need a further financial incentive to keep spending with us. They are our best customers through thick and thin. So from a promotions perspective, they probably aren’t the best target group for a BOGO discount. To deliver superior value to this group, our company would need to develop a set of services that would differentiate us from our competitors. This could be anything from priority customer service, exclusive event invitations, accelerated delivery, or early product access. The goal is to offer premium services to this select group to reward them for their top business. Make them feel exclusive or however you want them to feel about your business/brand. At this point, a suggestion would be to incorporate market research to survey what features would be beneficial for this select group of elite customers.
- As for the other segments, such as Needs Attention, Churn Risk, and About to Churn, specific tactical offers and promotions will need to be created to re-engage those customers. A Churn Risk promotion might provide a more aggressive discount and have wording that creates urgency to generate a sale vs. a Needs Attention tactic might be informational. The tactic could provide a little nudge to the customer to purchase complementary products to their previous purchase. A recommender system is an excellent tool to surface related products to the customer. In either case, MarTech, such as Persado is terrific at using Artificial Intelligence to create engaging content for your campaigns.
If you’re not sure where to start, the above customer segments can be rolled up into the 2x2 framework below as a starting point.
Airlines and Hospitality industry players do this extremely well. Airlines will offer a loyalty program as a B2C offense tactic to all customers regardless of spend. The Aeroplans, MileagePlus, SkyMiles, Marriott Bonvoy. Their BEST customers receive tiered benefits that increase in value the more you spend.
From a B2B perspective, hotel chains like Hilton or Marriott have account managers dedicated to their top tier accounts like AMEX Travel, FlightCentre, Carlson Wagonlit Travel, Expedia. The true role of the SAM here isn’t to sell, but to problem solve, project manage, collaborate and co-create solutions.
My last point on Customer Segmentation is that tactics don’t have to be set at the segment-level. Another benefit of CLV is the ability to compute churn probabilities. In the example below, we can see a specific customer’s probability to churn over the purchasing lifetime. CRMs can be programmed to initiate a marketing campaign/customer journey when a customer reaches a certain probability. We come back to our point about Customer Heterogeneity and how each customer’s purchasing pattern is different.
Bonus: Incorporate CLV and Customer Referral Value (CRV) Together
CLV does not take into account CRV, Customer Referral Value, which is a total different computation. If you think about it, the value of a customer isn’t 100% limited to the products they purchase. The value of a customer can also be determined by the referrals they produce which then lead to purchases. Technically, CLV = LV + RV. More on this concept here.
References
- How Valuable is Word of Mouth, Harvard Business Review
- Lifetimes Python Package, Cameron Davidson-Pilon
- Persado AI
- The Customer Centricity Playbook, Peter Fader and Sarah Toms
- Theta Equity Partners, Daniel McCarthy and Peter Fader
