August 20, 2026

Why Great Talent Stays: A Survey of Root Quantitative Science’s EVP

Written by Kyle Schmitt

Over the last four years leading Quantitative Science at Root, I have spent a lot of time thinking about why people stay.

This may sound like a human resources question, and in some sense it is. But for a technical organization, it is also a systems question. Talented data scientists, actuaries, analysts, and machine learning engineers have many options. They are often motivated by a complex mixture of autonomy, intellectual stimulation, peer quality, business impact, flexibility, trust, mission, and growth. These factors do not operate independently. They reinforce each other, decay at different rates, and can be strengthened or weakened by seemingly small organizational decisions.

Many companies describe their employee value proposition, or EVP, through the language of recruiting. The phrase often evokes career pages, benefits summaries, hiring pitches, and polished statements about culture. These things matter, but they are not enough. An EVP is not what a company says about itself. It is the lived answer to a more difficult question: what does this organization consistently do that makes exceptional people want to join, engage, grow, and stay?

At Root, we have tried to answer that question with more than intuition.

Treating EVP as something to be measured

Technical organizations are generally comfortable measuring the things they believe are important. We measure customer behavior, model performance, pricing accuracy, operational throughput, conversion, retention, and risk. We instrument processes, inspect failure modes, and update strategies when the evidence changes. Yet many organizations manage employee experience with a much looser standard of evidence. A few people leave, a few people stay, leaders gather anecdotes, and an informal narrative emerges about what the organization does well or poorly.

This is understandable. Employee experience is hard to measure. Human motivation is complicated, and a single team can contain many contradictory experiences. A highly autonomous environment may feel energizing to one person and ambiguous to another. A fast-moving organization may feel exciting in one season and exhausting in the next. The goal, then, is not false precision. The goal is disciplined listening: enough structure to identify durable patterns, enough humility to accept uncomfortable feedback, and enough consistency to understand how the organization changes over time.

Since 2022, we have conducted recurring EVP interviews across Quantitative Science. The questions have evolved, but the center of gravity has remained the same. Why do people stay? Why would they recommend Root to a friend or colleague? What gives them energy? What drains it? What factors, if left unaddressed, might eventually cause them to consider other opportunities? We categorize the resulting themes, compare them over time, and translate the findings into action plans that influence management development, onboarding, learning programs, planning practices, organizational design, and prioritization.

This process has changed how I think about leadership. It is tempting to treat culture as something that emerges naturally from good intentions and strong people. To some extent, it does. Within the strongest organizational attributes lie contradictions. Empowerment can become confusion. Talent density can become complacency. Collegiality can become siloing. Impact can become abstraction. If these strengths are not measured, discussed, and maintained, they can erode quietly or swing too far to an extreme while leaders continue to describe the organization as it used to be.

What has remained consistent

Across several years of EVP interviews, through changes in the business, changes in market conditions, and changes in the organization itself, four strengths have remained especially durable within Quantitative Science: empowerment, talent density, collegiality, and impact. These are not the only positives in our EVP. Direct management, intellectual stimulation, and confidence in Root’s direction, and a “work from where works best” culture also show up meaningfully. But the first four have become the hallmarks of the organization.

Their persistence matters. A single survey can tell you something about a moment. A recurring signal across years, teams, tenures, and roles tells you something about the system. When new hires, long-tenured contributors, individual contributors, managers, data scientists, actuaries, analysts, and machine learning engineers independently describe similar strengths, it becomes easier to believe those strengths are not accidental.

It also becomes easier to protect them.

Empowerment is the foundation

The strongest theme in our EVP is empowerment. In the 2025 interviews, this surfaced more prominently than any other positive theme, with contributors describing an environment where they are given autonomy, decision rights, end-to-end accountability, and meaningful ownership over important work. This has been a consistent thread in prior EVP work as well: Root’s Quantitative Science team is at its best when individuals are trusted to reason from evidence, make consequential recommendations, and carry work through to implementation.

This is not performative ownership. In many organizations, technical contributors are asked to produce analysis while decisions are made elsewhere. The work may be intellectually interesting, but the contributor is several steps removed from the outcome. At Root, that distance is often much smaller. Data scientist don’t influence pricing strategy, they set pricing strategy. Actuaries don’t shape the performance, they define the performance and hold us accountable. Analysts don’t inform decisions, they actively manage operational processes to maximize profit. Machine learning engineers help define the technical platforms that support our models and data products. The work is rarely just a slide, a model artifact, or a recommendation handed off into the void.

