Cracking Employee Lifetime Value (ELTV)

Reason #1 why I’m jealous of Marketing & Sales.

Jessica Zwaan
Incompass Labs
12 min readNov 16, 2023


Hello, Boss! Imagine…you’re the CEO of a large business. You’ve got a lot on your agenda. Product keep telling you they need some engineers to implement the seemingly endless list of customer needs before the next big enterprise deal will close. Your legal team have just let you know they’ve found a roadblock to the expansion into APAC. Sales are asking for yet another new outbound optimization tool that will revolutionise their pipeline (they promise!) And the People Team? They’ve got a list of requests you aren’t even sure when you’ll find time to mull over — a closure over Christmas? A WFH stipend?

In business, we’re all asking for something, and often with good intent. In People, however, we have a specifically bad habit of asking for things without using the words “so that”. And when we do, we often say things which don’t sound clear to CEOs, “so that we can improve engagement” or “so that we can offer management training”… hmm. These don’t sound so good juxtaposed against Sales coming with a, “so that we can increase revenue by 5%.”

When I’m looking at my P&L there are three big numbers I care deeply about:
A) Revenue today
B) Revenue soon
C) Costs, and primarily headcount cost (65%!!)

If you’re the CEO speaking to Product, Operations, and Sales & Marketing teams often the language we’re talking is related to A and B. We’re talking “line goes up” and when line goes up, that’s good. My investors are happy, my runway is distant, and my sleep is uninterrupted.

When People Operations are in the room, often we’re talking about C. Asking for more spend, or being told to spend less. Because, honestly — look, when that spend line goes up… things get rough. Sure, the recruitment team can hire 10 more engineers, and as long as Product are saying what they’ll work on will facilitate A (and ideally B) I’m all ears. But holiday breaks, new benefits, a god-danged 4-day work week!? Yikes. That all sounds like line goes down territory and, no offense, the board is telling me we’re in for the long-haul this bear market.

Time to talk “line goes up”

We’re in a tough situation in the People Ops world. We desperately want to be heard and respected in our businesses, but we lack the language to get the same priority revenue generation will naturally get just from talking growth rather than spend.

To further complicate matters, the world People Ops folks operate in is messy and chaotic. Humans at work are data-disasters. They hate being tracked, dislike surveys, loathe goal setting. If we started giving everyone a line in the company revenue targets they’d probably revolt (although…?)

Recently Benn Stancil (my favourite Substack at the moment) posted a really interesting article where he talks about democratising profit and loss within a company (no, I will not write in American English. I know I am in America.):

In some rough sense, the deal we sign with an employer when we take a job is that we’ll socialize our gains in exchange for socializing our losses. We get paid a steady salary, and the company takes on the risk that if we do something bad — like leaving an expensive computer running in the cloud for no reason — they’ll bear the cost. But if we do something good — like turn the expensive computer off — they get to keep the gains. They might fire us or promote us, but they lose — or make — the money.

There are lots of reasons I like the points he’s making in this post, but in short one thing I think is relevent here is that, in businesses, when we’re employing people, whatever their role, they do have some kind of impact on profit and loss. It’s our primary role in People Ops to maximise the profit, and minimise the loss. But tracking that is… wow it’s hard man.

But what if…

AI turning me into a math god

Customer Lifetime Value (aka CLTV)

CLTV is a metric used to estimate the total revenue a business can reasonably expect from a single customer account throughout the business relationship. The calculation can consider factors like the customer’s purchase history, the length of the business relationship, and other factors influencing their future buying behavior.

A lot of SaaS business do a very rough and ready calculation like the below:

CLTV=Average Value of a Purchase×Number of Repeat Sales×Average Retention Time

However, if we’re really trying to be accurate (but keeping it readable in a single blog post):

CLTV = ∑_t (rt×margin_t) / (1+d)^t

where r_t and margin_t represent the customer’s retention probability and profit margin given the customer is alive, respectively, at time t_2 and d is the discount factor.

Ok that’s pretty simple. But not all customers act like the average customer, and in fact acting on averages is generally considered a pretty dodgy idea; it makes it deceptively hard to get good numbers for the formula above for each of your customers.

