One of my main intellectual strategies is to explore important neglected topics where I can find an original angle to pursue.
As a result, I tend to lose interest in topics as they get more attention.
This is why I’ve avoided customer lifetime value.
Sure, it is plausibly important, but there are millions of articles written about it online, and I haven’t found a non-boring angle on it yet where I can explain our methodology without it getting lost in the noise.
So, I’ve let it lie.
But on the recommendation of one of my colleagues, I decided to publish a small sample of our secret sauce for building a Lifetime Value (LTV) model, also referred to as the Customer Lifetime Value (CLV) model or formula.
You can check that out here.
Today, I will expand on what I said in that article and go into a bit more detail about what lifetime value is, how to build a model, and discuss some common unforced errors people make in their calculations.
That said, if you landed here off of a Google search for “Lifetime Value” I’m sure you are probably wondering “what the f*ck is lifetime value and how do you use it?”
The concept of lifetime value is not new, but the nerds have yet to agree on a common definition (more on this later).
A lot of analysts refer to it as “the net present value of the profit stream of a customer.”
“The present value of the expected sum of discounted cash flows of an individual customer or cohort/segment of customers.”
This concept (which to the untrained eye may appear to be quite benign) is typically used to compare the costs of acquiring a customer, commonly called CAC (but also known as Subscriber Acquisition Costs or SAC) with the discounted positive cash flows that will come from that customer over time.
Here’s a simplified version of the formula:
The key variables are as follows:
- ARPU (average revenue per user)
- Avg. Cust. Lifetime, n (This is the inverse of the churn, n=1/[annual churn])
- WACC (weighted average cost of capital)
- Costs (annual costs to support the user in a given period)
- SAC (subscriber acquisition costs, sometimes referred to as CAC = customer acquisition costs)
LTV is critical for understanding the long-term value a customer brings, as opposed to a singular transactional value.
The formula, when used correctly, can be a good tactical tool for monitoring and comparing like-minded variable market programs, especially across channels.
Note: a lot of companies use “lifetime value” to let their shareholders know they can squeeze more revenue out of a customer than it costs to purchase that customer — and this is becoming a more common theme as customer acquisition costs rise across top sectors.
Note: In the context of business and marketing analytics, Customer Lifetime Value (CLV) and Lifetime Value (LTV) are two acronyms that essentially mean the same thing, but their usage can sometimes vary slightly in focus. CLV is often used in contexts where the focus is more on customer relationship management and/or customer experience and places a lot more emphasis on the customer aspect. This is a subtle difference, and we will not be getting that granular. For the purposes of this paper, we will use LTV to define the concept.
As I mentioned earlier, “lifetime value” is not a new concept, but it is rapidly becoming a new (annoying) buzzword you’re starting to hear everywhere as a lot of companies struggle with high costs and slow user growth; and the corresponding struggle to run the technical analysis required to solve for low demand and/or suboptimally priced/positioned products.
It doesn’t help that critical macroeconomic variables such as rising prices (inflation), interest rates, unemployment, and record consumer debt are squeezing disposable incomes, forcing many customers to ditch the products (very often subpar products) that they don’t actually need.
These lost customers are incredibly difficult (and expensive) to replace, and the cost to purchase a customer can rise 10-20% year-over-year (YoY) or more in many industries.
But let’s not turn this paper into a boring analysis of generic, blended advertising metrics.
Instead, let us assume we are working in a direct response environment and the bulk of the money spent is on digital advertising.
This medium, after all, makes the most sense for “lifetime value” in a modern technology sense because it is highly measurable, and attribution has a stronger quantifiable aspect than say physical advertising like billboards or cable TV.
But the competitive environment in digital advertising is in a state of chaos.
Google, Meta, and Amazon have the market in a chokehold as they have essentially monopolized digital ad spend.
If you are not using these platforms, you will very likely struggle to get customers, putting your business at a serious competitive disadvantage.
So, everyone is now forced to use these platforms and compete for the same pool of potential customers.
To the untrained eye this may not seem like a big deal — but the first rule of economics is scarcity — and a defining aspect of bad businesses is to always forget the first rule.
You see, digital customers have a finite number (regardless of the platform) so customer acquisition costs (SAC in our formula from above) are constantly rising as competitors fight to outbid each other for the best users to see their advertisements.
This is good for Google, Meta, and Amazon and very bad for you, the business, because this puts a tremendous amount of pressure on SAC.
You don’t need to be an economics scholar to know that a limited supply + growing demand = rising prices.
Rising prices can mean increasing revenues and good times for the ad platforms, and a cashflow disaster + bad times for you.
To make things worse, consumer protection measures have dealt a damaging blow to the advertisers, making an already bad situation much worse, because without having visibility into who your customer is has made it a lot more difficult to track/segment/measure campaigns.
So, now you have rising customer acquisition costs, and you are also unable to measure and/or scale your customer acquisition — and it’s all happening in a recessionary market environment.
Note: It’s no secret that marketing teams (especially the execs) love big budgets. The bigger the better. This makes it easier to grow the top line. Also, perfectly correlated, subpar managers love to use LTV to rationalize irrational spending. In today’s wild west of SPACs and overvalued IPOs, there is a tendency to put heavy emphasis on LTV as if it was some sort of magic metric. Quite often, you’ll find that many of these companies eventually experience massive losses as they scale (often through the IPO). This is a classic move from the overvaluation playbook, where a company tries to use LTV to “relax” the need for near term profitability and “justify” the ability to play it forward — so they spend today for benefits that are pushed into the future to be recouped later. In other words, “we will spend a lot on CAC today for a massive LTV payoff tomorrow. If your assumptions are wrong, those benefits will never come, leading to unsustainable economics and, often, insolvency.
Again, you don’t need to be an economics genius to realize that in a recession, customer acquisition efficiency is even more important than ever.
In this type of environment, many brands will hit a plateau, and eventually see their top line metrics take a nosedive, as clickthrough rate (CTR), cost per click (CPC), etc. all begin to put stress on conversion rate and SAC.
There are a lot of businesses out there right now that are starting to find themselves in this situation, and this phenomenon is even further strained by negative economic forces that are out of the company’s control.
So, it seems, the old playbook of cooking up some generic ads on Meta/Instagram, and then sprinkling in some Google Ads and “SEO” is no longer a viable (or sustainable) solution — plus most firms do not have the time preference to sit around and wait for their “SEO” optimizations to finally kick in.
The digital environment is very complex, and it’s evolving at a rapid pace — the old playbook is very difficult (if not impossible) to measure — and companies are spending more and starting to get less and less for each dollar spent.
What we often see (especially when analyzing these types of campaigns) is that a lot of companies will experience what they feel is “growth”, but their customer acquisition cost has probably 2x’d or 3x’d (or more) from baseline.
In situations like this, if your lifetime value is stuck at the point where you started (for example, SAC has increased from $200 to $400 and your LTV is only $300, you’re probably going to be in deep trouble).
Many such cases.
Note: the above case is just a fictitious example, but if you were to ask around, the general consensus from most formula bros is that a “good” LTV to CAC ratio is 3 to 1. But, when you look at survival rates (or the behavior of your unpredictable customers) across many industries, you’ll see that they are all very different, and that there is a lot of fluctuation. So, not only are general benchmarks like these lazy (LTV to CAC is theoretical and unrealistic) but they are also super unhelpful. Every industry/business is different and you will need to run your own analysis to see what works for you.
That said, given what’s happening with customer acquisition costs in the market right now, putting a significant effort into LTV analysis is paramount.
LTV can be very tricky, but there are a few companies that make it work.
The companies that get it right are able to:
- Allocate marketing budgets more efficiently
- Reduce wasteful spending on low-value customer segments
- Improve high-value customer targeting/personalization
- Optimize pricing/product development based on customer value perceptions
- Improve customer retention and loyalty
- Improve proforma and revenue forecasting
- Align business strategy with most profitable segments for max ROI
- Achieve profitability
The companies that get it wrong often end up in suboptimal cashflow situations and/or insolvency.
