Why are people commenting about "leaderless teams" as if the article defines that to mean, "A team where there is no leader present"? That's not what this article is about at all. I know it's "bad form" to accuse people of having not read the submission, but folks come on, go read the actual submission!
The article can be summed up by this line:
> What emerged from this exercise is that social class, education, gender and athletic ability were less important for leadership than the capacity for an individual to attend to others in the group.
So it's about how the best leaders are those who attend to the others in the group to create psychological safety so that those others can thrive without emotional risk.
The article also talks about how people seem to have forgotten this core concept but they haven't; it was the key conclusion to Google's own research into its most successful teams as well.[0]
If your team isn't primarily focused on providing psychological safety, then your leaders need to read Google's findings ASAP...
Good summary. Leadership can be shared and it's not about power or authority. As Google's data shows, teams can be autonomous and self organising but leaderless teams don't mean leaderless organisations.
I didn't find much of interest in the article, but I realized that we have the concept of "leaderless teams" in the company where I work.
The company produces several distinct but related products. All of the products have a shared set of concerns and resources. They are all built upon the same in-house operating system, they all use the same development and testing lab infrastructure, they all have UIs, automated testing, etc. Migration of CIs between teams (generally voluntary) is relatively common.
To avoid duplication of work and to coordinate work and improvements on these shared resources, "guilds" spontaneously formed. Guild members more or less represent their own team's interests while also contributing to guild efforts as a whole, if they can, and when they have the time. Each guild has semi-regular meetings. There is usually a guild "leader" but this person is more of a secretary in the sense that they only schedule meetings, take notes, and facilitate communication. Guild leaders are just CIs like everyone else in the guild and have no actual authority to direct work.
Company management is aware of guilds and will allocate resources to them if specifically requested but otherwise have no involvement in the guilds and never attend meetings. Guilds (along with non-asshat managers) are one of the things this company has that I wish was more commonplace in tech companies.
I work at a company that at one point attempted to form “guilds” around some unowned, decaying cross-cutting concerns that were still very important.
The guild was great at identifying issues to work on, but most members didn’t have the time to actually contribute on any of the more hairy issues. It was basically group therapy.
Ditto. If those concerns aren’t directly tied to profit or costs it’s hard to justify why time is required to work on them.
Also - and perhaps this is just me - but the world “guild” just makes me think of Dwarves and fantasy games. “The dwarvish guild is requesting more iron ore to make hammers”. Hard to get taken seriously.
Guild is not a term in corporate parlance. Perhaps “think tank” resonates a little better? IDK
Thanks for sharing. This sounds like a very sophisticated culture (though it may just seem obvious to those on the inside.) Whatever the structure or norms (e.g. guilds, squads, cells, lattice, project teams, etc) it sounds like leadership has given some guidance and more important empowerment. But, empowerment is an alliance. It must be taken up with given. By everyone in fact, as leadership is an act not a position, if you go by the definition of leadership as the act of driving change (aka doing real work.) Are you able to share the company? I'd be interested in studying it. info@brettmacfarlane.com
I think what has held HL3 back is Valve thinking they need to deliver something _more_ than just a continuation of HL2. HL2 was meant to showcase physics and the source engine, Valve needed something like that for HL3. Alyx I think was supposed to be that "thing": VR, the next frontier.
Unfortunately, I think most folks would have been perfectly fine with HL3 being a continuation of HL2, myself included, so now we have to wait for Valves hubris to subside.
Sometimes when I think about how anxiety-provoking it must be to even _think_ about making Half-Life 3, I wonder if Valve might not be better off building something that completely defies expectations about what "Half-Life 3" should be instead of trying to meet those existing expectations and then inevitably disappointing a huge number of people even if they produce something pretty damn good.
What could it look like instead? Anything! I think a Divinity-style RPG Half-Life would be heaps of fun. Or a hero-based RTS that still has Gordon Freeman centre stage. I think basically any genre that can tell a story would be fine.
Of course this guarantees disappointing a lot of people, but I feel there's a lot better chance of Valve producing something creatively worthwhile this way than trying to make a "vastly better Half-Life 2", which is what most people seem to be waiting for.
This pretty much exactly describes Alyx. They defied expectations about what Half-Life 3 should be and disappointed a large number of people (who were mad that it was VR exclusive) despite producing something incredible and unprecedented.
I have experienced leaderless teams in my career (46 years) a handful of times.
They formed around a problem that was unrecognized by Management (thankfully) but need to be solved. My best example was at a large SW dev firm in the late 80s. World wide electronic communications was need and faxes and leased lines connected to Decwriters were not good enough. I was assigned to the task as just one of my responsibility. A Single 1200 baud access to internet email at a university solved the problem. The problem expanded many people wanted the same service so we need someone (team member 2) who had access to a UNIX systems. Eight serial ports and 2400 internet connection to the first local ISP.
Rinse repeat 5 years later 10 people, part time, supported all internet connections and intranet/internet web-servers including application without management sanction. Everyone had a service role in there own department. Being connected to the others meant we performed at a high level. We didn't have a leader per se but we all moved in a common direction.
It lasted 8 years and finally several Directors discovered we existed as a team. Like sharks they each took a bite out of us, each one trying to become the dictator of the internet. Procedures were imposed. We were not allow to speak to each other with out going through the directors. Our individual productivity dropped. Angry internal clients screaming for the old way. Nothing got done.
Three months later the director that won got half the people back together but we were not the same any more. I still found interesting work for 2 more years before finally leaving.
Leaderless teams can work under the right conditions. I have tried to create them when I moved into management and it worked to some degree a few times. I did not lead or participate in the teams I just suggested people talk to each other know that the might click. I fed them problems (and resources) that I knew that they could solve that I knew would take more effort/time if we did it though channels.
The individuals got recognition and I got my problems solved.
Grat experiences, thanks for sharing. You experience with directors is a clear example I suggest of where their desire to control obliterated the autonomy and ability of the team to function effectively. Mature and self aware senior leaders restrain the natural impulse to control.
I'd love to learn more, maybe even write a teaching case study if you're interested.
Each department had a person service their internal or external client. The could provide the service the old slow way or our way.
EG You ask the Tech Pub dept for the 5000 page Word perfect documents and wait a month to get it printed and shipped through formal channels. Or a couple of days through our team to have it on a secure download location on our public server.
Marketing, New Business, Sales and Engineering each had a rep. If you want a Schematic out of engineering, several weeks until it was plotted or a day or two to get the right Autocad or Bentley Microstation cad file into the secure down load.
I was the goto guy if you needed data off our Mainframe. It was a massive data sink but it was hard to get anything out.
Web dev let me turn the MF into a FTP server were I could pull data down to the Web servers. Web application were written for each need.
Oh if you did have your own departmental server I would ask H/W services to get me some old PC hardware and have a machine up and ask the Computer center to host it. None of this had official sanction.
Expectations were low as the competition were Green screen MF program.
The problem was there was 3 to 5 year waiting list for new MF projects.
I could do new apps in a week. It was written prototype level code with no expectation of maintaining it.
One subcontract tracking app stayed up for 7 years (3 years after I left).
The alternative was to talk to someone who got a monthly 1000 page printout and manual marked up the updates that were being sent to Daily Data Entry. I download the big monthly output once and each Daily update and merged them.
Probably less the 50 lines of Perl. Set up a Perl backed web page the searched the file and display matching records.
This is an empty, hollow article and its not like how I experienced groups for 20+ years.
Just like with siblings in families or groups of friends, everyone naturally gets a role in a group; that's just how groups of people let alone animals work. Need to learn how to help others is fine and dandy, until it affects your own autonomy, let alone your identity, let alone the groups' autonomy. Than it will eat at you.
Also what happens in groups is that if there is someone thats is truely not fit for a group, the whole group dynamic will change for the worse and it will try to oust it like a cancer. Until there is someone that speaks up (a leader) things will stay the same because the group will value the "being a group" above all else, even the goal of which the group was founded on.
So? You actually need a good natural leader to guide you through some of the bad stuff. I think leaders are often misunderstood of being only authoritarian, but thats not the case.
I will agree it's hollow in that it is story of an experiment not a specific real world situation. :-)
You are right that in families everyone gets a role. Unlike business you don't join a family at will, you're born into it and roles are very fixed.
Companies are much more dynamic and our roles (formal and informal) change. We also work "at will" and can change companies, teams and colleagues.
There are different types of leadership roles. As you say, one type gets through the bad stuff. Others do really well when times are good. Some only in certain situations.
