Flo Health’s path to the hiring strategy in engineering
Like other high-growth start-ups, Flo Health adapts its organizational structure to meet the demands of growth and opportunities. As the internal structure changes, processes evolve, including the hiring process.
This is a story about how and why we changed our approach to hiring through strategy and culture.
Background
A year ago, we had the typical hiring problems of the hyper-growth stage. If you formed a new team, hiring could eat up all your time. On the other hand, managers on stable teams did not participate in hiring, and we had an organizational bias.
While staffed teams improved their processes, new teams constantly looked for candidates instead of impacting teamwork.
Overall, our time-to-hire was long and showed us a clear direction for improvement.
The more efficient your hiring process is, the shorter your time-to-hire will be.
Hiring Improvements RFC
We decided to unify the pipelines and revise the interview stages to improve time-to-hire.
When we deep-dived into it, we realized that it wasn’t possible to identify and quickly solve a series of problems. There were a lot of contradictions.
With our CTO, I started working on a request-for-comments document (RFC) and involved the talent team and other engineering managers in the discussion. We tried to find an approach that fits our engineering strategy.
We clarified our goals, culture, the companies we want to look up to, and why. As a result of offline and online discussions, we devised the following strategy, added the desired cultural aspects, and made some bets on execution.
Strategy
The main idea behind our strategy was to build sustainable hiring practices in the face of the growing organization. Our advantage would be a quick response for candidates.
What do all these statements mean?
Flo strategy
- Hiring into a company, not a team. This gives us flexibility. The company changes, new teams are created by splitting the current ones, and transitions between them happen.
- Fewer intermediate steps. They are rarely helpful and harm time-to-hire.
- Feedback loops. It is the most natural way to improve any process.
- A pull system instead of a push system. This is preferred in lean processes. We must build enablers to work as a system rather than through personal contact.
- More practical interviews. Interviews around system design and coding tasks increase confidence in evaluating a candidate.
Flo culture
- Bar raising. The quality of the hire is more important than speed. We have to avoid overhiring and focus on strong candidates.
- We only win as a team. One of the company’s core values is undoubtedly essential for improving the process in which recruiters, engineers, and managers are involved.
- Everyone improves their part of the interview. Each step must be enhanced by those involved.
- Shift left by education. Managers in the middle of the process help recruiters sell the vacancy and learn from engineers what to check earlier during the leadership interview.
- Simple process. It must be comfortable for all participants: candidates, managers, engineers, and recruiters.
How do we ensure that our changes help us move forward and not break something else?
Metrics
- Time-to-hire (the time between when the candidate enters the pipeline and accepts an offer) is the most accurate metric of the internal hiring process. It has to be evaluated on trends per vacancy and per location. Our north star is 10 days.
Health metrics
- It is a simple and comfortable process for all participants, especially our candidates. Evaluate feedback.
- Interview quality. Evaluate the ease of decision-making.
- It is a scalable process. Evaluate feedback from new interviewers.
Having all this, we aligned what we needed to achieve and autonomy on how. And the best way to prove that our strategy made sense was to experiment.
Experiment
We decided to start an experiment with those most interested. These were primarily managers from cross-functional teams who had common problems with hiring and dealing with many vacancies. Together with a talent team, they formed a working group.
During impact mapping sessions, we discussed possible ways to change the situation according to our strategy. Finally, we bet on the following:
- One pipeline for similar positions
- Designating pipeline owners instead of relying on the hiring manager for everything
- Pipeline visualization
- Predefined slots for interviews
- Leadership interviews in the middle of the process to gather feedback
Some of the changes were related to the quality of the interview:
- Pair all interviewers
- Create a bar-raisers list and involve them in the process
- Hiring managers make an offer
Taking metrics from our strategy, we started using check-ins to validate how we were progressing toward our goal.
Let’s look more closely at the most impactful changes we identified.
One pipeline for similar positions
This was a fundamental part of our changes. Let me explain why.
Most companies show all their hiring plans in open vacancies, and the engineering manager is responsible for how that progresses. Each available vacancy has a separate pipeline, which does not allow you to work systematically with similar positions.
Hiring managers usually deep-dive into every vacancy’s details. It is not only about participating in interviews. They have to open positions, describe the job, and post them. It takes a lot of time.
We also followed this process before we realized that we needed to work with similar roles in one pipeline.
Imagine a situation where the hiring manager doesn’t need to open any position and participate only in the offer stage. Most vacancies are evergreen. Only priorities matter.