My friends and co-founders Professor Peter Fader and Professor Dan McCarthy, are renowned experts in customer lifetime value (CLV). They uses a more sophisticated approach to calculate CLV that involves probabilistic models. Their method focuses on predicting customer behavior based on past behavior, considering factors like the frequency and recency of a customer’s purchases, as well as the monetary value of these purchases. The plain vanilla version of their approach is referred to as an RFM (Recency, Frequency, Monetary) model. Here’s a simplified overview:

  1. Recency and Frequency: The model begins by calculating a customer’s recency (how recently a customer made a purchase) and frequency (how often they make purchases).
  2. Statistical Models: Then, the model employs a generative, probabilistic model such as theBeta-Geometric/Beta-Bernoulli (BG/BB) model for predicting repeat purchases. This model estimates the probability that a customer is still “alive” and makes future purchase predictions based on this.
  3. Monetary Value: Alongside the BG/BB model, the Gamma-Gamma model is often used to predict the monetary value associated with those purchases.
  4. Lifetime Value Calculation: The output from these models is then used to calculate the expected number of transactions and the expected monetary value of those transactions over a customer’s lifetime, which together give the CLV.

These two are good at this stuff. So good in fact that they built a company called Zodiac to do just this and then they (very cleverly) sold it to Nike who wanted the magic all to themselves!

They’re also very kind to me and listen to me go on and on about how we need something like this in People Operations so we can finally get a piece of the “line goes up” action. That magic would be called ELTV.

Employee Lifetime Value on the Aggregate

A few months ago I wrote a handy little LinkedIn post talking about the white whale I’m chasing; ELTV. And specifically, ELTV to ARC (Aquisition and Retention Cost). I made a very simple little average calculation, not dissimilar in simplicity to the formula shared above for CLTV, it involves calculating the profit for each dollar invested in employee compensation (including benefits and equity) over the average tenure across your team.

Look at that📈

There are many ways to improve ELTV. You can speed up onboarding, make your team more efficient, increase their tenure.

All the stuff we’re asking for drives more $$… right?

Equally, many of the decisions we make to increase ELTV should have a mirrored ARC. If we’re spending, we should see more performance, longer tenure, or quicker “ramp ups”. That’s the big sell, anyway.

Everyone in your team, regardless of their role, should have some meaningful contribution to revenue. That’s what a well designed org structure should do. If we thought hiring 50 VP Sales would drive 50x more revenue we would do it. So that’s why we hire an EA for the VP sales, to enable them to spend X more time on making $Y more money, but also why we can’t really explain why spending $ more money on Y more L&D folks will make it work the other way around.

It’s worth noting here that people that are already thinking about ELTV have a lot of challenges with how this metric (and specifically the averages) can and should be used. In short, lots of issues with this one, but it’s a useful little way to get the People Team to start thinking about the investments we’re making as having an impact on revenue. It’s also a helpful way to think more long-term, focussing on maximising the investments in your people rather than on short-term outcomes or “putting out fires” in People Ops. I’ll call it an optimistic first step.

(A nice second step is breaking your ELTV and ARC into segments or functions, but that’s a step I won’t get into here.)

So what would the next step be? Calculating ELTV using a methodology similar to Customer Lifetime Value (CLTV), as proposed by experts like Peter and Dan, is theoretically possible, but it’s more complex due to the different nature of employee relationships compared to customer relationships. In a perfect world, the concept of ELTV would aim to quantify the total revenue value an employee brings to an organization throughout their specific tenure.

It’s also very helpful to overlay that with what we’re spending in our ARC. It’s spectacularly compelling to make an argument to your CEO that spending $X more on ARC results in $Y more in ELTV. Now we’re really cooking with gas. All of a sudden we’re not just the person-who-asks-for-$ but the person-who-can-make-that-65%-of-my-budget-really-worth-something.

That. I want that.

Yeah, me too.

Good news is we have some things, and we can probably work some things out now that we have our supreme AI overlord hanging above us.

So what do we already know?

Thanks to tools like Xero, Pento, Gousto, Zenefits…Charlie, HiBob, Remote, Oyster, Humaans, Workday, Paycom, Bamboo, Rippling… Lattice now too? (wait is everyone building HRIS systems these days????) we know what we’re spending. We have an increasingly granular view on how much we’re spending on each person, and on our HR budgets; payroll, benefits, overheads, training, offices, the list goes on. We also have quite a lot of complex data around the profiles of those individuals: their role, tenure, reason for exit, values, projects, goals, managers, interview feedback, and much more than we probably give ourselves credit for really.

Beneficially, due to the continious improvements in revenue tracking, we’re also able to get very clear views on our overall ARR (annual recurring revenue), the segments that comes from, and often even the initiatives and product features that drive that growth.

What are we missing?

In a nutshell: performance.