This is a high stakes game.
It’s not for the faint of heart.
Your ass is on the line.
Your job is on the line.
Now, if you’re new to this stuff you may have looked at the lifetime value formula and felt a little intimidated.
Perhaps you were thinking “oh sh*t, this is like math, I’m totally f*cked.”
Yes, there is a formula (don’t worry it’s a relatively basic formula) but the most important thing to keep in mind is that you can’t just plug in numbers to get this right — the key is in your assumptions.
This may be a challenging concept to fully grasp at first, but just like any model, getting it “right” is 100% dependent on the assumptions used in that model.
If your assumptions are wrong, you’re going to get the entire thing wrong.
And that brings me to my next point: a lot of “lifetime value” nerds don’t actually know what they’re doing, they take shortcuts and use voodoo math, which provides a greater degree of freedom to make their numbers look better than reality.
These folks often end up using the model incorrectly, and ultimately, abusing it.
But an even more dangerous group than the incompetent nerds are the LTV zealots — the folks that worship the formula and believe it is actual science.
If your company has been infiltrated and overrun with LTV zealots, the formula will slowly begin to take on more importance than it should, management eventually becomes obsessed with it, and the whole thing usually ends up devolving into something useless, and in many cases, destructive.
With that said, I don’t want you to get the wrong idea here, I think LTV is an incredibly useful tool (we build our own models that work quite well) but these models only “work” when used correctly and when placed in the right hands.
Most companies, however, in their LTV obsession and mathematical delusion, tend to use the formula incorrectly, and they put the wrong people in charge of it, which leads to bias, blinders, and bad outcomes.
Note: depending on how your company is structured, it can be tricky to track down who actually “owns” the LTV calculation. For example, let’s say you want to check on and/or make adjustments to your LTV to CAC ratio. Who is in charge of it? Who do you talk to? Do you contact the head of customer experience who owns churn? Do you contact the marketing department who is spending all your money on customer acquisition? Do you contact finance and/or sales who owns ARR? Many times, there is no clear owner, and when you cannot determine who owns a metric, the metric often becomes meaningless. But often, what you’ll find is that usually the same department that is always arguing for more spending is the one that typically “owns” the LTV calculation.
When the obsession reaches apogee, these managers often slip into a common delusion where they begin to confuse the model with reality and begin developing a voodoo-math-based superiority complex where they become fixated on the dogmatic execution of the formula, losing sight of the more important elements of business strategy.
In such cases, the formula can be confused, abused, and misused, much to the detriment of the business (and sometimes the customer as well).
As a prime example of said misuse, you will hear the common argument that “as long as the sum of the discounted future cash flows are significantly higher than the SAC” then “we need to push the accelerator.”
When you translate this from Office Speak into normal English it basically just means: “we need to burn capital by over-aggressively spending on marketing.”
A common [smoov.bra.in2023] reasoning error.
The companies that do this often get into serious trouble.
They do not understand that LTV is not a linear concept, and it is never quite as simple as “flooring the gas”.
The future of your business can’t be calculated like an annuity, because humans don’t behave predictably.
You can never expect your customers (past/present/future) to behave predictably.
Humans are predictably irrational, and if you get your assumptions wrong, you’ll get the entire thing wrong.
So, you need to be especially careful about how you use LTV.
Note: there is a predisposition (especially with Tech/SaaS companies) to use “big models” and “quantitative methods”, so they tend to overvalue their LTV model and undervalue more efficient forms of customer acquisition. When you open the hood and get a closer inspection of these companies, however, most of the time you’ll find they are much less skilled at more leveraged marketing techniques like social media, word of mouth, and public relations. It’s usually the scrappy “disruptors” with no marketing budget who are working to take customers away from these bloated entities by finding clever, inexpensive ways to get their name out there and scale growth organically.
There are many common pitfalls in the typical LTV calculation.
Most of the formulas are wrong and full of errors, especially when it comes to the generic ones that you usually find online.
There are a few useful ones, but most do not speak to the true predictability that you want in a real LTV/CLV model.
