One of the Categories for Business Model Analytics is Behavioural.
It fits into the thesis around CLV (Customer Lifetime Value), whereby the behaviours of the most loyal and highest-spending customers are identified and extrapolated into strategies and actions that optimize the business model.
Customer Analytics – Behavioral
- Power Laws and The Pareto Principle
- Customer Analytics – Behavioural
- Behavioural Analytics – Higher AOV, More Repeat Buyers
- Business Model Impact
- More Business Model Analytics Posts
- Case Studies Cited
Power Laws and The Pareto Principle
The Pareto principle is a mathematical precept (not a formula) that states that ~80% of the consequences (outputs) come from 20% of the causes (inputs). A general example of the Pareto principle in relation to the business model is that 80% of the profits come from 20% of customers.
While by no means an exact science, understanding this precept through the lens of Behavioural Analytics can help a company optimize its business model and shape strategies around the customers who derive the highest amount of Customer Value.
Where does the Pareto principle come from? The Pareto principle was developed by Italian economist Vilfredo Pareto in 1896. Pareto observed that 80% of the land in Italy was owned by only 20% of the population. He also witnessed this happening with plants in his garden—20% of his plants were bearing 80% of the fruit. This relationship is best mathematically described as a power law distribution between two quantities, in which a change in one quantity results in a relevant change into the other.Asana
The Pareto principle is a Power Law distribution based on observed relationships in many different facets of life – from nature, to finance, to sports, etc.
Anecdotal observations around this Power Law help create new ways of thinking about the best way to optimize performance. As such, the idea is not to apply it rigorously and formulaically, but rather to observe certain customer behaviours as a bell curve (ie. distribution) and look for patterns.
Power Law Distribution vs. Normal Distribution
In the case of the Farfetch business model, for example, a large % of their GMV (Gross Merchandise Value) was driven by a small handful of retailers in the early years.
But 22% of the company’s GMV is from its top ten retailers, even though it had over 614 on the platform, showing some centralization and power law-like dynamics.LooseThreads
This example exemplifies how the thought process behind Power Laws work. Not that 10 retailers (only 1.6% of retailers on the platform at the time) drive 80% of GMV, but that a notable Power Law exists on Farfetch’s platform as it pertains to the Pareto principle.
Extrapolating this in relation to theories behind Business Models and Behavioural Analytics, we can see that much of the data that relates to revenue, margin and profitability are averages across the whole business. Statistics helps to break down averages into more granular and actionable data.
Specifically, if we look at ARPU (Average Revenue Per User), CAC (Customer Acquisition Cost), and LTV (Lifetime Value), then we can see that Average Revenue Per User is typically an average of all customers. This average feeds into the ultimate LTV calculation.
In most cases, however, there will be certain Cohorts or segments of customers who drive large % of the profitability across the whole segment of customers. These customers will likely have very specific behaviours/attributes that can be measured and tracked in order to better understand how to acquire more of these types of customers and optimize the business model.
Customer Analytics – Behavioural
The problem is that most businesses do not have a simple and sustainable way to measure Customer Value beyond headline metrics such as revenue and margin.Measuring Customer Value
Thinking about typical metrics of non-contractual businesses such as ARPU (Average Revenue Per User), AOV (Average Order Value) and others helps us gain surface insights into per-customer revenue statistics.
These are typically highlighted against the backdrop of total revenue, margin, and profitability (or loss). But if those numbers are trending in the wrong direction – as is happening with many companies at the current moment – then the typical solutions normally revolve around two levers: generating more revenue or lowering costs.
Without more granular Business Model Analytics, however, these types of generic strategies can often lead to actions such as raising prices, cutting support, or outsourcing, which can often end up alienating the core customer base who drive the profitability behind the company.
Customer Analytics Visual
To look at the problem holistically can be overwhelming and require analysis through a seemingly endless sea of data, which is why we like to think about looking at 1 core Analytic/Metric from each of the 3 Categories listed above – Behavioural Analytics, Retention Analytics, Channel Analytics.
From Retention Analytics, we know that building a large Cohort of Repeat Customers is essential for becoming profitable in the long-term, and that many of those top customers often come by referral.
