LTV Fundamentals


Customer Lifetime Value (CLTV, or just LTV) is one of the fundamental means by which a VC or investor can understand a Startup, and one of the best means of understanding how and when a Startup should grow. It is particularly useful when compared to Customer Acquisition Cost (CAC) in the ratio of CAC:LTV.

The concept of LTV is relatively simple, but the implementation can be complex and dependent upon several factors. Most articles on this subject to date have used relatively simplistic measures and explanations of LTV to calculate it, but this a disservice to Startups and can lead to misunderstanding or misrepresentation

This article talks about the fundamentals of LTV; starting with what it is attempting to achieve as a measure; the components of that measure; and ending in an overview of the various means of calculating LTV. It will be useful to reference "A Lighthearted Intro to CAC-LTV" if you are not familiar with the concept of LTV.

LTV Fundamentals

LTV attempts to estimate the likely value that a Startup will derive from a Customer, over the likely lifetime of an average Customer.

The following points are important to bear in mind when calculating LTV: -

  • LTV is a prediction of the future value...
  • ...of an average Customer...
  • ...which is an estimate...
  • ...based upon Customers' historical activities

In this regard, LTV is both a forward-looking and backward-looking metric. It is attempting to predict the future activities and value of a Customer, but it does so by looking at historical activities and value and using these to estimate and predict future trends.

LTV Concepts

The core concept of LTV is that a Customer's activities will benefit the Startup in some way for the duration of that Customer's lifetime. These activities generate value which can be captured in some way by the Startup. This value generation may be direct or indirect, depending upon the business model, for example: -

  • Customer Purchases - Customers may directly purchase items and the Startup receives a Profit on that purchase
  • Recurring Subscriptions - Customers may sign up to use a product on a recurring subscription basis, and the Startup benefits from the Profit on this subscription
  • Periodic Transactions - Customers may periodically transact on a platform (example, rides on Uber) and the Startup derives value from these transactions
  • Eyeballs - Startup may not directly transact with the Customers, but those Customers can be monetised in more indirect means, such as via ad-revenue

These activities may be one-time (example, the sale of an item on a webstore that is unlikely to occur again), or recurring on a periodic or aperiodic bases into the future to a given point in time. The core aim of LTV is to estimate the cumulative future value of those recurring value-deriving activities to the Startup

Components of LTV

LTV is concerned with the value derived by a Startup from the activities or interactions of one Customer, and so it is concerned with Profit rather than Revenue. For example, if a Startup sells a widget for $100, which costs the Startup $85, then the value derived by the Startup is the Gross Profit of $15.

The Profit versus Revenue component of LTV is complicated by two key factors, namely

  • Approximating Gross Margin to 100% - Most software Startups have virtually zero CoGS, resulting in a Gross Margin of 85+%. This is often simplified to 100%, allowing the LTV calculation to be performed on Revenues as a proxy for Profit
  • Non-Accounting Definition of Gross Margin - The Gross Margin measure in LTV does not comply with the strict accounting definition of gross margin

This can add significant nuance and complexity to calculating LTV - which in turn can confuse Founders when pitching VCs, or worse, can lead to mistrust and misunderstanding. A good guiding principle here is to think clearly about what $-value the Startup receives after it has done everything to service that Customer and transaction.

What $-value does the Startup receive after it has done everything to service that Customer and transaction?

Specifically, this means that the Gross Margin component of LTV should factor in the following costs: -

  • Cost of Good sold
  • Software hosting costs
  • Customer Support / Success, and tools / products / costs associated with that (e.g. Intercom, CS salaries, overhead)
  • Account Managers, personnel or other costs directly associated with servicing a Customer's needs
  • Cost of tools used to support, guide or educate the Customer (e.g. FAQ, blogs, training, tutorials)
  • Future improvements to the product, offering or service that impact Gross Margin over a meaningful part of the lifetime of the Customer

However, the following costs should not be factored into the Gross Margin when calculating LTV: -

  • Cost of Sales, Marketing, sales salaries, CRM etc. (these are all in CAC)
  • Software development or R&D costs
  • One-time or non-recurring costs, for example, promotions, discounts or giveaways (components of CAC)
  • Excess costs due to partial utilization. For example, the unused cost of a Customer Support operative who currently supports 20 Customers because the young Startup has few Customers, but will be able to support 100 Customers at scale

Methods of Calculating LTV

There are many nuanced and different methods that can be used to calculate LTV, and each method may be more or less suitable depending upon the type of company and business model. In general, I recommend the following three methods for calculating LTV (with a zeroth addition for completeness): -

  1. Estimated of Customer life x periodic Revenue
  2. LTV Equation
  3. Cumulative Revenue Retention
  4. Discounted Cashflow

Each of these methods requires deeper discussion, but are summarised below: -

0. Estimated Customer life x periodic Revenue

Estimate the likely lifetime of a Customer, multiply that by the periodic revenue, then multiply by the Gross Margin

Example, a Startup believes that their Customers will be active for 3 years on average, spend $100 per month, and their Gross Margin is 85%. LTV is consequently 36mths x $100/mth x 85% = $3060

This method is very quick and dirty, but fundamentally falls apart because the likely lifetime of a Customer must be estimated. If it is possible to measure the likely lifetime of a Customer, then the following methods are inherently superior.