This level of empowerment is not always comfortable. Consequential work involves uncertainty, tradeoffs, and accountability. It requires contributors to understand not only whether something is statistically defensible, but whether it is operationally feasible, economically meaningful, and strategically aligned. For many people, though, that is exactly what makes the work engaging. Ownership is motivating because it creates a direct relationship between judgment and outcome. It tells people that their thinking matters.

Talent density compounds

A second durable strength is talent density. This is one of those phrases that can sound generic until you experience its presence or absence. Strong colleagues change the day-to-day texture of work. They make technical discussions sharper, reviews more useful, decisions more durable, and ambitious projects more plausible. They also raise the standard for everyone around them, often without any explicit management intervention.

Across our EVP interviews, people repeatedly cite the caliber of their teammates as a reason they stay engaged. This is especially meaningful because many members of the team came to Root with substantial academic, technical, or industry experience. When experienced contributors describe the organization as unusually strong, it is a signal worth taking seriously. Talent density is not just about credentials, though we have historically attracted many people with deep quantitative training. It is about judgment, curiosity, ownership, communication, and the ability to improve the thinking of those around you.

This strength is also self-reinforcing. High-caliber people want to work with other high-caliber people. A strong hiring process improves the team, and a strong team improves the hiring process because candidates can feel the quality of the environment. Once someone joins, talent density accelerates development. It becomes easier to learn new methods, absorb domain knowledge, navigate ambiguity, and calibrate what excellent work looks like. Over time, this compounding effect becomes one of the organization’s most important advantages.

Collegiality makes rigor sustainable

There is a common belief that high-performance technical cultures require sharp elbows. Our experience has largely been the opposite. Quantitative Science at Root has historically aspired to be intellectually rigorous without being hierarchical. People should be able to ask basic questions, challenge assumptions, admit uncertainty, and disagree on technical grounds without turning every discussion into a contest of status.

This is easy to say and hard to preserve. Technical organizations naturally develop hierarchies of expertise. Some people know the data better. Some understand the models better. Some have more context on the business, the architecture, the regulatory environment, or the operational constraints. The challenge is to benefit from that expertise without allowing it to harden into an environment where ideas are filtered too heavily by title, tenure, or personality.

The EVP interviews suggest that collegiality remains one of Root’s distinguishing strengths. People describe a low-ego environment where colleagues are approachable, questions are welcome, and collaboration is sincere. This does not mean we always agree, or that every debate is frictionless. A healthy technical culture should contain disagreement. The goal is not to avoid tension, but to make disagreement productive and minimize the downsides of ego. Collegiality matters because it allows rigor to scale. It creates the conditions under which good ideas can come from anywhere and difficult problems can be solved without exhausting the people solving them.

Impact is visible

The fourth hallmark, and close cousin to ownership, is impact. Many technical professionals spend years working on models, analyses, systems, and recommendations without seeing much evidence that their work changed anything. This is not always a failure of the individual or even the organization. Large companies naturally create distance between technical work and business outcomes. Decisions pass through layers of translation, prioritization, and implementation until the original contributor may no longer recognize the final result.

Root’s Quantitative Science organization has largely avoided this abstraction. The connection between technical work and business performance is often direct. Contributors see pricing changes affect outcomes. They see operational analyses influence decisions. They see models and data products shape customer-facing systems. They see research become production. This creates a form of motivation that is difficult to manufacture through slogans or recognition programs. The work matters, and people can see that it matters.

Tallied across all teams (something we do rigorously via pro forma evaluations that are tracked), the Quantitative Science organization is behind ~$50M in incremental profit each year, upward of a 5x return on investment.

This is also one of the reasons insurance remains such a rich environment for quantitative talent. From the outside, insurance may not always be viewed as glamorous. From the inside, it is a dense applied decision science sandbox involving prediction, optimization, economics, regulation, behavioral signals, operational constraints, and long feedback loops. Root’s mission to build a more modern and fair insurance experience gives that technical work a clear purpose. The mathematical and statistical problems are challenging, but they are not merely academic. They sit close to the business and, ultimately, close to the customer.