In tech, where the nature of work and expertise needed day to day, week to week, can vary a lot at times, a variety of leadership capabilities and styles, along with the universal ability to give a shit about colleagues, makes for an awesome team.
In fairness to the article, it’s not actually about leaderless teams. Leaderless teams were just experiments allowing them to identify better leaders, who then went on to take positions of command.
It’s so just-obvious that it needs no argument. It’s just natural.
> I think leaders are often misunderstood of being only authoritarian, but thats not the case.
Bosses (as in formal authorities, people who have formal control) have been rebranded as “leadership” exactly because “leader” sounds much more voluntary and consensual.
> What emerged from this exercise is that social class, education, gender and athletic ability were less important for leadership than the capacity for an individual to attend to others in the group.
So, not actually leaderless teams, but a self-organizing structure from which emerged a role that is easily mapped to leadership, though not called that by name?
That we incorrectly define leadership, and attach a lot of ideas to it that have nothing to do with leadership itself, should be obvious to anybody who has worked inside a badly run organization. And it's a trivial observation that doesn't get us any closer to forming better teams.
I used to think I wanted to work in a completely flat organization, with nobody who thought they could boss me around. Having done that, I now realize what I actually want is someone who is good at being the boss, and isn't a prick about it. I want somebody to have a synoptic view of what's going on, keep all the parts running smoothly, and act as a shield against all the various distractions coming from outside the team or the company, so I can do my job well. Such people are multipliers for effectiveness and happiness on teams.
What you usually get is someone who has no particular skill in these areas, but likes to have authority, and likes making more money.
I sometimes wonder if the problem isn't that there are only a few actually good leaders in the world, but we need to act like there are 10x or 100x more than that in order for our economy to run. Every once in a while you meet a great one, but the majority of "leaders", from the C level to the PM level, should really be doing something else instead.
Leaders will naturally acquire de facto power in a group setting. Power attracts the power hungry, so avoiding a buildup of power is important. This build up can be offset greatly by limiting the introduction of de jure power into the equation.
You list attributes of bad managers: no particular skill in the area, likes to have authority, and likes making more money. We can mitigate the likelihood of attracting those managers by providing the opposite.
1. Let teams self-organize, where team members will value those who are both competent and effective cooperators.
2. Don't pay them more, it's just a role within the team, like being responsible for some feature. Have pay scales that value your experienced members, just don't pay your lead more than their more senior reports.
3. But who'll do the paperwork and other secretarial work that managers are normally responsible for in addition to their decision making/leadership duties? Hire a team secretary: someone who does not have hiring and firing power over the rest of the team.
You can't. You live with that. Everyone around us does it. Bad politicians, bad owners, bad managers, bad leaders, etc. A bad director in your company temporarily affects you, but finally is just a job.
In general, I agree 100% with your comment, in my career so far I would say I met only 1 leader, the rest is good talkers or people who know how to keep their positions.
Unfortunately I have come to the realization that it's just how it is, and you move on. Life is not just work, and some people know better how to climb the ladder than others, some are just luckier. Etc. Etc. Therefore I learnt to be happy with what I have, knowing that it could be better or worse. When it's too bad, it's probably time to move on.
I think I agree with that. The more actionable question might be "how do I identify teams with good leaders, as opposed to the other kind, and only work with them from now on?" I'm fully on board with life being not just work, but about 20% of it is, so it's worth optimizing.
The most bureaucratic and toxic cultures I've ever seen up close are flat or describes as "family." Personality and power take over in the lack of clear ways of working, huge wastage financially and emotionally.
Perfmafrost is unhealthy and unsustainable for an org, so in permathaw where everything is perpetually fluid.
Leadership isn't power and control. Leadership is the ability to drive change with others. Leadership is an act, not a position. Flat organizations tend not to be leadersless teams and rather leaderless organisations.
There are many good leaders, but the tend to be in the middle not the top. Based upon research evidence.
Leaderless models work very well in cases where tasks are stable, tasks can be assigned 1:1 to team members, team members more-or-less understand which tasks are for them, and any major changes to tasks can be coordinated on timeframes of months or years.
The classic example is the traditional academic department at a non-research university.
1. The main deliverable are courses.
2. Each professor has some assigned courses. Once assigned, the professor can teach the course mostly in isolation. Additionally, the department tends to hire faculty in a way that makes sense for their curriculum; ie, the one Systems professor in a small CS department is probably going to be teaching the Systems electives.
