Growth Units

Growth Units

Author

Paul Orlando

Year
2020
image

Review

A short book that focuses entirely on the calculation of (and importance of) CAC (Customer Acquisition Cost) and LTV (Lifetime Value). This is a super simple introduction to the two most important growth metrics. If you can’t define or calculate CAC or LTV then this is a great place to start.

You Might Also Like:

image

Key Takeaways

The 20% that gave me 80% of the value.

  • Sustainable growth requires an understanding of unit economics, customer acquisition costs and lifetime value.
  • Acquisition Cost Ratio = Lifetime Value (LTV) : Customer Acquisition Cost (CAC)
  • CAC is incurred up front (leading) and LTV only happens over time (lagging). We understand CAC before we understand LTV - which can cause problems.
  • An Ideal customer: as defined by Steve Blank in 4 steps to the Epiphany
    1. Has a problem (that you can solve)
    2. Knows they have that problem
    3. Is looking for a solution
    4. Has hacked a solution to the problem already
    5. Can get budget to solve the problem
  • The vast mass of humanity does not have a problem that a specific business solves
  • CAC metrics include:
    • Cost to react, conversion rate, cycle time, payback period etc..
  • Unhelpful CAC = total marketing spend per period / number of new customers in period
    • 3 reasons this is unhelpful
      • People that didn’t convert in that period but are going to convert later are not included
      • It doesn’t tell us which marketing activities led to new customers
      • Doesn’t give us insight into what we can approve
  • CAC = cost to get a potential customer “in the door” / conversion rate to becoming a customer
  • Where possible break out CAC separately for each channel. This provides more insight into what channels currently work and which we might need to tweak or abandon.
    • Facebook Ads = $1.50 per click / 3% conversion rate = $50 (40% of acquisitions)
    • LinkedIn Ads = $2 per click / 5% conversion rate = $40 (25% of acquisitions)
    • Social Media Post = $0 per post / 6% conversion rate = $0 (35% of acquisitions)
  • Double down too hard on a marketing channel and you can saturate it - only customers who are less likely to convert will be left - driving up your CAC
  • A conversion funnel is a representation of the way potential customers gain awareness of your product, how they then sign up, start to pay, generate referrals, and more
  • 100% (0% loss)
    Acquisition
    Getting someone on site
    70% (30% loss)
    Activation
    Sharing their email or signing up
    55% (21% loss)
    Retention
    Keeping people engaged for some time
    20% (64% loss)
    Referral
    They tell others
    5% (75% loss)
    Revenue
    They become paid customers
  • Scaling: when something increases in size but also becomes more efficient
  • Scaling business are relatively rare. They often have network effects that make it easier for them to maintain customer acquisition and retention. They might also benefit from other advantages (e.g. funding, cash flow and market power advantages)
  • Relying on paid acquisition can be dangerous. Treat customer retention and referrals as a priority
  • Looking at single, static numbers for LTV, or for payback periods such as Months to Recover CAC, can produce misleading numbers.
  • Make your customer acquisition model:
    • Specific: by understanding difference by each channel
    • Predictive: by understanding how costs and conversions change over time (if they scale)
    • Fit with the business model
  • Lifetime Value is a measure of contribution margin a business earns from selling to a customer over time.
    • Unhelpful LTV calculation = Price of the product X Number of times the customer buys
  • LTV = (Price – Cost to produce) X Number of times the customer buys
    • Price per unit – Cost per unit gives us the Contribution Margin
    • Do not simplify the formula. Showing the component parts of contribution margin is valuable
  • A static measure of LTV is always going to be misleading. It’s more helpful to think of LTV as a flow of inputs and outputs. Chart it out over time
  • Assumptions in this model:
    • Customers start with a product’s limited free version, upgrade, and then are charged monthly while churning away.
    • Churn is kept constant at an additional 5% a month and not compounded
image
  • This shows when flows come in, when we breakeven, and what modifications could improve the situation. A single static LTV number doesn’t help as much.

LTV With Cohorts

  • LTV by Cohort and CAC by channel
  • Retention is an important part of LTV - it shows the loss of customers over time.
image
  • Customers are shown by the month they joined - what % remain is shown in the following months
  • It’s easier to see what’s going on if we shift everything left, so each cohorts relative month is shown in the same column
image
  • And now shown on a graph - it’s even more clear how much retention has been improving.
image
  • If you had distinct user types or customers (e.g. paid/free) - it might be worth duplicating the above tables for each type (to show different characteristics)
  • Tracking LTV by cohort can also help you predict what revenue or gross profit flows to expect.
  • Rule of thumb ratio for CAC : LTV of being 1 : 3 or 1 : 4
    • Maintains a buffer between spending on customer acquisition. Allowing for all the other costs in your business and some risk to LTV too
    • It also takes time to pay off the CAC. A higher ration allows for a longer payback period
    • Slow down your investment in growth if your ratio drops too low
  • Retention is a measure of how many customers or users stay.
  • Churn is a measure of how many many customers or users leave.
  • Retention = 1 - Churn
  • Churn = 1 - Retention
  • To calculate retention you need:
    • To agree what you are measuring retention of
      • e.g. member of a cohort
      • e.g. customer of a certain product
    • What qualifies as retained or churned.
      • e.g. monthly product usage?
      • e.g. repeat purchase frequency?
      • e.g. not renew their annual payment?

Negative Churn

  • If churn is just 1 - retention, can there be such a thing as negative churn?
  • Churning away some bad customers can be good news
    • LTV can increase (per channel or overall)
    • Referrals can increase
    • Improved efficiency (happy customers take less effort)
  • Some churn can be good news
  • Negative churn occurs when even after losing customers in a period the customers who remain end up spending more (perhaps by upgrading or buying more) and the business ends up with more revenue per customer than previously.
    • It’s churn as measured on revenues or gross profit, not individual customers.
    • Negative revenue churn is also known as positive net revenue retention
image

Deep Summary

Longer form notes, typically condensed, reworded and de-duplicated.

Introduction

  • Sustainable growth requires an understanding of unit economics, customer acquisition costs and lifetime value
  • Unit economics: marginal profit of one unit sold (collectively gross margin)
  • Don’t assume that a higher price will result in less demand / volume
  • Business goals can be different: E.g. to max unit sales, profit or market share
  • Key questions we want to be able to answer:
    • Given unit economics - are we doing enough volume to cover fixed costs?
    • Is our pricing/cost strategy allowing for retention and sustainable market share?
    • How much does it cost to acquire a new customer?
    • How many times do customers buy? Over what period?
  • Acquisition Cost Ratio = Lifetime Value (LTV) : Customer Acquisition Cost (CAC)
  • Calculating unit economics can be difficult
  • Estimating your competitors CAC and LTV - will help you understand their strengths and weaknesses
  • CAC and LTV will vary in precision - depending on what’s known about your customers
  • CAC is incurred up front (leading) and LTV only happens over time (lagging)
    • We understand CAC before we understand LTV - which can cause problems
  • To increase accuracy and insight we go into the next level of detail:
    • CAC by channel, LTV by user segment
  • It’s also worthwhile looking at payback periods

Customer Acquisition Cost (CAC)

  • Customer → pays a business for what they sell
  • Non-customer → don’t want or need what the business provides
  • Potential customers → want what’s provided, but don’t know it exists or haven’t converted yet
  • An Ideal customer: as defined by Steve Blank in 4 steps to the Epiphany
    1. Has a problem (that you can solve)
    2. Knows they have that problem
    3. Is looking for a solution
    4. Has hacked a solution to the problem already
    5. Can get budget to solve the problem
  • The vast mass of humanity does not have a problem that a specific business solves
  • Of those who do have the problem they often don’t realise they have it in the first place
  • CAC metrics include:
    • Cost to react, conversion rate, cycle time, payback period etc..
  • CAC can be made more specific by calculating it across different segments
  • CAC can be contextualised by connecting it to LTV.
  • CAC is the measure of the cost to start a relationship with a new customer

Unhelpful Ways CAC is calculated

  • Unhelpful CAC = total marketing spend per period / number of new customers in period
    • Quick and dirty calculation, but often misleading
  • 3 reasons this is unhelpful
    • People that didn’t convert in that period but are going to convert later are not included
    • It doesn’t tell us which marketing activities led to new customers
    • Doesn’t give us insight into what we can approve

Customer Acquisition Done Better

  • A better way to calculate CAC is by thinking on an individual customer basis.
  • Basic Components of CAC
    • The cost to get a potential customer ‘in the door’ (e.g: cost per click on paid ad)
    • The conversion rate of those potential customers that convert to actual customers
  • Better CAC Calculation
    • CAC = cost to get a potential customer “in the door” / conversion rate to becoming a customer
    • CAC = $1 / 5% = $20
  • This is the industry standard way to calculate and compare - allowing you to benchmark to other businesses