At the same time, managers participate in the leadership interview. However, it is a balanced load calibrated on the company level with other managers.
This is similar to the approach described well by Jade Rubick:
I’m sure it’s more convenient for candidates because they don’t have to choose between many teams, just a position. Because we constantly ask about the candidate’s preferences throughout the interview, we have the opportunity to select the team they fit, not the one they applied to by chance.
As a result, we are less biased because the system pushes us to assess the candidate without reference to the specific team.
Pipeline owner & feedback loops
Instead of managers being responsible for all positions, we introduced the role of pipeline owners. Their main goal is to influence the time-to-hire and interview quality inside one pipeline.
They follow a systematic way of working with recruiters, engineers, and other pipeline owners by leading these feedback loops:
- The hiring funnel stages and drops between them
- Close percentage
- Candidate feedback
- Interviewer load
- Debrief rejections
- Poor-quality scorecards
- Sourcing statistics
All these points were perfectly described by Gergely Orosz, and we organize our routine around them.
Although this is an additional responsibility for some managers, it takes less time than when engineering managers had to deal with several positions.
However, feedback loops are not effortless. You will be asking many uncomfortable questions, and you must have enough energy to go through some over and over again. But the results are worth it.
Pipeline visualization
I usually wouldn’t recommend specific instruments because tools are secondary. But in the hiring world, we often lack the basic stuff for collaborative work.
TalentWall was a breath of fresh air for us. With this tool, we can see all our pipelines in one place and can work in Kanban style. It helps spot and remove bottlenecks with excellent visualization of data and metrics.
Also crucial, one pipeline per role simplifies company-wide metrics calculations, such as time-to-fill or time-to-offer. So, we could use these stats right out of the box.
Experiment results
If we look at the results of our experiment, we see that our changes showed excellent results according to time-to-hire.
And it’s fair to say that we saw a plateau after some success, like the fact that we’re just getting better at tracking metrics.
What about the health metrics?
The process has become comfortable for all participants:
- Engineers began to interview less and have more fun during interviews because the leadership stage rejected weak candidates.
- Managers began to interview with a balanced load and stopped bureaucracy with the opening of new positions.
- New managers were onboarded in this process with positive feedback.
- We cut a lot of administrative overhead for the talent team with unified pipelines and processes, predefined slots for interviews, interviewer pools, and automated metrics calculation.
The plateau indicated that it was time to make improvements inside pipelines and reinforced the need for the role of the pipeline owner.
Considering that we experimented on the most in-demand vacancies, we decided to roll out the experiment to the entire engineering.
Current state
In the year after we started, we reduced time-to-hire by half. According to onboarding trends, the quality of hiring has not gone down. We continue to hire strong engineers even in current times.
We also had a plateau, just as we had after the experiment. However, we knew what to do with it and were able to move past it.
It is worth noting that the second part of the plateau was during the Russian aggression in Ukraine, which greatly affected hiring in general. Without these changes, I would expect this metric to degrade.
Instead of 50 open positions, we now work with 10 evergreen ones, making the process lightweight and predictable. Of course, we have unique positions, but we use the same principles and stages of the interview for them too.
Today, we focus more on the quality of interviews, specifically:
- Ease of decision-making around candidates. Evaluate by the amount of complex debriefs.
- Show off our tech stack and get interviewees interested in the product and technical challenges. Evaluate by candidates’ feedback.
- Continue to improve time-to-hire by breaking down our hiring stages, locations, and roles. The main difference is that now it is business as usual and we know how to do it.
Most importantly, we leave room for experimentation. We accept that we will have unique hiring processes for a certain percentage of applicants. If that percentage grows, we must map it into the process.
Side effects
Feedback loops raised a lot of questions:
- How do we link a career ladder with hiring?
- How do we work with product owners around priorities on a system basis?
- How does the concept of bar raising change as the organization grows?
- Based on interview results, how do we use our salary formula at scale in different locations?
- How do we onboard a new interviewer into the hiring process?
We found answers to most of these questions, and they formed a valuable playbook.
Summary
Hiring is hard, and there is no universal solution to do it best. That depends on your goals.
Here’s what I would suggest:
- The key to success is strategy formalization. It helps balance tradeoffs and provides opportunities for autonomy and experimentation.
- Don’t forget about culture: value systems and ways of communicating.
- Building feedback loops helps you move forward and make adjustments. But be prepared that it is energy consuming.
I hope the story of our experience will be helpful and can inspire you to design your beneficial strategy.
More about Flo Health and Careers