As an example, here is a formula that is traditionally used to calculate CLV for many recurring revenue businesses:
But there are 3 major issues with this methodology:
Issue #1: The Generic One Size Fit All Aspect. When you are building your LTV model, there are two inherent variables that should influence the methodology: (1) Contractual vs Non-Contractual: are you working in a contractual business (e.g. SaaS, Corporate Gyms, Cell Phone Service Providers, Netflix, etc.) where churn is explicitly observed? Or are you working in a non-contractual setting (e.g., hotels, restaurants, grocery stores, etc.) where you are uncertain if the customer’s last purchase was truly their last, or if they will return for additional purchases? And (2) Continuous vs Discrete: Are you operating in a continuous payment setting (Credit Cards, Amazon, etc.) where payments are less predictable? Or are you operating in a discrete payment setting (Gyms, Cell Phone Service Provider, Netflix, etc.) where customers pay (i.e., monthly, annually) on a fixed cycle? Not knowing these variables and payment settings can ruin your model from the start.
Issue #2: The Failure to Account for Uncertainty. This formula delivers an expected customer lifetime value in a singular, quantified expression, but it clearly leaves out the inherent variabilities and uncertainties in the projections. This lack of accounting for uncertainty/fluctuation becomes especially problematic in the context of online subscription based-models such as usage-based frameworks or SaaS models. In these businesses, the variability in customer revenue can be substantial, a problem that is further exacerbated when you’re working with different tiers of service offerings, like basic and premium plans.
Issue #3: The Constant Retention Rate. You don’t have to have a PhD in data science to see that the most critical problem with this methodology stems from using an aggregate customer retention rate. A superior and more nuanced approach would involve a comprehensive analysis and account of the varying retention rates observed across all customer cohorts, ensuring a more robust and comprehensive model.
Look, here’s the deal…
Applying LTV is a black art, and most businesses get it wrong.
When businesses get it wrong, they usually pay a heavy price.
When businesses pay a heavy price, they start looking for people to blame.
When blame starts going around, people tend to get fired.
When people get fired, sometimes the entire company goes bye-bye right along with them.
You don’t want to get this stuff wrong; it always ends badly.
Despite all the meticulous “planning” and “quant” it almost never works out the way you thought it would in the long run.
The final destination, this great promised land, imagined by the LTV disciples often never arrives.
Either you start to see diminishing returns (growth slows down) or you run out of capital/credit to fund losses.
If you are a large corporation and this happens, the Wall Street sharks will begin to circle the waters and ask to see profitability.
This is usually when most folks begin to see the horsemen pull-factor (more on this below) begin to show up in the data and the fragility of your model begins to rear its ugly head.
You hit your growth target, but you took heavy losses.
SAC is higher than your projections.
Your churn numbers are out of control.
But it’s too late now to do anything about it.
Many such cases.
But there is good news.
At the time of this writing, this 18th day of January, anno Domini 2024; where we sit Neanderthal on one hand and Singularity on the other, where any such event of significance may catapult us forward into the future or a thousand years back into the dark ages, we have access to more high-tech customer telemetry and data science tools than ever before.
You no longer need to be a data science genius to apply the “scientific method” to managing a customer relationship.
So, there is a small glimmer of hope.
But if your assumptions are wrong (even with these cool new tools) it may be garbage in and garbage out.
Your customers (i.e. human people) are involved, and nests of complex adaptive systems are at play.
It is not as simplistic as many of the LTV nerds and formula bros online always seem to make it.
Every variable in the formula is interdependent.
The outcomes may vary significantly from the math.
The reason for this, however, is simplistic: we live in an obscurantist world with imperfect information.
The humans who exist inside this world (your customers) are interesting creatures.
Their behaviors aren’t random.
They’re systematic and predictable – making the species predictably irrational.
This can make it incredibly difficult to properly define assumptions, measure risk, and predict behavior and outcomes properly and effectively.
This may destabilize the system away from its Nash equilibrium.
So, does this mean you should avoid LTV altogether?
I definitely don’t think so.