Customer Acquisition Cost By Channel

  • Channel examples: online ads, social media, conferences, print, word of mouth, and many more.
  • Where possible break out CAC separately for each channel. This provides more insight into what channels currently work and which we might need to tweak or abandon.
    • Facebook Ads = $1.50 per click / 3% conversion rate = $50 (40% of acquisitions)
    • LinkedIn Ads = $2 per click / 5% conversion rate = $40 (25% of acquisitions)
    • Social Media Post = $0 per post / 6% conversion rate = $0 (35% of acquisitions)
  • Double down too hard on a marketing channel and you can saturate it - only customers who are less likely to convert will be left - driving up your CAC
  • It’s not as simple as choosing the lowest cost channel - we should look at LTV per channel and payback period.

Conversion Funnels

  • A conversion funnel is a representation of the way potential customers gain awareness of your product, how they then sign up, start to pay, generate referrals, and more
  • 100% (0% loss)
    Acquisition
    Getting someone on site
    70% (30% loss)
    Activation
    Sharing their email or signing up
    55% (21% loss)
    Retention
    Keeping people engaged for some time
    20% (64% loss)
    Referral
    They tell others
    5% (75% loss)
    Revenue
    They become paid customers
  • Think of new potential customers becoming aware of your business and then flowing down the funnel. By the end of the funnel, only some are left.
    • You can show % remaining at each stage of the funnel, plus the loss from the previous stage (in parenthesis).
    • Showing the loss from the previous stage helps you target where you can have the biggest impact
  • In the Example above:
    • We may be casting too wide of a net and wasting time on people who are not our target segment.
  • To simplify things you can look at just Acquisition, Activation and Revenue
  • Look for an area to improve, make changes - and then re-asses where your effort is best spent.
  • Some channels will have less reach but better conversion.

LTV through the lens of a simple funnel

  • Assume a paid ad campaign costs $1 per click and it gets 1000 people into the top of the above funnel.
  • Let also assume that activation phase costs us nothing extra
  • Only 10% become paid customers
  • CAC = $1 / 10% = $10
  • If your product is $10/month, with costs of $5/month, so contribution margin is $5/month.
  • We breakeven on a paid customer after month 2 (assuming 0 churn)
    • Total number of paid customers = 1,000 x 10% = 100
    • Total spent on campaign = $1,000
    • Total earned after 5 months (assuming 0 churn) = $5/month x 5 months x 100 = $2,500.
    • That’s a 1 : 2.5 ratio of spending to earning and a total $1,500 difference

Growing and Scaling Customer Acquisition

  • Growing: when something increases in size
  • Scaling: when something increases in size but also becomes more efficient
    • Businesses that don't scale have variable costs that rise in line (or more) with growth
    • Businesses that do scale use their fixed costs as a way to serve more customers without a similar increase in variable costs
  • Scalability is a sign that the business becomes more efficient per customer as it grows.
  • Scaling business are relatively rare. They often have network effects that make it easier for them to maintain customer acquisition and retention. They might also benefit from other advantages (e.g. funding, cash flow and market power advantages)
    • They push their customer acquisition cost higher and go after less ideal customers. That hits CAC as lower conversion and as a higher cost to get somebody in the door
  • Relying on paid acquisition can be dangerous. Treat customer retention and referrals as a priority

What about when the relationship to the customer is less direct? (e.g. for SEO content)

  • Look at how many customers are needed for break even.
  • Break even # of customers = cost of asset creation / contribution margin of an additional customer

Scaling Customer Acquisition Channels

  • If our CAC to LTV ratio is good we might decide to spend more on customer acquisition, either in aggregate or on an individual interaction basis.
  • Different channels have slightly different scalability dynamics:
  • CAC increasing over time can be problematic - unless you can grow LTV
    • Early customers are acquired for less as they have the most pressing need.
    • Later customers may take more convincing or may have lower LTV

Customer Acquisition Cost by Cycle Time

  • Almost always marketing spend does not result in an immediate customer relationship.
  • The CAC formula can be misleading (cost to get someone “in the door” / conversion rate)
  • Fast Conversion Business
    Product is understandable Product is functional Product is an impulse buy Product is bought frequently at low price points
  • Some business have a big time lag before conversion, months or years.
    • Mitigate with reservation (Tesla Model 3 - $1000)
    • Extend your product line with licensing
  • A CAC of $50 but people spend $0 for the first seven months and then generate $10/month in profit margin after that, then in one year you break even on their acquisition.
  • Time to break even = CAC / time until total customer contribution margin equals CAC

