|
Growth

Google for everyone…one segment at a time

Hey friends - after attending an intense week-long training course at Wharton, I have summarized my key takeaways in a series of posts. Subscribe to The Debrief to receive my learnings directly in your inbox.

Morning of Day 1 - We heard from Professor Eric Bradlow of The Wharton School on the importance of Customer Lifetime Value (CLV) in the field of Marketing

tl;dr

  1. Marketers should always optimize for Customer Lifetime Value (CLV) and not for Cost Per Acquisition (CPA)
  2. The key to customer segmentation is sacrifice (Yes, Google should be for everyone…just 1 segment at a time)
  3. STP - First segment your users, then target a specific segment, and finally position your product for the segment you decided to target

Optimizing for CLV

Some companies price their products based on their CPA (i.e. “it cost us $5 to make this product, let’s sell it for $6 to make 20% margin”). However, optimizing for a low CPA often correlates to low value and retention because if users are cheap to acquire for us, they’re also cheap for our competitors

💡
Professor Bradlow borrows an accounting acronym: FIFO = First in, First out

Breakeven Analysis

  • Time to breakeven needs to be as short as possible or there’s a very real consequence of the product being killed (even if it might succeed later on!)
  • CPA/CAC varies because of heterogeneity (every customer is different)

From a purely business standpoint, you shouldn’t focus all your efforts on the highest CLV customers since they will be buying your products anyways.

Imagine if Apple could stack rank every customer that walks through the door from highest to lowest CLV, they should NOT serve the highest CLV customers first because they’re already die-hard Apple fans!

💡
Therefore, we want to maximize the change in CLV, not CLV itself

Optimal Resource Allocation

If we look strictly at the CLV equation, we see there are 4 main contributors to CLV:

  • Revenue at a given time
  • (Recurring) Cost at a given time
  • Probability of acquiring the customer
  • Customer Acquisition Cost (which is also CLV at time = 0)

As marketers, we have limited marketing dollars to spend, so which area should we spend that additional $1 in marketing budget? Hence the need for optimal resource allocation

CLV and Acquisition

The primary reason to calculate CLV is NOT to minimize customer churn, but to optimize for customer acquisition. And if you don’t have workable data on hand to calculate CLV, make assumptions.

Obviously we’d like to acquire high CLV customers, but how do we calculate the CLV of these “unknown” customers?

First, we calculate the CLV of our existing customers, and identify the ones with the highest CLV. Then, we look for commonalities of those customers (demographics as an example) to “predict” what types of new users would turn out to be high CLV customers.

💡
Transactional data + Accurate demographics + CRM system = Holy Trifecta for Marketers

The key to customer segmentation is sacrifice

The concept of “versioning” means to put out different types of products and let the customers decide which “version” they want. Airlines do this perfectly: First Class, Business Class, Premium Economy…and Economy.

We know exactly what we need to do to drink limitless champagne and sleep on a king-size bed 10,000 miles in the air.

Huge companies like P&G and Unilever create “Flanker Brands” for the purpose of segmentation; they want to capture both the high-end market (high price, high quality) and the low-end (low price, low quality) without affecting their parent brand.

Segment, Target, Position

This is an extremely basic marketing concept, but one that is crucial to get right. I work on the Google Ads product at work so let’s apply this in an oversimplified way for my marketing campaigns -

Segment

We have (1) Large advertisers (2) Medium-sized advertisers (3) Small-business advertisers (segmented based on their quarterly spend)

Target

One of my marketing campaigns specifically targets a sub-segment of (3) Small-business advertisers who have the potentially to grow quickly (within 1 quarter) to (1) Large advertisers

Position

Although Google Ads is suitable for businesses of all sizes, the messaging of my marketing campaign positions the product - Google Ads - as an advertising platform that that relies on Machine Learning to help advertisers find high value users using a very limited data set (small advertisers may not have that much existing customer data).

This is another example of “the key to segmentation is sacrifice” since although Google Ads can be and is used by businesses of all shapes and sizes, my marketing campaign will be more effective if I target 1 specific segment.

This wraps up the key learnings from the morning session of Day 1, stay tuned for takeaways from the rest of the week!


Got feedback on this edition of The Debrief? Let me know!

Thanks for being a subscriber, and have a great one!