Thanks to technology, we can quickly and seamlessly collaborate with coworkers — whether they’re across the room or across the globe. Location and time of day simply aren’t the barriers they once were. But despite all the efficiencies we gain, replacing in-person interactions with Slack messages and emails can have unintended consequences.
When we work remotely, the emotions we infuse into our face-to-face communications are less accessible to coworkers and leaders. Communication is reduced to fragmented, tweet-length digital messages, and it’s easy for employees to hide little signs of unhappiness.
As more and more organizations support both remote and on-site work, traditional paths for advancement are becoming less accurate. It’s easier to establish trust and give the perception that you’re providing value for employees who work on-site. Often remote workers feel that they have to work twice as hard to get the same recognition.
It’s no surprise that modern workplaces suffer from low employee engagement, faster burnout, and job hopping. Traditionally, leaders have tried to measure their team’s cultural health with surveys and assessments. They use these tools to collect quantitative and qualitative data and use it to try to understand the broader health of the company.
But surveys at scale are not only costly; they’re often incorrect. Employees might withhold critical information if they feel like it could result in shame or disciplinary action. Other employees might even try to game surveys in an attempt to advance their career. Anonymous surveys can help, but they’re not an adequate long-term solution.
Because we’re a remote-first organization at Very, we believe there has to be a better way to understand the health of our team, our culture, and the effectiveness of our leadership. For the last 18 months, we’ve been investing heavily in developing feedback loops and insights into our culture by applying data science and machine learning to measure our team’s asynchronous communications through Slack.
Instead of biased feedback loops, which can easily be gamed, we're using the hundreds of day-to-day communications across our dozens of Slack channels as the baseline dataset. By doing this, we've been able to gain insights that would otherwise be impossible.
A few notable highlights include:
- Visualizing the flow of information among people and working groups; as well as the strength of connections within each.
- Visualizing information silos, key communication bridges, and disconnected individuals.
- Understanding how these change over time as the organization evolves.
- By applying sentiment analysis and anomaly detection across several dimensions, we've been able to identify real company issues and potential team member burnout and address the root cause.
- By clustering team members into their working groups based on their natural communications, we’re able to identify stress and burnout at the group level, which provides a safety measure against individuals being unfairly targeted.
- Using graph theory and PageRank, we are able to have a more representative understanding of true leadership across the organization compared with traditional organizational psychology models.
- Predicting turnover rates 3-6 months into the future.
While we've made significant strides towards a better system for measuring cultural health, there are still several issues to work through before it can be adopted at scale. Specifically, the notion of individual privacy within public Slack channels left some early adopters concerned about how the tool would be perceived within the broader team.
And while we’ve developed some powerful diagnostics tools, they need to be coupled with equally powerful solution tools. It’s relatively easy to diagnose a symptom, but it’s much more complex to recommend a “treatment” when it comes to corrective measures an organization should take to combat trending cultural issues.