Benchmarking

  • You can benchmark against those in a similar industry or those who sell to similar customer segments, but your results may be different (as there are many contributing variables)
    • Talk with people who see a lot of different companies and get their perspective
  • Looking at single, static numbers for LTV, or for payback periods such as Months to Recover CAC, can produce misleading numbers.
  • If increasing CAC drove an increase in LTV that might be good - it’s not as simple as minimising CAC

CAC Summary

  • Make your customer acquisition model:
    • Specific: by understanding difference by each channel
    • Predictive: by understanding how costs and conversions change over time (if they scale)
    • Fit with the business model

Lifetime Value (LTV)

  • Lifetime Value is a measure of contribution margin a business earns from selling to a customer over time.
  • LTV is a forecast - you need historical data to accurately predict it. It can’t be measured in advance.
  • LTV is made up of unit price, unit cost, and customer repeat purchases. Other factors like cycle time also come into consideration.
⚠️
LTV can be calculated in different ways - but still reported as LTV
  • As with CAC - there are common unhelpful calculations
    • Unhelpful LTV calculation = Price of the product X Number of times the customer buys
    • Unhelpful LTV = $1.95 X 10 = $19.95
    • This one doesn’t account for the cost of provision

Lifetime Value Done Better

  • LTV = (Price – Cost to produce) X Number of times the customer buys
    • Price per unit – Cost per unit gives us the Contribution Margin
    • Do not simplify the formula. Showing the component parts of contribution margin is valuable
  • The downside of this approach is that it’s static and it ignores timing

Lifetime Value Done Even Better

  • Price doesn’t have to be the same for every customer. Price can be dynamic. Price can depend on what customers receive. Price can change based on location, when they use it, how much they use it and more.
    • Apple used the price of the Edition watch to drive press
  • Cost represents the sum of inputs needed to produce a unit of the product. For some products costs are easy to conceptualise - in others they aren’t. Don’t include fixed costs.
  • Retention: a helpful metric for tracking repeat purchases and engagement. Some business don’t have predictable purchases from the same individuals (e.g. ] restaurants). You need long-term user retention and relatively unchanging products can sometimes measure retention accurately (a high degree of predictability).
    • If someone doesn’t pay - they aren’t a customer. Facebook’s customers are the advertisers, the people on the network are users
    • DAU (daily active users) is a user related retention number.
  • A static measure of LTV is always going to be misleading. It’s more helpful to think of LTV as a flow of inputs and outputs. Chart it out over time
  • Assumptions in this model:
    • Customers start with a product’s limited free version, upgrade, and then are charged monthly while churning away.
    • Churn is kept constant at an additional 5% a month and not compounded
image
  • This shows when flows come in, when we breakeven, and what modifications could improve the situation. A single static LTV number doesn’t help as much.

LTV With Cohorts

  • LTV by Cohort and CAC by channel
  • Retention is an important part of LTV - it shows the loss of customers over time.
image
  • Customers are shown by the month they joined - what % remain is shown in the following months
  • It’s easier to see what’s going on if we shift everything left, so each cohorts relative month is shown in the same column
image
  • And now shown on a graph - it’s even more clear how much retention has been improving.
image
  • If you had distinct user types or customers (e.g. paid/free) - it might be worth duplicating the above tables for each type (to show different characteristics)
  • Tracking LTV by cohort can also help you predict what revenue or gross profit flows to expect.
  • You can show how much yearly cohorts contribute to revenue:
    1. image
    2. 1 to 5 at the bottom of the bars in the chart are years. Cohorts 1 to 5 track customers that joined in different years. In the early years there’s negative revenue churn, in later years its positive.

LTV Is Misleading - it doesn’t help you…

  • Predict future changes to price, cost, and retention.
  • Predict entrance of new competition (can affect your prices or costs)
  • Predict whether an individual customer will be retained or when they will churn.
  • Predict whether churned customers will one day return
  • Predict what customers you can upsell on new products that you haven’t developed
  • Predict future business decisions that may prioritise market growth, leading you to accept a lower LTV to CAC ratio

About CAC to LTV Ratios

  • Rule of thumb ratio for CAC : LTV of being 1 : 3 or 1 : 4
    • Maintains a buffer between spending on customer acquisition. Allowing for all the other costs in your business and some risk to LTV too
    • It also takes time to pay off the CAC. A higher ration allows for a longer payback period
    • Slow down your investment in growth if your ratio drops too low
  • When to ignore the rule of thumb:
    • When you are focused on growing market share.
    • When your payback period makes it difficult.
    • When there are strategic considerations.