The supporting conditions

Although empowerment, talent density, collegiality, and impact are the hallmarks, they do not stand alone. Several supporting conditions make them more durable. Flexibility continues to be a meaningful part of the EVP, especially in a distributed environment. People value the ability to work where they are effective and to structure their time with a degree of personal agency. Flexibility is not a substitute for good work design, but it is an important expression of trust.

Direct management has also become an increasingly important contributor to retention and engagement. In the 2025 interviews, many individuals highlighted their immediate managers as supportive, technically fluent, and invested in both outcomes and development. This is not incidental. As organizations grow, the quality of frontline management becomes one of the primary ways employees experience the company. Strong managers resolve ambiguity, create focus, identify growth opportunities, and help people navigate tradeoffs without removing ownership.

Mental stimulation remains a quieter but consistent strength. People generally describe the work as interesting, varied, and intellectually rewarding. This matters because technical talent is rarely motivated by comfort alone. An organization can have good benefits, kind colleagues, and reasonable processes, but if the work itself is dull or disconnected from meaningful problems, engagement will eventually suffer. Root’s problems are difficult enough to stretch people and applied enough to matter.

What we are still improving

A credible EVP cannot be only a list of strengths. One of the benefits of recurring measurement is that it forces us to confront the areas where the organization is not yet where we want it to be. In recent interviews, the most persistent opportunity areas have included work-life balance, administrative load, career trajectory, and technology investment. These concerns are not surprising in an ambitious technical organization operating in a complex and competitive business. Still, they are real.

The way we respond to these concerns matters. We do not want to manage toward an environment with no urgency, no pressure, and no hard problems. That would be inconsistent with the kind of organization many of us joined Root to build. But there is an important distinction between intensity and unnecessary drag. Context switching, unclear priorities, excessive administrative load, brittle tooling, or opaque growth paths can consume energy without creating commensurate value. Our action plans over time have reflected this distinction: improve planning without eliminating ambition, create more clarity without reducing ownership, invest in management without creating bureaucracy, and strengthen technical foundations without losing speed.

This is where EVP measurement is most useful. Without it, leaders may overreact to the loudest anecdote or under react to a pattern that is visible only in aggregate. With it, we can separate isolated frustrations from systemic concerns and track whether interventions are actually changing the lived experience of the team.

Retention as a signal

One reason we continue to invest in EVP measurement is that retention provides an important, if imperfect, feedback mechanism. From 2022 through 2025, Quantitative Science navigated a period of considerable change: a volatile technology market, shifting business priorities, organizational evolution, and increasing pressure for efficiency. Through that period, the organization retained exceptionally well and has generally remained highly engaged.

Retention is not proof that everything is working. People stay for many reasons, and strong retention can sometimes mask hidden dissatisfaction. But when retention is paired with recurring qualitative evidence, it becomes a more meaningful signal. In our case, the reasons people say they stay are consistent with the organization we are trying to build. They stay because they have ownership. They stay because they work with excellent colleagues. They stay because they can see the impact of their work. They stay because the culture is rigorous without being unnecessarily political or ego-driven.

That consistency gives me confidence, not complacency. The moment an organization assumes its EVP is permanently strong, it begins to put that strength at risk.

EVP is never finished

The most important lesson I have learned from four years of EVP work is that an employee value proposition is not discovered once. It is built, measured, maintained, and rebuilt as conditions change. The same practices that attract people in one phase of a company may not be sufficient in the next. The same autonomy that energizes a small team may require more scaffolding as the organization grows. The same informal culture that once enabled speed may eventually need clearer mechanisms for development, prioritization, and communication.

This is not a failure. It is the normal life cycle of an organization that is trying to improve without losing what made it special. At Root, we have treated EVP as a first-class concern because the quality of the team determines the quality of the work, and the quality of the work determines what the company can become.

The best employee value propositions are not marketing copy. They are the accumulation of thousands of decisions about whom to hire, what to reward, how to manage, where to invest, how to listen, and when to change. For Quantitative Science at Root, the evidence continues to point to a strong foundation: empowered people, exceptional colleagues, a collegial culture, and work whose impact is visible.

That foundation is worth protecting.

Follow us to learn more about our 2026 EVP interviews. We look forward to hearing how last year’s action plans have made a difference as well as how our EVP has held up over time.