Xsolla fires 150 employees using analysis of chat and email activity

Reminds me of a story I heard from a friend.<p>Some consulting firm came in to the company and decided, based on number of commits (or some such metric), that one particular engineer was the lowest performing engineer on the team. So, management fired them.<p>Turns out that engineer was the one who everyone side-channeled with to get help when blocked. They were the one who knew the system best and were enabling everyone&#x27;s productivity, it just didn&#x27;t show up in the metrics.<p>I wonder how many false positives they got here.

<i>Y.T.&#x27;s mom pulls up the new memo, checks the time, and starts reading it. The estimated reading time is 15.62 minutes. Later, when Marietta does her end-of-day statistical roundup, sitting in her private office at 9:00 P.M., she will see the name of each employee and next to it, the amount of time spent reading this memo...<p>Y.T.&#x27;s mom decides to spend between fourteen and fifteen minutes reading the memo. It&#x27;s better for younger workers to spend too long, to show that they&#x27;re careful, not cocky. It&#x27;s better for older workers to go a little fast, to show good management potential. She&#x27;s pushing forty. She scans through the memo, hitting the Page Down button at reasonably regular intervals, occasionally paging back up to pretend to reread some earlier section. The computer is going to notice all this. It approves of rereading. It&#x27;s a small thing, but over a decade or so this stuff really shows up on your work-habits summary.</i><p>--Snow Crash by Neil Stephenson.

This reminds me of Ted Talks about the original thinkers by Adam Grant: <a href="https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=fxbCHn6gE3U" rel="nofollow">https:&#x2F;&#x2F;www.youtube.com&#x2F;watch?v=fxbCHn6gE3U</a> Just because these people are not behaving or contributing the same way like others, it doesn&#x27;t mean they are worthless.<p>If company design the work &quot;game&quot; like such, people will try to cheat the game like automatically producing whole bunch of crap on daily basis - is that productive? Then, they will have to figure out an anti-cheat, on and on. Play stupid game, get stupid prizes.<p>For people who might want to work for them, this really a turn off. What kind of potential metrics they could be looking for next? If a person made a 2000 line code that impacted the whole company, but that&#x27;s all the code that person produced, should he&#x2F;she get fired for being unproductive? What a rabbit hole!

Busyness is not productivity. But organizations have a history of using metrics that measure &#x27;busyness&#x27; when they say they are measuring productivity. Things like how often you are seen at the water cooler or chatting with a co-worker, how quickly you respond to communications, how quickly can you be reached, any time, anywhere, whenever someone wants to reach you.<p><a href="https:&#x2F;&#x2F;maxfrenzel.medium.com&#x2F;in-praise-of-deep-work-full-disconnectivity-and-deliberate-rest-e9fe5cc50a1d" rel="nofollow">https:&#x2F;&#x2F;maxfrenzel.medium.com&#x2F;in-praise-of-deep-work-full-di...</a>

We’re still incredibly limited in what behaviors do and don’t increase a teams productivity. I’ve worked in groups where some of the least productive people were also maybe the most valuable. <i>Very</i> strong in keeping cohesion and moral up inside the team. Whether it was the way they smiled or greeted people when they saw them or the way they just knew how to spot someone who was tangling with a problem and instinctively knew how to nudge that person away from the edge. I still cant even slightly pin down precisely what people like this bring, but I know I know this …thing… exists and it’s something important.<p>And when given the choice I’ve seen teams almost battle to get these “low producers” into the group even though their output wasn’t top-tier.<p>I don’t know how we measure for these slippery traits without falling into woo traps but until we figure out what to look for, great team building will remain an art.<p>We’re nowhere near a point where we can accurately measure the weird and chaotic quirks that make up a top-tier team. Building a team is still far more of an art more than a science.<p>I get letting a person or two go who have clearly demonstrated they’re not a good fit, but I can’t help but wonder how much of the kool-aid this company has drank to think firing this many people based on weird data measurements was at all a good idea.

A bit of Russian color. The CEO FB posts in Russian:<p>- the 1st, sounding great in Russian, yet hard to translate, is &quot;work energetically like f&amp;cking or f&amp;ck off&quot;<p>- the 2nd is &quot;looking for a great PR professional in Russia&quot;<p><a href="https:&#x2F;&#x2F;vc.ru&#x2F;hr&#x2F;277765-glava-xsolla-obyasnil-massovye-uvolneniya-nachal-iskat-piarshchika-i-necenzurno-predlozhil-vsem-rabotat-ili-uhodit" rel="nofollow">https:&#x2F;&#x2F;vc.ru&#x2F;hr&#x2F;277765-glava-xsolla-obyasnil-massovye-uvoln...</a><p>While without directly offensive words, not like the FB posts above, his message firing the people is still written in a very insulting disrespectful tone. &quot;Write me a long letter&quot;. Russian term is &quot;izdevka&quot;. The message is written with a lot of mistypes and Russian language mistakes. It doesn&#x27;t look to me like it was written in a rational and&#x2F;or sober state.<p>Some points from <a href="https:&#x2F;&#x2F;app2top.ru&#x2F;industry&#x2F;xsolla-ob-yasnila-uvol-neniya-zamedleniem-tempov-rosta-nizhe-40-urovnya-188990.html" rel="nofollow">https:&#x2F;&#x2F;app2top.ru&#x2F;industry&#x2F;xsolla-ob-yasnila-uvol-neniya-za...</a><p>- slowdown of growth is the reason for layoffs, and thus decision to cut 10% of salary budget<p>- primarily firing rank-and-file, not management<p>- time spend in git&#x2F;IDE wasn&#x27;t considered in the evaluating employee efficiency<p>- most layoffs in Perm is because salaries there are lower than in US<p>I get a feeling that he is losing it, like he probably got teared a new one by the investors or something like this, got drunk and got that epiphany ... The metrics collection was probably going for some time, and finally they found a use for them.