3. The courses offered are mostly stable. When the curriculum changes, the changes can be managed in a distributed fashion because the change will take at least 6 months to 18 months before rollout.
Therefore, a traditional academic department can run in a mostly leaderless fashion, with a department chair who plays a supporting role and perhaps a bit of a coaching role for newer hires.
This doesn't work as well when tasks require close collaboration or when tasks evolve rapidly. It's not impossible. Just much harder. In those cases you tend to end up with a recognizable leadership vacuum, leading to sub-optimal outcomes mentioned in other comments.
Anyways, that's all a bit of an aside and not directly germane to the article. But I do wonder if selecting leaders via leaderless teams exercises selects for leaders who work well in well-defined tasks with teams that are already partitioned into quite specific roles, but flail in environments with more dynamic tasks/goals.
I don't think this was the main thrust of the article. The "leaderless team" concept was used to identify individuals within that group that were well-suited to leadership, specifically by identifying (and coaching) those individuals in the group who possess "the capacity for an individual to attend to others in the group".
So more of a question of how to identify the emergence of organic leadership skills within a setting where there is no "assigned" leader.
You touch on a few good points. The key is how leadership is defined. I use the definition that leadership is an act not a position, and driving change is the act of leadership.
In roles of routine tasks, execution, efficiency and repetition is the task. It take management or monitoring but not necessarily leadership.
Leaderless teams actually are the best way to deal with highly ambiguous challenges where the path is unclear and the best people to figure it out are the people doing the work, not a manager or monitor.
So the leaderless team exercise identifies individual who do well in ambiguous situations. They may actually fail or get bored in situations without dynamic tasks/goals.
Leaderless teams always start to build a hierarchy on its own and the longer they exist the harder it gets to get new people into the group. It's really hard to manage.
"The Tyranny of Structurelessness" is a great read on this. TLDR: structure will always form, so it's best for formalize it from the beginning.
That said, I've been interested in the idea of "adhocracies" for a while now, and it seems to me that it would be possible to form structureless groups that, once they have fulfilled their purpose, disband thus becoming immune from the sort of rot that would otherwise happen.
And I'll also say: the same rot can happen in a group with a well-defined structure too. So it's not all cut and dry.
If the culture is adhocracy then leaderless teams are extra effective. Though a self organising or autonomous team can thrive in any organisation. For example Toyota in the glory days of TPS. Where individuals took up their role in driving change and attending the needs of colleagues.
This is exactly what the article seems to be talking about -- how to identify the emergence of leadership where no formal hierarchy exists, and how to coach those individuals to find leaders.
Yep, leaderless teams are either ad-hoc for a short time, or they develop an unofficial hierarchy (often based on the complaining of certain people, not anything else).
To me authority and leadership are related but orthogonal.
Authorities come from the ability to author. It is a creation right. The creation could be a story, a software and, probably what most most people have in mind about authority, a rule. On the other hand, leadership is the ability to, well, lead. You go where ever you want, but if you want people to follow, you need leadership.
One can be a leader without authority. Think of a manager, while it might not be obvious, it is ultimately up to the "managee" to decide whether to carry out what the manager requires. You may fire, you may hire, but none of those guarantee you things getting done. Manager has no authority.
Now, write a software. People wants to add a feature to YOUR software. You have absolute authority in deciding whether to accept.
Very much there is a delineation between authority and leadership.
Authority is the right to do work either formally through policy/power or informally through influence/trust.
Authority has to be granted whereas leadership is something anyone can take up regardless of role. Leadership is the ability to drive change, an act not a position. Just because someone has a lot of authority doesn't not mean they act as leaders, if they operate purely for individual means. In my eyes there is a morality, but that's my personal approach.
Thanks, I really connect with your point that it's the managee to decide whether to carry out what the manager requires. In the knowledge economy we work at will (though may be constrained by perceived needed.)
People follow those who are deemed legitimate leaders. The legitimacy may come from power (I can fire you if you don't do what I say), or it may come from any of a number of other things.
[edit]
I should also note that everybody has the authority to create rules, just not the legitimacy to get people to follow those rules.
Article is interest, but miss important thing - military at war is very special type of collective, near impossible in civilian collectives.
