Remote Working Productivity Software? Please Don’t Spy On Me…
When we wanted to introduce remote working software we got a very clear message from the company “please don’t spy on me…”
https://knowyourday.com/ is a productivity suite, time and motion study, security tool and more, which we created internally. It was born out of necessity and later expanded into a fully featured platform. We created it because we could not find anything which did the job and also met our criteria.
In choosing a solution we faced several major challenges:
- Our organisation is spread out across the globe
- Its very difficult to understand when people are under worked, over worked or frustrated
- Software developers and geeks are not always the greatest of communicators
- By the time we found out we had a problem with a team or team member it was often too late to address it
One thing management was sure about was that if we wanted a software toolkit to help us manage users, then we had to be comfortable using it on ourselves too, and that’s where things become a little tricky.
Most if not all of the software available went straight for the jugular of “employee monitoring”, and generally this genre of software prided itself on being highly secretive and invasive. This included things we felt were completely unacceptable such as CASB solutions which intercept and decrypt all cloud traffic (Cloud Access Security Broker https://en.wikipedia.org/wiki/Cloud_access_security_broker ) to out and out keystroke loggers, screen recorders and command and control systems. In fact most of the solutions felt like they would not be out of place in a Black Ops movie.
Apart from the fact we did not want to be spied on, equally concerning were the security risks associated with these solutions. The truth is that when you run all emails and keystrokes through a central monitoring system you create an enormous vulnerability.
Perhaps best summed up by this “security gate” , which could also be described as a “ladder”…..
I think it was fair to say we, knew what we did not want and that led on to us trying to figure out what we did want. Our list of what we did want included
- Better understanding how people in remote locations were working
- Being able to see whether changes were incrementally positive or detrimental
- Protecting people who were overworking and stressed
- Early visibility of people who were in trouble
- Spotting anomalies in activity
- Helping teams work more efficiently together
- Create monthly board reports that identified trends
We discovered very early on that we could get all the information we needed for the above and more if we collected high volumes of low quality data and ran it through a Neural Network to clean it up. So we started, by counting but not recording keystrokes, looking at which apps were in use, and looking at the machine network activity. In many ways we approached the problem more like a machine Antivirus / Firewall endpoint than a personnel monitoring platform. We looked at what the software (being driven by the users) was doing. The data we collected was deliberately stripped of anything identifying or personal, and typically we collected 60,000 minimal information data points per user per day. Once we collated that data, we grouped it into 30 minute segments for a minimum of seven days before starting to analyse it.
When the first batch of data came through, we knew we had something. I think it is fair to say that we were all blown away. With a simple seven day view we could see a real life time and motion view, and the insights we gained were immediate.
Once we started to look at whole teams though, the data was too much. To tackle this we used a Convolutional Neural Network Auto Classifier to sift through the data. What we found was that whilst the vast majority of people classified as expected, there were two scenarios where people classified as anomalous
1) When they were struggling with their role
2) When they were being given tasks by another department or performing tasks that were a-typical for their role
Armed with these insights we were able to unlock teams and individuals in ways too numerous to mention. One example was a huge workflow issue we discovered, where the QA team always worked late and were very unproductive in the morning. It is something we had never realised. It came from a culture of finishing the day’s work before going home. The problem was that there was nothing to do for QA the next morning. A simple pipe-lining piece of software and a change of approach, reduced their excessive hours and made them more productive within days.
The final piece of the jigsaw puzzle for us were the automated monthly board reports.
Month on month reports were critical for us. As the teams grew and evolved, we were able to see whether the efficiency improved or decreased (it rarely stayed the same). Bigger screens gave us a 8% uplift in productivity, a new VPN client reduced our productivity by 15% until the IT team figured out more efficient split tunneling to control which traffic did and did not route to the different networks.
Perhaps the biggest measure of success for us was the analysts team. This team performs a broadly measurable task that remains the same month in and month out. They add and repair media monitoring sites. By changing workflows, improving pipelining and investing in better hardware, we were able to show a 34% improvement in productivity over 5 months. In the face of rising costs everywhere, these improvements have allowed that team to remain competitive.
Looking after your staff and spying on them do not have to be the same thing. I can say with certainty that the latter is not needed, whilst the former certainly is. So after a lot of work we ended up with a system that everyone from trainees to directors was comfortable to use.
I can safely say because we discuss it frequently, that if these tools were taken away from us now, we would literally feel as though we had our sight taken away.
