Tech companies (ab)using personal data is today’s 1984. And while Facebook is at the forefront of our outrage, a growing industry in the HR Tech space is also experiencing challenges overcoming the creepiness factor of mining all our interactions at work.
Over the past years, there has been a minor explosion of startups working in the datafication of workforce performance. This burst of modern-day HR tools is a result of 3 trends:
- our work lives being captured on communication and productivity software tools
- the continuing social media indoctrination that feeds our obsession to swipe, like and comment
- price wars leading to tumbling fees for cloud services
This confluence has led to so many new startups that HR Tech has made its way into the eye sore infographics world.
But despite these new startups and their claims of creating productive and engaged work environments, when I talk to employees of companies who have tried some of these tools, they express concern about the creepiness factor of having so much data about their work activity collected and used.
One Tech Company Novel Experiment in the Datafication of Workforce Management
Among the noise of this fragmented landscape, I recently learned about an initiative piloted by a company I have been advising for years.
Satalia combines machine learning with optimization algorithms to solve complex problems in scheduling, routing and delivery. If you’re a service provider with thousands of employees that you need to keep 100% productive, each employee with different skill sets, and each project with different staffing requirements, you end up with a multi-variant problem that can’t be managed efficiently by humans – although most service providers try to do just that. Satalia Scheduler automates the process, and then continuously learns from data feeds to produce more effective team scheduling over time.
Satalia is evolving the exercise of workforce scheduling.
But that’s not why I’m writing about them. What fascinated me is how Satalia combined innovative organizational structures with its technology platforms to improve its organizational development and growth.
From its inception, Satalia embraced values of radical transparency and psychological safety through trust.
This led Satalia to operate much like a Holacracy. However, unlike other attempts at Holacracy, Satalia married its unique management structure with technology and data in search of real organizational change.
When HR Tech Meets Organizational Development
“‘Semantic’ is an internal organizational analysis tool developed on top of Satalia’s AI and data platform”, Jai Clarke-Binns told me. Jai, like other Satalia employees, doesn’t have a traditional title. His interests and contributions mostly lie in Satalia’s “Humans Circle” with a focus on “Humans Inclusivity”.
Using Semantic, Satalia collects data from the systems and applications used by its employees – everything from Slack, to calendars, to github, to HR and project management tools, and more.
Other HR analytics companies claim to offer similar tools – with two really important differences.
- Most vendors talk about where they could pull data from, without developing the technical backbone to analyze and draw insights from that data, even if they actually did pull it all.
- Most companies haven’t invested in a culture that supports the broad collection and use of employee data. So implementing these HR analytics tools tends to feel creepy to their employees.
It is Satalia’s investment in its culture that made it okay for it to use data from its employees’ digital interaction.
You can’t use the HR tools of the future without also investing in the organizational culture of the future.
If you meet Daniel Hulme, CEO of Satalia, you’ll see that he doesn’t shy away from complicated, sensitive challenges. It wasn’t surprising to me that he and his team set out to apply their organizational structure with their technology to one of the toughest HR discussions – pay review.
Is A Data-Driven Pay Review Process Possible?
Satalia believed it was possible, and so set out to prove it.
With the data collected in Semantic, Satalia used its own AI platform to automatically mine that data for insights. The insights they were looking for differed from traditional HR analytics tools. Instead of using this data to understand who was most productive, they used it to decide whose “opinion” of another’s work held more “weight”.
This weighting algorithm then fed their semi-Holacratic pay review process, that looked like this.
- Every employee’s pay was made transparent.
- Each person publicly requested the pay raise she thought she deserved.
- The aggregated data about each person’s contributions to various projects were presented.
- With that data made transparent, each person could then vote on any other person’s pay raise request, indicating whether they think it should be approved at that level, below that level or even above that level.
This is where the Satalia weighting algorithm kicked in. If Angela (who is also in Satalia’s “Humans” Circle) votes Jai’s pay request downward, but Angela’s weighting with respect to Jai is really low, her vote has very little impact. If her weighting with respect to Jai is high, her vote has a great deal more impact.
I asked Jai and Angela how they felt about this process that they helped design. Neither of them seemed concerned that others in the company were able to see their pay and vote on their pay request. There were concerns that the voting itself hadn’t been made transparent. While the company had some good reasons for keeping votes private, Satalia is looking at new ways to drive decentralized pay decision-making without having to resort to any information secrecy.
Satalia learned some interesting things from its experiment with transparent pay and decentralized, data-driven pay review.
They found that women didn’t request as high a salary as their male counterparts did.
The good news is that through this process, the company managed to close its gender wage gap.
Sometimes, there was a wide range of opinions on the appropriateness of pay requests.
Even when presented with data about an employee’s varied contributions to the company, an evaluator might still evaluate that employee largely based on just the projects they have been involved in together.
Both these findings suggest that an even more data-driven, and less heuristic driven, approach to setting pay would lead to more equitable pay levels.
“As a technology company devoted to self-organization, non-hierarchical structure, employee happiness and data science, Satalia strives for AI that makes difficult decisions without succumbing to the bias of humans,” says Jessie Petrova, a data scientist at Satalia.
How Not to Be Creepy
I believe Satalia has shown why existing HR Tech startups haven’t yet cracked the code on re-inventing people management. To make significant progress that also doesn’t feel creepy, HR Tech solutions need to do more than just collect data about our interactions and graph them for us. They even need to do more than use machine learning to find patterns and draw insights from the collected data.
A new paradigm for people management needs to combine data science with organizational development for a solution that invests and motivates employees to be transparent about their activities in service of developing and growing as a team.