Let’s start where it’s “easy” (hint: it’s not). Understanding how a sales person tracks to revenue is often already somewhere in your data. Sam makes $100,000 per month in NNR (Net New Revenue). Jane makes $90,000 in NRR. So Jane’s ELTV is 10% less than Sam’s. Easy. Let’s fire Jane, find another Sam.

Well, hold on there, Buster. Jane actually trained Sam. In fact, Jane trained… well, everyone in the sales team. And… dang. They’re all doing really well. Ok, so maybe Jane isn’t less impactful than Sam to revenue. Maybe Jane is just really good at other things? But are those things more important? 🪱🥫.

We know behaviours matter.

Layering in the RFM model, but giving it some HR spice.

One of the things I love about Pete and Dan’s work is that they’re not just sticking to the averages, they’re looking at customer behaviour as well. Just like we would in HR. The beauty of data in HR is that it tells a story beyond numbers — its about people, their journeys, and their growth.

We need to understand, at the corner of all of this, a few things to really get under the skin of ELTV.

  1. Performance according to goals
  2. Performance according to behaviours that drive top-line growth
  3. The organisational network analysis of your team


As I mentioned above, in Product and Sales, many teams are already directly connecting the work they do day-to-day to revenue outcomes. We can understand from within that work if what they are doing (Sending 500 outreach emails, launching a new landing page, driving 10,000 new visitors) will have a positive, neutral, or — god forbid — negative relationship with revenue growth. Very sophisticated companies may even have connections from People, Operations, Legal, and Customer directly to top-line revenue targets.

Within this, there is a lot of really valuable data we can begin to collect from the team (and programmatically) around the appropriateness of goals, their actual value, and those who worked on them.


Now we’re getting a bit more in the weeds. For this, ultimately, we need to know which behaviours in your company (values, competencies) truly drive performance and improve your chances of success long term.

If you’re operating purely on goal-based on output, you lose the detail of how impactful collaboration, safety, resilience, and motivating others truly is to building a high-performing company sustainably (not to mention a nice place to work). Almost every business I’ve ever worked in has acknowledged this, it’s why performance is often measured across the axis of the “what” and the “how” and why we all agree that brilliant jerks are generally a bad idea.

I’m not going to get into why and how I think we can do this, but — well — we can.

If we can point to a single person in your team of 1,000 or 10,000 and say that they are, objectively and with a low chance of error, the person who is strongest at Collaboration, we’ve got something we can hang our hat on.

Organisational Network Analysis

Finally, and something we should already have from the vast-swathes of valuable metadata we have in our HR systems and within our goals and behaviour data, we can begin to build complex organisational network analysis showing who in your team is the strongest performer, their impacts and “halo effect” on others, and how they work day to day.

All three above helps us get the most important (and messy) part of the RFM model Pete and Dan use within their predictive CLTV model, the Monetary.

How would we apply the RFM model into ELTV?

  • Recency: The employee’s most recent performance evaluations and goal achievement.
  • Frequency: The frequency of goal achievement, specific milestones (such as promotion, onboarding etc.)
  • Monetary: As above, the estimated economic value of the employee’s contributions to the organization.

The final thing I’ve had framed to me recently by someone very clever was the idea of Capital Intensity around the specific work different folks are doing, including which tools they do to alleviate human labor. I think that’s probably a next thing keeping me up at night wandering Google Scholar, and I should let our weary minds rest for now.

Wait. So that’s it? We don’t know what we don’t know?

Yeah, sorry. That’s it for now. We don’t know what we don’t know. But we’re getting closer to knowing it. ¯\_(ツ)_/¯

I’m partially writing this blog to get more folks thinking about ELTV even as it currently exists (messy, in the aggregate), because I personally still feel it’s incredibly helpful and specifically helpful in places like Sales, Consulting Services, Marketing and any revenue-generating roles.

I’m also writing it because I know that one day a post like this will crack something open in my beloved Pete and Dan, and the whole rest of the Incompass Labs team, and we can sit down and build what we’re aiming to do at Peerful: give People Operations teams the data we need to take truly unignorable action. It’s not just a question of how for me really, it’s just a question of when.

Ok that’s all from me, folks. 👋

👉 Buy my book on Amazon! 👈
I talk plenty more about this way of working, and how to use product management methodologies day-to-day, I’ve been told it’s a good read, but I’m never quite sure.

Check out my LinkedIn
Check out the things I have done/do do
Follow me on twitter: @JessicaMayZwaan

Me and my cat, looking professional



Jessica Zwaan
Incompass Labs

G’day. 🐨 I am a person and I like to think I am good enough to do it professionally. So that’s what I do.