The LTV formula (when used correctly) can be a great tactical tool for analyzing, comparing, and monitoring variable market programs, across multiple channels.
But the model is only as good as the human working on it, and for that reason I suggest using it with extreme caution unless you really know what you’re doing.
That said, if you don’t know what you’re doing (and there is a lot on the line) I strongly recommend outsourcing your project.
But, if you’re still intent on proceeding on your own, here are a few critical factors to keep in mind when building your own model:
1. The variables are interdependent and stochastic. When we first started building these LTV models years ago, we noticed some interesting movements in the variables: they all seemed to tug on each other.
Note: from above, our variables are: ARPU, Average Customer Lifetime, WACC, Costs, and SAC.
We call these variables “the five horsemen”.
Imagine, if you will, a pentagram (a five-pointed star polygon) where each horse is standing at the tip of each of the vertices, all facing different directions.
Now, picture a string connecting them all.
When one horse pulls (let’s say ARPU), it pulls on the other horses, making it more difficult for them to go the direction you want.
So, let’s say your ARPU horse starts pulling one way (e.g., price), it pulls on churn, so churn rises.
Now, for the sake of this sh*tty analysis, let us assume your company is in growth mode. But the overlords at your company want to grow even faster, so the LTV zealots decide have the classic “very serious strategy meeting”.
This meeting always leads to the same inevitable outcome:
“We have decided we need to put our foot on the gas and spend more on marketing.”
It happens every time.
At first, everything looks like it’s going well (yay) but then you start to see one of your horses start to behave erratically (oh f*ck).
Your SAC horse is out of control and starts pulling in a very strange direction – and this action starts to rile up your churn horse, who has been asleep (minding its business) the whole time.
But now, your out-of-control SAC horse is now pulling hard on churn (as a result of your aggressive marketing program capturing lower quality customers) causing it to rise.
Then the zealots usually convene yet again in nervous meetings of panic and finger pointing when everything goes off the rails.
This is a common inevitability of the variable-pulling methodology of the LTV framework.
Note: a major constraint to keep in mind as you work your way through the var-pull matrix is the assumption that there is a finite amount of opportunities to purchase customers.
Now, you may be thinking: “Ha. That stuff only happens to inferior minds! My campaigns will be different! I’ll just beef up my customer service to improve churn and solve the puzzle…”
Not so fast my friend.
When one horse pulls one way, it makes it more difficult for the other horse to go in the direction he wants to go.
So, sure, you can hire more reps to lower churn, but this will only pull on your cost horse, which will not only directly impact future costs, but also potentially sabotage your cash flows.
If you try to raise ARPU (price) to offset negative cashflows you will also tug on churn, naturally causing it to increase.
This often leads people to the same conclusion: we must spend more on marketing to get more customers into the funnel, which pulls on your SAC horse again, putting you into what I call the lifetime value circle of death.
Note: remember, basic economics tells us that there is a finite amount of opportunities to purchase customers, so in theory, your over-aggressive marketing program will capture lower quality customers, which increases churn, which means you have to constantly feed the beast (e.g., consistently increase spend over time, which pulls on SAC) putting your entire model at risk.
As I mentioned above, the winning is in your assumptions, and one critical mistake I see a lot of people make is making the assumption that every critical metric will magically improve month over month (MoM), year over year (YoY).
But in the real world, this is rarely how it plays out, even though most company presentations show all metrics improving every month/quarter/year.
What usually happens when you find yourself inside the horse pull matrix is your burn rate increases, your cash flows decrease, and a lot of people end up getting fired in the end.
The more you work with LTV models (especially the more complex ones), the more you will see that the variables are not independent, but interdependent, and that the formula (whichever one you choose to use) presents an overly simplified abstraction of reality.
In my humble opinion, the horse pull matrix is quite arguably the most important concept to understand when it comes to lifetime value, and the failure to balance/optimize it correctly is central to understanding why most LTV models ultimately collapse and cannot be scaled ad infinitum.
2. LTV is math but it’s not rocket science. There is a tendency among LTV zealots and actors who play business experts on TV to think of LTV as “hard science” and apply it as if we were calculating Newtonian physics.
