What’s driving employee satisfaction

One thing I love about startups is that they’re essentially big “natural social experiments.” Like, ask any other startup what they’re calling their OKRs/all hands meetings/product artefacts/roadmap/etc., and you’ll get wildly different answers – which is cool because it means startups aren’t just testing products, they’re also testing how to work together, and sometimes that leads to good ideas.

It’s interesting though that most startups probably don’t take data collection on how they work together as seriously as data collection on their sales, product-market fit or whatever, even though it’s arguably as important to their success. Here at Oscar, we aim to take a data-driven approach to everything we do, including employee engagement.

One of the best ways we can gauge how employees are feeling about the work they’re doing here is through the employee engagement survey we conduct twice a year (we use a platform called Glint for that). This has been especially important throughout this pandemic, as we moved to a remote work model, and as we’ve grown to over 2,000 employees, distributed across the U.S.

Our survey asks employees about things like how meaningful they find their work, how they feel about Oscar’s culture, how they perceive Oscar’s leadership team and whether they believe the leadership team will take action based on the survey, how much they believe Oscar values DI&B, the level of innovation their team demonstrates and more. We’ve been conducting our surveys in the same format since 2016, so we have a great trendline of how our employees feel about a variety of topics and key employee engagement drivers over time.

Whenever we do surveys, the most important objective is to know where the issues are and what we have to fix internally – but I also always play around with the data a bunch, to see what we can learn about the psychology of our interactions. One way in which I’ve analyzed the data is to look at it by department and correlate the departments’ answers to each question with all other questions. (The data is anonymized and only available on a departmental level, so you have to believe that this represents a reasonably effective subsampling.)

For each driver, we also look at the standard deviation of the series over the years to get an idea of how “stable” a correlation has been over time and throw anything out with stdev > 0.2. This isn’t meant to be particularly scientific, but here are some points that I might not have expected as intuitively:

  • Employee perception and trust of the leadership team is one of the most important drivers for overall employee satisfaction. Employees expect the leadership team to play a key role in creating culture and a sense of belonging within the company, and also connect a company’s performance and future prospects with their belief in the leadership team.
  • The other key drivers are how excited employees feel about the future of the company they work for, and whether they believe the work they’re doing is meaningful to them. Obviously a company can’t run on purpose alone, but empowering employees to make decisions on their work, and connecting them to the overall mission behind the work, can help leaders retain and engage their team.


Employee satisfaction, or eSat, is the score assigned to a company based on the answers to the question of “how happy are you working at Oscar?” Not too surprisingly, it correlates strongly with “I rarely think about looking for a job at a different company” (Retention). Still, good to see that confirmed! It also correlates strongly with “I am excited about Oscar’s future (Prospects), which isn’t surprising either.

But what good is that if you’re a founder? Well, I for one found it nice to see that “I feel like I belong at Oscar” (Belonging) and “I have confidence in the leadership team” (Leadership) and “I feel satisfied with the recognition or praise I receive for my work” (Recognition) show up here in the top correlations as well, because those are more under your control as a startup than perhaps your Prospects at times (if times aren’t really bad at times, you’re not a real startup, just lucky). It’s often said “people don’t quit companies, they quit managers,” and I would not say that that’s borne out by this here: the macro stuff really matters.

Now, the lowest stable correlation with eSat is just Role (“My role is an excellent fit with my strengths”) at 0.5. You could see that as indicative of the same broader point.


The score related to the statement “I rarely think about looking for a job at a different company” has one of the highest correlations with eSat, as we saw above. This is also true for other markers related to employee retention: Growth (“I have good opportunities to learn and grow at Oscar”), Career (“I have good career opportunities at Oscar”), Prospects, and Purpose (“The work I do at Oscar is meaningful to me”). Looking at the standard deviation of these over time, these are pretty stable relationships (the lower the red bar, the less fluctuation the correlation figures saw over the years). 

Ok, so this is where the personal career and growth correlations show up: so it seems that your overall satisfaction with a job is driven more by the macro picture, but whether you think “maybe I should pick up my head and look elsewhere” is driven more by your personal career experience. But, Prospects and Purpose show up here prominently too!

Interestingly, all the low-correlation questions exhibit high variance over time. It suggests that what’s driving people to rate this question negatively actually has fluctuated a lot over the years. It’s like the old saying: all happy families are alike, but each unhappy family is unhappy in its own way.


This is the score related to the statement “I have confidence in the leadership team.” The highest correlations are with eSat and Prospects – meaning, employees seem to connect a company’s performance with their trust in the leadership team. (I think there is data for public companies that suggests that CEO performance is actually pretty random and not meaningfully correlated with company success as measured by stock returns – whether or not that’s true, employees definitely seem to think differently!) 

Belonging and Communication are the next most-correlated, and “Oscar has a great culture” (Culture) right after that. To me, this provides interesting data on what employees seem to consider the job of the leadership team: to create culture and a sense of belonging.

The lowest-correlated relationships were Role, Work Life Balance and Equal Opportunities. It could be that more people associate those with their immediate work environment, their team and their manager. (Obviously I think leadership has a huge impact on Equal Opportunities, including by creating the right frameworks and systems.)

Work Life Balance:

Startups are hard places to work, particularly during the pandemic and remote work. So how does work life balance factor in all of this? It’s highly correlated with Resources (“I have the resources I need to do my job well”, nice sanity check on the methodology here). Empowerment (“I feel empowered making decisions regarding my work”) is an influence too. 

That suggests that the feeling of being in control of one’s work is almost as important as the actual resources one has available. This is an interesting lever in company management: I’d posit that budgets are always going to be constrained in startups – but generating empowerment is not. Empowering people to make more decisions themselves can help alleviate resource constraints. Somewhat curiously, Team and eSat are both at a fairly low correlation – Work Life Balance is less correlated than quite a few other factors with ultimate satisfaction.

The two lowest correlations are Purpose and Prospects, at 0.38 and 0.4, respectively, meaning if you have a strong connection to what a company is trying to achieve, you are less focused on work life balance. It’s also interesting to see that Innovation has a similar effect as Prospects – if you feel like your team is innovating, you perceive your work life balance as ok. And finally, the same is true for Voice and Inclusion.


The highest correlations with Purpose are quite a bit lower than for the previous questions. This could mean that purpose is a more intrinsic quality to a particular company. Yes, Leadership and even Prospects influence it, but it’s still generally weaker correlated with everything else than, say, Belonging.

Purpose is least correlated with Work Life Balance, Rewards, Resources. Obviously a company can’t run just on purpose, but it appears to be helping people in some tangible ways get through tougher times.

Again, as I said, this isn’t meant to be especially scientific, but I hope this gives people – especially fellow leaders – something to think about!

Leave a Reply

%d bloggers like this: