Double Down on the Most Valuable Customers
Leveraging data science to define your ideal customer profile (ICP)
In Seizing the Low-Hanging Fruit in Business with Data Science, I mentioned that companies would gain tremendous growth opportunities by cutting back on underperforming initiatives. Across industries, these failing initiatives start with the wrong audience. For example, if a gadget startup tries to win over conservative consumers, its campaigns will yield little returns.
Discovering the ideal customer profile (ICP) is often more challenging than anticipated. Having interviewed hundreds of companies, few could name their ICPs with enough details that sales and marketing teams can take action to attract visitors fitting ICP right away. Without a clear ICP, businesses unconsciously waste precious resources to chase impossible prospects.
In this article, I will talk about how to define ICP with the help of data science.
What is your Ideal Customer Profile (ICP)?
As Prof. Robert Kelley discussed in his book Critical Path Manifesto, all business activities should align with the critical path: generating more revenue or reducing costs for a company. As such, ICP describes the customers who bring in the most long-term value for a company.
Since every company has a different business model, its definition of customer long-term value varies. For simplicity, I’ll use the following formula to measure the value.
A customer’s long-term value = Revenue from the customer’s repeated purchase — its acquisition cost
With a proper data-tracking mechanism, you will likely see customer value distribution as below.
Visitors in the Negative-Value Cohort and Low-Value cohort outnumber the ones in the High-Value Cohort and Medium-Value Cohort. The split between the two groups generally follows the 80/20 rule. Only 20% or less of the visitors who have received ads or browsed a company’s website complete one or more purchases.
What does the 20% look like?
Finding out the 20% is every company’s dream; even those close to finding it can scale their businesses quickly. Let’s start small and analyze the High-Value Cohort to get a clear picture of the 20%.
Customer segmentation is the most common approach in this type of analysis. To illustrate this approach, I will continue with the gadget startup example mentioned in the beginning.
For over two years, the startup has tracked each visitor’s interactions with marketing campaigns, browsing history, purchase history, records of using coupons, etc. For each visitor, the startup collects 100+ attributes. With principal component analysis, clustering, and other algorithms, the High-Value Cohort can be further segmented into three smaller cohorts, and each cohort’s attribute distribution is shown below.
While algorithms can generate cohorts automatically, interpreting cohorts relies on business operators’ industry and product knowledge. In this example, they quickly conclude that cohort 1 likes to buy the latest model, cohort 2 prefers to buy with discounts, and cohort 3 often makes purchases as gifts.
How to attract the 20%
In addition to these labels, attributes like visitors’ interactions with marketing campaigns hint at how to attract them; their browsing histories reveal details about their preferences toward customer experience. With these insights, companies can tailor their approaches to attract each high-value cohort.
- Align product and messaging with customer preferences
Leveraging data, the gadget startup can personalize its marketing messaging for each product and each customer cohort. For example, historical data shows that cohort 1 reacts positively to campaigns about cool geeky features. On the other hand, cohort 2 engages more with campaigns spotlighting the practical benefits of using a new gadget.
- Present the product where customers show up
Cohorts 1 and 2 probably turn to different places to buy their electronics. For example, cohort 1 may trust YouTubers who talk about the latest and greatest tech products and make purchases after watching unboxing videos. Meanwhile, Cohort 2 likes to shop at Costco and buy gadgets at a discount. Therefore, partnering with tech influencers helps the company attract cohort 1 while working with wholesale stores speeds up selling to cohort 2.
Focus on the most valuable customers
You may have found that discovering and attracting each high-value sub-cohort takes substantial effort. However, acquiring customers is not the end of the story. Retaining these customers after acquisition also requires excellent service. Due to resource constraints, which every company has to deal with, a company’s service towards its valuable customers will be negatively impacted if it spreads resources across the board.
Do you still want to spend attention on the 80% cohorts?
In the following articles, I will continue diving into customer acquisition topics. If you want to chat about them in the meantime, feel free to contact me on Linkedin.