This sh*t isn’t rocket science; it isn’t absolute.
In the real world, businesses are intricate, complex, adaptive systems that cannot be modeled with absolute certainty.
The only thing you should be certain of is that there is uncertainty.
As I mentioned earlier, your model is only as good as your assumptions, and your projected LTV results are merely predictions hinging on assumptions, which may or may not prove valid.
Are you willing to bet your job on that?
For founders, are you willing to bet your company on that?
If you don’t know your numbers as well as you know your ABCs, you should not be making that bet.
It’s also important (critical) to keep in mind that LTV is not a “strategy”, nor does it create a sustainable competitive advantage.
LTV is merely a tool of measurement.
It should be used to test and analyze the effectiveness of your marketing spend — nothing more, nothing less.
One should never confuse output with input.
Also, if you’re keeping score at home, it’s important to note (again for the record) that LTV is not fundamentally grounded in science.
One should never walk around under the delusion that their “model” — which is nothing more advanced than buying customers below what you charge them — is actually “strategy” — because in reality, it is nothing more than a basic game of arbitrage, and arbitrage is not sustainable.
Yes, you are doing “math” but just because it is math does not mean it is good math.
If you get sloppy and let the brain worms overtake your mind, it is easy to become convinced that you are actually in charge, forgetting the inherent stochastic nature of the critical variables (SAC and ARPU most notably) and how difficult they are to control, even for a large corporate team.
Also, unless you have a really sophisticated model, there is nothing really preventing your competitors from executing the exact same “strategy”.
So, where exactly is the competitive advantage unless you build something that is incredibly sophisticated?
At the end of the day, you can build whatever you wish, but it’s important to keep in mind that LTV is not rocket science, it’s a simple formula that any average business school grad can calculate.
Fooling yourself into believing it gives you any sort of proprietary advantage is the first mistake.
Convincing yourself that you are a superior being because you built a basic LTV model will be your undoing.
3. The difficulty of acquiring a customer changes over time. You cannot blindly spend money on the same marketing source and expect to avoid hitting diminishing returns over time.
For example, the SAC for your first 1,000 customers is going to vary significantly when compared to your next 10,000 customers.
So, if you spend $1,000 to get your first 1,000 customers, the metrics you see for the first thousand will not necessarily carry over to the next 10,000 customers.
You cannot recklessly expect a 1 to 1 ratio of spend to customer to sustain as you steadily increase your ad spend/marketing budget (e.g., spending $10,000 will probably not land you 10,000 customers the way $1,000 landed you 1,000 customers).
That $10,000 spend will probably get you around 7,000 customers. That is a reasonable assumption.
So, as you scale your ad campaigns, it’s important to keep in mind that your return on investment will generally decrease to a point where it is no longer profitable.
Many factors can cause this (e.g., competition, audience saturation, the algorithm optimizing for low-cost users first to keep you spending money on the platform, etc.).
The mistake I see a lot of newer companies make is: running a small test, calculating a LTV to CAC ratio, and then making the assumption that the ratio will hold across a much larger ad spend.
What usually happens though, however, is that these companies will burn through the “cheap” first wave of new users, purchasing many early adopters at a low price, and then slowly come to the realization that every new user after that becomes incrementally more difficult to acquire.
This becomes even more difficult when you’re battling over laggards in a saturated market with tons of competition where users have multiple firms they can choose to attach themselves to.
4. It can take quite some time to get an accurate read. Let’s assume you do everything right. You build the perfect LTV model, you procure the capital you need, you craft the perfect marketing campaign, your segmentation analysis is perfect, your unit economics are perfect — everything is going according to plan, right?
Not so fast my friend.
Even if you do everything right, you will still need to let enough time pass in order to get enough data to make a reasonable calculation.
How long should you wait?
Well, it depends.
You will need to bake this into your assumptions, but you could reasonably assume you’ll need around three years of selling history to get a solid measurement.
Why will it take this long?