Agapitov later gave a long interview to Meduza explaining the decision. It&#x27;s surprisingly honest. Here is an interesting quote:<p><i>In America, in recent years, both in our company and in the media, various minorities have been actively protected and people are very cautious about the dismissal of gays, blacks, and more recently, Asians. But in Xsolla, we did not give immunity to these groups, because all decisions that a human makes, from the point of view of Americans, have a risk of being biased, and our algorithmic solution is as unbiased as possible. Therefore, from the point of view of American media and American civil society, for us it is much better and more indisputable than if some manager fired someone.</i><p><a href="https:&#x2F;&#x2F;translate.google.com&#x2F;translate?sl=ru&amp;tl=en&amp;u=https:&#x2F;&#x2F;meduza.io&#x2F;feature&#x2F;2021&#x2F;08&#x2F;06&#x2F;my-formalno-esche-nikogo-ne-uvolili-my-skazali-kompaniya-vas-ne-tsenit" rel="nofollow">https:&#x2F;&#x2F;translate.google.com&#x2F;translate?sl=ru&amp;tl=en&amp;u=https:&#x2F;...</a> (In Russian, Google translated)

I wouldn&#x27;t take this at face value. It&#x27;s more likely that they already wanted to do the layoffs and just put the burden on the employees.

The translation of the letter has lost some of its undertones, which betray some psychological disorders of the author.<p>For example, &quot;... If you want to stay in contact with me, please write me a long letter...&quot; should be better translated as &quot;... If you want to <i>remain my acquaintance</i>, please write me a long letter...&quot;<p>(source: I&#x27;m russian)

In my first job, I took standup to be a pace keeping device, so for a while I would hack together something to have it ready, or if I did two things in a day, I would withhold one to have something to report the next day.<p>I did not continue to do that as when a task carried over into the next day I wasn’t hassled about it, but whenever things like this occur I have to wonder if there are places where this strategy would be required.

There was a round of layoffs I went through at a consulting company where we are pretty sure some pure overhead MBAs pulled a report of who was missing &quot;cloud&quot; skills from an internal skills tool and had had anything less than a top box review laat year. Without consulting even the GM dozens of senior&#x2F;principle people were let go.<p>This naturally ended up a disaster because A) those people weren&#x27;t updating their skills in the tool because they were high demand rock stars and B) they had deep customer relationships, to the point where many customers threatened to (and I&#x27;m some cases did) cancel their contracts and stop doing business with us. My understanding is somewhere between 1\3 and 1\2 of everyone let go was hired back at higher pay after keeping all their secerence pay, and many of the rest just refused to do so.

I know everyone will think I am crazy, but I actually like to get work done at work, and I am so fucking sick of everyone bunking off and never being available with the change to remote work!<p>It is a real problem! People are just bunking off and not working!<p>I also saw this when people were in office, however. Eventually, some people just stop doing work completely and coast until they are noticed or find a new org to be a parasite of, a surprisingly rare occurrence.<p>And I say this as an extremely lazy employee who often checks out at 4pm. It gets BAD with some people.

&gt; we will help you find a good place, where you will earn more and work even less<p>Wow you&#x27;re really selling it for us. PR nightmare.

And it&#x27;s funny because in my experience the people not chatting and checking email all the time are the ones actually getting work done.

I am often surprised how often email and slack chat volume corresponds to how productive an employee is. I’ve often had employees I’ve had to fire or resigned that after the fact I check their activity and they basically aren’t doing anything.<p>But using this data without the context is pretty foolish.

The only thing surprising about this is it happened in a tight labor market. Perhaps it isn’t tight for this industry&#x2F;market?<p>I was in the room one time when HR decided to let go of 250 people based on the sq footage rate of the office they worked in. It’s not at all odd that in the post office world they are looking at other dumb metrics for these decisions.

Want to bet that there was a round of human review applied after the algorithm ran where, at a minimum, employees judged to be high-performing (or otherwise favored) were saved even though they were initially flagged by the AI&#x2F;data analysis?

They probably let go some valuable employees, given that things like this lack nuance entirely. All of the services listed could be considered distractions from work in several situations depending on the engineer.

This reminds me some people who open Jira tickets for every tiny things, then have fixes for these tickets.

The letter is so unprofessional it makes me question the validity of the decision making more than I otherwise would. An email dripping with sarcasm has no purpose here except to make the CEO feel tickled inside, it&#x27;s masturbatory.

Depending on the nature of the employee’s role, looking at hard metrics gathered from the systems the employee works in seems like something every line manager already does. One could argue that a system capable of essentially automating part of the manager’s job is a net positive and could increase productivity and possibly even fairness in evaluating some aspects of performance.<p>But at most, this should be supplemental information considered in a proper evaluation process, not the sole source.<p>Things quickly go south when they start to involve “softer” signals like email activity.<p>And further still that the direct outcome was immediate termination and not some kind of performance improvement period.<p>Perhaps this last part is not present in the Russian workplace, I’m not familiar.<p>I can’t help but feel like this is a real life slippery slope that came fully to fruition.

&gt;Nadia and her care team partnered with seven leading HR agencies, as we will help you find a good place, where you will earn more and work even less.<p>Earn more and work less?<p>That&#x27;s weird.<p>Those poor folks who still work for him will be working more and earning less than those he fired.

We&#x27;re having a field day with this one!<p>I know it&#x27;s rare relatively speaking, but between this, Blizzard and others, it seems insane how disconnected a CEO is from their company.<p>While short sighted, thankfully this doesn&#x27;t represent the norm of companies.

I know employers can see your internal &quot;private&quot; communications, but it is also illegal, at least in most Western countries, to actively watch and spy on your employees. Would they have some recourse if it was say the US?

This looks like the inverse of &quot;everyone needs to come back to the office so that we know you are working&quot;.<p>I doubt this will go well since if it keeps happening unproductive people will become the best at gaming the system.<p>But at the same time, there has to be a better way than being observed by a boss in a physical office. I don&#x27;t want to commute just because people on my team who were unproductive in the office are now just as unproductive at home and not seeing their face daily makes it hard to tell if they took another job and just haven&#x27;t told anyone.

Wow that&#x27;s crazy. 150 at the same time? I&#x27;m not sure how long did it took him to review all of those before taking that kind of decision.<p>But again not sure if that&#x27;s the whole reason behind the action, maybe the company is not doing so well at the same time? That&#x27;s why they needed to change.<p>Regardless of the reason, I&#x27;m sure firing 150 people at the same time will give you some backslash in any way. Good luck to all those 150 employee for their future

Funny, I can draw a pretty substantial parallel between “getting stuff done” and “hours of time to work on a problem uninterrupted by pointless chat and E-mail”.

For me this whole story sounds like an excuse to fire bad developers.<p>It’s a bit easier to delegate to some authority (data) than just say “you’re bad, we dont want you”.

I know that all corporate data are owned by the companies, but i would leave any company that regularly checks my communication activities to “measure” my performance. Jira, confluence, git commits, etc. is acceptable and can be quantified somehow, but slack dms or emails? Do you also want my camera footage? This is toxic.

Wonder if what they really did was tap employee webcams, then measure the employees attendance that way? Similar to what&#x27;s done for student attendance measurement in remote classes.<p>It kind of sounds like it from the description:<p><pre><code> ... you were not always present at the workplace when you worked remotely.</code></pre>

&gt;“We want all our employees to think daily about how their actions and decisions affect the company’s fate and success because we have very ambitious goals in the coming years; it is one of Xsolla’s values and it is reflected in everything — from operating standards to compensation system,”<p>Oh the irony

I&#x27;m not entirely surprised considering their reputation<p><a href="https:&#x2F;&#x2F;mmos.com&#x2F;news&#x2F;chronicles-of-elyria-class-action-lawsuit-has-been-split-into-two-separate-cases" rel="nofollow">https:&#x2F;&#x2F;mmos.com&#x2F;news&#x2F;chronicles-of-elyria-class-action-laws...</a>

The letter screenshot has a lot of grammatical errors. Seems like it is google translated from english to russian.

speaking from experience this isnt going to end well. the employees that didn&#x27;t get fired will feel betrayed and you will lose alot of good people with the institutional knowledge of your systems. some people that remain will figure out how to game the system while contributing as little as possible.

I guess that beats doing Pulse Check surveys.<p>I mean obviously there are internal problems if that many people are disengaged to the point of needing to be let go, but after reading the letter, if the company follows through on helping everyone find a new job, it could have been worse?

That&#x27;s 150 employees that just dodged a bullet.<p>I feel more sorry for the 151st.

In post-soviet russia, correlation IS causation, apparently…

Hello new Russia, fundamentally quite similar to old Russia. Things like this strengthen my conviction that it probably wasn&#x27;t communism that wrecked Russia, bur vice versa.

&quot;You received this email because my big data team analyzed your activities in Jira, Confluence, Gmail, chats, documents, dashboards and tagged you as unengaged and unproductive employees. In other words, you were not always present at the workplace when you worked remotely.&quot;<p>Sweet.<p>Maybe we are hitting the day when all of your code commits, all communication, location in the building, telecommute meetings, etc. etc. could be run through the Big Machine and give you a grade. All automated.<p>A parrot could be trained to fire people in a special HR chamber.