For example, remember covid-liners cruise ships, appeared on first weeks of pandemic, where it was impossible to get out, but they have all needs satisfied.
This also remember me parts of USSR system, where people moves where limited, but there where no financial crisis, no recession, free food, free apt, for some significant share of people, this is dream life.
You touch on a deeper truth about change in organisations. It can't just be important, it has to be vital. That is why military is such a pure testing ground for leadership. Or why Edmonson's psychological safety research was initially carierd out in hospitals. In less life or death situations, the fact change is only important not vital allows more emotional needs to distort things. Leaderless teams only thrive when senior authority cultivates or tolerates such autonomy. If they try to control due to some individual or systemic needs then autonomous leadership dies.
This is an interesting article. My takeaway is that this British team during the war seems to have streamlined a process for surfacing leadership traits that are relevant in a military context.
Lately, my view on what makes a good leader and what I'm trying to practice myself is simplifying down quite a bit. I've come to the conclusion that if there is one attribute you should optimize for, either for yourself as a potential leader, or in choosing someone to follow when selecting a venture, it's ... technical competence.
Hopefully it's clear that I'm defining technical competence broadly to mean deep knowledge about the details of your particular domain. I think this definition is applicable to engineers, attorneys, doctors, entrepreneurs, etc.
I think generalizable leadership attributes are important but purely supplementary to technical competence.
I used to think technical competence was a desirable trait in a manager, but I'm starting to come around to the idea that technical competence doesn't really matter all that much and could even be a negative.
Instead, I feel that the most important trait in a manager is trusting the technical competence of his or her subordinates. The best managers I have ever had gave the team (and individuals) a direction to go in, and then sit back and let the employees work out how to solve the problem. They are always available to answer questions, facilitate communication, remove roadblocks, cut red tape, insulate the team from corporate politics, and provide resources to the team.
None of that requires technical skill, but loads of soft skills. I'll go so far as to say that having technical skill is often a detriment, as it is tempting for a manager to say, "I know how to solve this," and thus begins the slippery slope to micromanagement.
Change in an organisation comes from collaboration at minimum and cooperation at best.
It also needs a working alliance between leaders and followers. In teams with shared leadership, different people contribute to team goals in different ways or can shift between leader or follower over time depending on how the task evolves. I developed a concept called quantum leadership that tries to study and convey this point. It's a bit esoteric but I've experience it.
> Instead, I feel that the most important trait in a manager is trusting the technical competence of his or her subordinates. The best managers I have ever had gave the team (and individuals) a direction to go in, and then sit back and let the employees work out how to solve the problem. They are always available to answer questions, facilitate communication, remove roadblocks, cut red tape, insulate the team from corporate politics, and provide resources to the team.
This works as long as the subordinates are acting in good-faith. The primary reason why technical competence is important in a manger is to enable them to identify the subordinates that are bullshitting them.
> This works as long as the subordinates are acting in good-faith.
Presumably they are judged on their ability to perform their jobs instead of trying to read their minds and determine whether they're acting in "good faith."
And one would hope their managers do not call them "subordinates" and instead choose a more respectful moniker.
> Presumably they are judged on their ability to perform their jobs instead of trying to read their minds and determine whether they're acting in "good faith."
1. One example of not acting in good-faith would be spending more effort on looking like they are performing their job well than actually performing their job well. This makes it harder for a manager to judge people's ability to do their jobs
2. Another example would be spending effort to make it look like ones peers are doing #1, thus making it harder for a manager to properly detect #1.
3. Criticizing the term "subortinates" in a reply to my comment is a bit of a non-sequitur considering it was taken directly from the GP post.
I've read a lot of material about leadership, but rarely have I read on how to follow, something that is just as hard to do effectively.
I've come to a more complex position, where the task needs to align with the team for them to be leadable. With the right amount of incentives, any group can be pushed to deliver, but those incentives might differ by task (shit work needs different encouragement and support compared to creative work) and sometimes the needed incentives cost more than the end result is worth.
Understanding this alignment between task and team helped me triage how to approach the situation, or if it's even possible with the given people.
In the context of the reserach it's helpful to know they were all psychoanalysts. They focus on the human needs we acheive through work that are overlooked from the functional or material needs/outcomes of work.
In a modern context attend means would be more in line with EQ. Being about to sense, contain or facilitate someone scratching the itch of a need. It could be simply to help clarify, to hear, to reassure, to encourage, etc. All situationally specific based on the individuals in the group.