Retention, Churn, and Ways to Calculate Them

  • Retention is a measure of how many customers or users stay.
  • Churn is a measure of how many many customers or users leave.
  • Retention = 1 - Churn
  • Churn = 1 - Retention
  • To calculate retention you need:
    • To agree what you are measuring retention of
      • e.g. member of a cohort
      • e.g. customer of a certain product
    • What qualifies as retained or churned.
      • e.g. monthly product usage?
      • e.g. repeat purchase frequency?
      • e.g. not renew their annual payment?
  • Typical standards:
    • For a subscription service billing monthly - retention is measured monthly
    • For other fast-changing and high churn services (e.g. mobile games) daily or weekly is more common
  • Use cohorts to measure retention:
    • What happens after the first month? What % remain as customers?
  • With enough data you can make predictions…
    • You can assume you’ll lose a fixed number, or a fixed % of users a month

Retention Behaviours

  • Estimating retention becomes easier over time, as you learn more about your competition, your customers and their behaviour
  • Three different types of retention curves:
    • Constant. Churn is constant month to month.
    • Annual. This type occurs when customers need to make annual decisions to continue as a customer. That could be the requirement to make an annual payment.
    • Cliffs. High churn in the first couple of periods, then low churn afterward. Those who stay are a good fit
  • You won’t be able to predict retention well until you have enough historical data
  • Issues that can cause variance in retention:
    • You haven’t reached statistical significance.
    • You have changed something substantial.
    • Different renewal patterns (e.g. for subscription products)
    • You generate attention among a large group of non-customers (e.g. Major press)

What impacts retention?

  • ⬆️  Better understanding of the target customer, provisioning a better solution for them and finding them where they are
  • ⬇️  Media coverage will lead to target and off-target customers being acquired
  • ↕️  Changes to price
  • ↕️  Changes to customer comms
  • ↕️  Changes to product
  • ↕️  Public opinions
  • ↕️  Entrance of competitors, changes in fashion, changes in demand, changes in economy

Off-target customers can cause interesting affects to metrics and prospects.

Negative Churn

  • If churn is just 1 - retention, can there be such a thing as negative churn?
  • Churning away some bad customers can be good news
    • LTV can increase (per channel or overall)
    • Referrals can increase
    • Improved efficiency (happy customers take less effort)
  • Some churn can be good news
  • Negative churn occurs when even after losing customers in a period the customers who remain end up spending more (perhaps by upgrading or buying more) and the business ends up with more revenue per customer than previously.
    • It’s churn as measured on revenues or gross profit, not individual customers.
    • Negative revenue churn is also known as positive net revenue retention

Matching LTV and CAC

  • Time to Recover CAC:
    • CAC / (Average revenue per customer per month - Cost to serve customer per month)  
  • This simple calculation is per retained customer. Even a small amount of retention would push back the payback period.
image

Using Discount Rates

  • A discount rate is an assumption about the future value of payments. We take the value of getting paid $1 in the future to be less than getting paid today, for a couple reasons:
    • The Time Value of Money
    • Uncertainty Risk
  • If you’re working at a high risk startup - discount rates might be less important than at a retailer that is lean with it’s capital
  • Discount rates may be below 10% for a low-risk public company, or much higher for riskier companies
Future Value Discount Formula
  • You can model inflation too if you think it’s going to be important.

Trying to improve CAC and LTV

  • CAC
    • Lower the cost to bring someone “in the door”
    • Increase conversion rate (focusing on those that are a good fit)
    • Track and understand the differences in CAC across different channels
    • Track changes across time
    • Understand the CAC payback period for different segments and at difference prices.
    • Understand the connection between customers acquired through different channels and their eventual LT
  • LTV
    • Improve pricing
    • Improve margins (costs
    • Improve retention
    • Show it over time and CAC payback
  • Overall
    • Understand the value of segmenting vs averaging customers
    • Appreciate outliers as opportunities to gain insight.
    • Graph your data to help you notice other changes.