There is no “right” amount of time to wait, so you will have to feel it out based on your own situation, but you need to give the cohorts in your measurement groups enough time to potentially churn.
For example, do you have annual contracts? How long are they? Are you determining a LTV to CAC ratio on customers that haven’t had the chance to make a renewal decision?
Sure, you could model it based on contract duration and form a soft estimate, but this is not a true measurement or solid indicator of true customer behavior; you have no idea what that customer will do in the future (critical cashflow implications here) and you should not overvalue data like this in your model.
So, be sure to capture enough data of an “n” sample size, and then extrapolate it out as best you can over the entire customer base.
5. Growing rapidly with strong unit economics is very difficult. Let’s assume you are the founder of a deep tech space company.
And for the purposes of this sh*tty analysis let’s call it AstroBlast.
You are in year 3 at AstroBlast and at the beginning of what your analysts are predicting to be a growth stage.
Now, let’s assume your analysts are forecasting the company will do $20 million in revenue this year (year 1), $40 million the next year (year 2), and $80 million the year after that (year 3).
You, excited at the prospect of such high revenues (but also confused how such a thing could be possible) ask the analyst “how can we accomplish such a thing?” to which the analyst replies, “it’s simple, we will take a big chunk of revenues, 40 to 50 percent, and use it to invest in marketing.”
So, your “growth phase” budget for the next three years is $10 million (year 1), $20 million (year 2), and $40 million (year 3), respectively.
Note: it is common practice for many startups to execute what is known as a “land grab” (especially in the early days and/or in new markets) to try and acquire as much market share as possible to subsidize growth.
Of course, this sort of concept always sounds good in theory (most theoretical fantasies always do) but basic economics tells us this plan will most likely end in disaster.
For example, if you 4x your marketing spend, is it realistic or delusional to make the assumption that your SAC will decrease?
Another critical error that I see a lot of companies make is: the assumption that platforms like Google/Facebook/Meta/Amazon Ads can deliver an infinite supply of customers.
But being the great economist that I know you are — you know that the number one rule of economics is scarcity — and what you’ll often find is that these platforms are increasingly finite resources.
Users exist only in limited quantities.
So, what happens when you continue to buy more and more of a good that has a limited quantity?
Well, the price goes up.
You do not need to run a complex supply and demand analysis to figure this out.
This is remedial economics.
High school level stuff.
What often happens in these auction-based environments (e.g. Google, Meta, Amazon ads) is that in the competition for user attention — firms compete to outbid/outrank each other to gain advantage — and this incrementally increases the cost per user over time.
Once you hit diminishing returns, you are forced into the inevitability of having to spend more to reach parity from your previous baselines/successes, pulling on SAC, churn, etc.
As the game progresses, the tougher it is to keep your 5 horses in line.
And it can end in complete disaster.
These sort of “land grab” strategies only work if the underlying business model is sound.
And this means you need to know exactly what you are doing.
You’ve really got to know your numbers.
You’ve got to have your segments dialed in.
You’ve got to know your product market fit.
And you’ve got to know your unit economics.
There is no room for voodoo math.
The companies that make it work are able to figure out paid acquisition that’s profitable on a unit economics basis.
This means they have a deep understanding of LTV and CAC, and how they directly tie in and affect the success of your business.
In today’s market environment — with an ongoing downturn and a hypercompetitive digital marketplace — it is more important than ever to know your numbers and follow an economics program that makes sense.
- Fader, P. S., & Hardie, B. G. (2012). Customer-based corporate valuation. Journal of Marketing Research, 49(1), 40-56.
- Fader, P. S., Hardie, B. (2009). Probability Models for Customer-Base Analysis. Journal of Interactive Marketing.
- Gupta, S., & Lehmann, D. R. (2006). Customer lifetime value and firm valuation. Journal of Relationship Marketing, 5(2-3), 87-110.
- Smith, A. (2018). Data-Driven Marketing. Harvard Business Review.
- Hardie, B., Fader, P.S. (2006). How to Project Customer Retention.
- Hardie, B., Fader, P.S. (2014). What is Wrong with This CLV Formula?