Through Channel Analytics we know that there are many different marketing channels that can be used to target these customers, and that optimizing for a low CAC is not as simple as finding the lowest cost marketing channel.
The strategy needs to be linked to optimizing LTV and ensuring customer loyalty remains at the core of the strategy.
But the questions from a Behavioural Analytics perspective become:
- who are these Repeat Customers? (profile, geography, income characteristics, etc.)
- what do they like/not like in relation to the product/service? (as determined by actions like NPS or other surveys)
If theoretically, 20% of the top-performing customers could be linked towards some kind of underlying behavioral analytics, this could then, in turn, be used to drive the top-quartile of the distribution and increase AOV without having to rely on traditional blanket strategiesBusiness Model Analytics
Many businesses will have this type of data in one form or another, but it is often fragmented and uncontextualized relative to better understanding and quantifying Customer Value.
Behavioural Analytics – Higher AOV, More Repeat Buyers
The Pareto principle states that 20% of consequences (inputs) can lead to 80% of the causes (inputs).
Looking at Farfetch, for example, we can see that they had an AOV (Average Order Value) that trended up from ~$600 to ~$750 over time, with many customers ordering several times per year.
Contextualized against Farfetch’s core business model, where they earn a 30% Take Rate, the following breakdown is roughly how the model functions:
- 50 – 55% Gross Margin on platform revenue
- 20% of Digital Platform Revenue for Customer Acquisition
- 30 – 35% Contribution Margin
- LTV that has trended upwards over time, towards more than 4X by some estimates
Management has alluded to much higher Contribution margins for older customer cohorts (50-60%).Partnership Investing
But we know that Farfetch – who had 55% Repeat Customers when they filed their F1 in 2018 – can generate much more profitability from Repeat Customers.
Flourishers Case Study
When looking at the Flourishers Case Study in the Business Model Analytics post, we could see that when a more granular study was done on the core customers who drove the majority of profitability, it yielded many insights that could be shaped into Behavioural Analytics. For example:
- “have a high lifetime value, spending an average of $468 a year in the category, versus $235 for other customers”
- “Fully 46% of Flourishers shop key fashion categories at least monthly, versus 21% of all shoppers”
- “Flourishers are 1.4 times as likely as other customers to recommend retailers to their friends and family members”
- “They are 2.3X as likely as other customers to say they are “willing to pay more for the best fashion products”
Presumably, Farfetch would have similar levels of data about its Top 20% of customers from within the Cohorts of Repeat Customers. A deeper analysis of that data would likely yield insights – similar to the Flourishers above – that could be used to identify strategies to acquire and retain more of those customers.
When factoring in the combinations of behaviours – highers AOV, multiple repeat purchases, likelihood of recommendations/referrals, etc – it is clear these customers can produce effects on the business model that are consistent with a Power Law distribution.
Business Model Impact
Many non-contractual businesses are facing an uphill battle in the current market right now. Growth has flat lined for many businesses, consumers are dealing with issues related to inflation, fears of a recession, and other factors depending on the market.
Historically, these types of downward cycles – while devastating for many businesses – can open once-in-a-generation type opportunities for other businesses.
Understanding Power Law distributions and Behavioural Analytics in relation to a customer base may sounds tedious and time consuming, but the ability to use this knowledge/data to optimize ARPU (Average Revenue Per User), CAC (Customer Acquisition Cost), LTV (Lifetime Value) and Churn (more Repeat Customers) should not be underestimated.
The Goal – Target High LTV Customers
To develop strong Cohorts of Repeat Customers – thereby mitigating the need to constantly acquire New Customers at scale – requires a deeper view as to what behaviours drive and motivate these customers towards endorsing a certain product/service.
But the core concepts around Pareto principle help to move away from simple averages and headlines metrics in order to dig deep and target segments that will drive the most profitability in the mid to long term.
Overall, Behavioural Analytics can help zone in on the behaviours and motivations of the most profitable customer Cohorts for a given business. In combinations with other Analytics, this can help optimize customer segments for LTV and the business model for long-term profitability.