1. LTV Equation

Gross Margin x Periodic Customer Revenue / Periodic Customer Churn

Example, an average Customer spends $100 per month on a subscription package, monthly Customer churn is 3%, and the Gross Margin is 85%. LTV is consequently 85% x $100/mth / 3% = $2833

The LTV Equation (or LTV formula) is generally applicable to all Startups, is quick to estimate and gives a clear measure of LTV. It is also effectively the mature, "big-brother" of the method above, because 1 / periodic churn = likely Customer life. Churn can often be clearly measured and quantified from historical data, resulting in a more justifiable calculation.

However, this method makes a number of simplifying assumptions and becomes distorted, innaccurate, or incorrect in the following situations: -

  • Negligible, zero or negative Customer Churn
  • Revenue expansion or contraction that is distinctly different to Customer Churn
  • Aperiodic interactions
  • Varying Revenues
  • Churn that cannot be approximated by exponential decay - i.e. it implicitly assumes that the product lacks Product-Market Fit

2. Cumulative Revenue Retention

Sum average Revenue per average Customer over time since first activity, extrapolate to the point at which it flatlines, multiply by Gross Margin

This is the most complex and most illustrative means of calculating LTV but requires significant calculation via Excel / database / algorithmic implementation. However, the results are incredibly illuminating and illustrative (see Jonathan Hsu's description) and can help prove the following: -

  • Significant network effects
  • Meaningful expansion Revenue
  • Strong economies of scale

This method is most applicable to newer business models, as it is inherently able to deal with infrequent Revenues of varying sizes, which cannot be easily accommodated for with the LTV Equation. This makes it particularly applicable to Startups with the following features: -

  • Varying frequency of interactions, sales or Revenues - e.g. Uber riders taking trips on an infrequent basis
  • Varying Revenue amounts - e.g. varying sized orders on JustEat
  • Revenue expansion / contraction - e.g. additional seats and add-ons purchased for HubSpot

We built because of the complexity of performing this calculation, and the realisation of what it can uncover and prove for Startups - please contact us if you want to learn more.

3. Discounted Cashflow

Find public comparables Beta, derive WACC for your future likely debt-equity ratio, use this as the discount rate in a DCF model of per-Customer Revenues

Don't worry if that sentence doesn't make sense - this method of calculating LTV is applicable to perhaps 0.01% of Startups. This method should be used when a Customer's lifetime on a platform is significant, for example, 10+ years, which means that the following factors must be considered: -

  • Time-Value of Money - $1 received today is more valuable than $1 received 10 years from now, so we need to use a discount rate in order to perform an NPV calculation
  • Discount Rate - the discount rate to be applied should reflect the average discount rate over the lifetime of the Customer (i.e. 10+ years), which means it needs to account for both the future risk profile and future capital structure of the Startup as a mature company
  • WACC - in order to account for the future risk profile and future capital structure, we need to calculate the WACC for that future mature Startup
  • Industry Beta - the expected return on equity in the WACC calculation requires that we consider industry comparables for the future mature Startup

Clearly this is a complex and cumbersome calculation, with some inherent assumptions that can massively inflate LTV. For example, this method generated an LTV for one consulting client that was almost 10x larger than that calculated via the LTV formula above. Hence this method must be understood clearly, implemented when the reason is clear and compelling, and interpreted with a clear understanding of the reasoning and methodology above

Which Method to Use?

In general, the LTV Equation gives a meaningful and simple means to estimate LTV, and I recommend that it be used for almost all Startups as a first pass. However, it is unable to deal with a variety of relatively simple complicating factors, and so becomes highly inaccurate when analysing modern Startups.

Despite the inherent complexity of calculation, the Cumulative Revenue Retention method is both universally applicable, and by far the most illustrative means of calculating LTV. For this reason, I recommend it be used for all Startups.

Cumulative Revenue Retention method is both universally applicable, and by far the most illustrative means of calculating LTV


LTV is one of the fundamental means by which an Investor or VC will understand a company, and it is also one of the best indicators of when or how a Startup should be scaling growth. However, it is not a simple or easily calculated measure, and the most suitable method to use is dependent upon the type of company and business model being analysed.

There are several implicit assumptions and complex components to LTV, and an accurate calculation requires an understanding of these in order to apply it correctly to a particular case. There are also several different methods for calculating LTV, each of which presents various advantages and disadvantages, and is variously applicable to different business models.

This has been an overview of these methods, and we will dig into each method further in subsequent blog posts.

Posted by Phillip Gales

Phillip is a serial entrepreneur who specialises in Operations, Data and Metrics. He applies AI and Machine Intelligence to old, antiquated and/or forgotten industries that are ripe for disruption.

Phillip holds an MBA from Harvard Business School, and an MEng in Electrical Engineering from the University of Cambridge, specialising in Machine Intelligence.