"What emerged from this exercise is that social class, education, gender and athletic ability were less important for leadership than the capacity for an individual to attend to others in the group."
If this experiment was done 80 years and they selected military leadership positions based on this experiment (as it implies by reduced failure rate measure), I don't understand why OP used "gender" in the summary statement above? Did UK actually field troops in WWII led by women?
> There were over 640,000 women in the armed forces, including The Women’s Royal Naval Service (WRNS), the Women’s Auxiliary Air Force (WAAF) and the Auxiliary Territorial Service (ATS), plus many more who flew unarmed aircraft, drove ambulances, served as nurses and worked behind enemy lines in the European resistance in the Special Operations Executive.
Thank you. Definitely not questioning if women took part in armed forces in UK. I've seen enough WW2 movies. It does makes sense in a wider sense of leadership. I took it for granted they tested only for combat related teams which prompted the question.
Though less of them, there were women combat related teams. Special Operations Executive was a special forces organization with women. British women were also in anti-aircraft units. The limitation was on frontline combat. They still engaged in combat on the home front and behind enemy lines.
I feel like I've suffered through this type of "leadership" for the past decade. When it works, it's great (although I feel you've kind of traded traditional authority for slight manipulation/scheming on part of the leader in such instances). These instances have also always involved more experienced team members.
When it doesn't work it's useless. Instead of just getting told to drive straight off a cliff you get to wander around in a forest for a while and when you eventually find out you're lost it's way too late.
As always (and as TFA points out) the person you pick for a leader still needs to actually possess whatever skill set your flavor of the month leadership style requires.
Suffer is the right word for when it's not carefully executed or understood.
Self organising or autonomous teams is not a replacement for strategy, vision or a clear task. That's leaderless organizations which isn't a good thing.
Who grants the authority to the team is a skill to test and develop. The point of the research is to find such leaders by what they actually do and how, not what they say they do.
The worst thing is that you end up in interviews and garbage like this is somewhat important. This is at most an opinion that matches poorly to the real world.
Hierarchies are ingrained in our genetic code. Its how most organisations have evolved for hundreds of millions of years starting from crustasians like lobsters. You might not agree politically with Jordan Peterson, but his book 12 rules for Life, is an excellent book and the first chapter talks about how hierarchies are fundamental to any social structure across most organims. There can be some outliers, but if there is no hierarchy to start with, a new hierarchy will most certainly evolve over long periods of time.
It's a fallacy that "flat" org structures have no hierarchy, they just become more opaque. A team that shares responsibility can have a dynamic hierarchy based on relevant capabilities at relevant stages of work. It's also important to note that I addressed teams in this article not organisations. To cultivate autonomous teams you also need to understand authority, where it comes from and who it's granted to and by whom. The more intentionally authority is understood the less likely power becomes a corrupting force.
This is really a bad strategy for promotion. Attending to others means not doing your own shit, which would compound as you help others more and more. Hence they would all be promoted whist you're on a PIP.
Makes sense though, basically need to balance individual desires and group goals. Helping the newbie get up to speed won't really benefit you, but will mean the team now can do more.
It's actually a great strategy to get promoted, and in fact, past a certain point, it's pretty much impossible to get promoted by trying to focus on your self-achievement rather than making your org (and surrounding orgs) as successful as you can by helping them.
Cultures that are creative and innovative in general are good at rewarding such behaviour. I kind of realise in a way I've only ever works in shared leadership situations as experts needed to contribute and be motivated to contribute to developing innovative products and work. Only through collaboration and generating their own motivation could I deliver and thus get promoted. However, I would say some who were more power orientated shot up faster at times, only to eventually fizzle out in the long run.
The article can be summed up by this line:
> What emerged from this exercise is that social class, education, gender and athletic ability were less important for leadership than the capacity for an individual to attend to others in the group.
So it's about how the best leaders are those who attend to the others in the group to create psychological safety so that those others can thrive without emotional risk.
The article also talks about how people seem to have forgotten this core concept but they haven't; it was the key conclusion to Google's own research into its most successful teams as well.[0]
If your team isn't primarily focused on providing psychological safety, then your leaders need to read Google's findings ASAP...
[0] https://rework.withgoogle.com/print/guides/5721312655835136/