Leadership communication catastrophes: Starcraft Ghost
The vast majority of business news, and business writing is utter tripe. No, let me restate that: almost every bit of “business news” you see on television, and most of what you read isn’t business news at all. It’s speculation news, mixed with investment tips (that are not uniformly tripe, but are mostly tripe.) Real business news is a rare commodity because the nub of almost every business story is proprietary, and usually hidden from any reporter’s sight by confidentiality agreements. If we readers ever do get the real story, or even a big chunk of it, that usually only comes years or decades after the fact. So the media usually just doesn’t bother reporting real business stories.
So it was my pleasure yesterday to read “Starcraft: Ghost: what went wrong”, by Patrick Stafford on polygon.com. Long story short, it seems that the smaller company Blizzard outsourced to, Nihilistic, was slow-boiled by a long string of “suggestions” that took their original stealth game far away from its original path. Eventually the game missed the console generation it was designed for as a result, and was scrapped.
Pardon me if I shamelessly use my loose interpretation of that article’s reporting as an excuse to illustrate leadship lessons that that I’ve learned painfully, albeit in more petty ways.
Let’s start with outsourcing. Outsourcing is a species of “delegation.” Ask someone who’s been a boss for any length of time about delegation and how to do that right, and you’ll likely get an emotional blast of some kind. I vividly remember getting a good answer on the subject from the owner of a steel smelter. This was early in my career, and as a far tinier sort of boss, I was running into problems with delegation repeatedly. This included sub-projects that had to be scrapped entirely after much work and investment because they violated the most basic directives the team or individual had been given. My father was an engineer. I can be precise. Disambiguating wasn’t the problem. I couldn’t figure out what was the problem. So I was eager to learn the magical secret of delegation from a real high mucky-muck businessperson. Instead, she said to my disappointment with great finality, “There is no such thing.”
And that’s the correct answer. Exactly. If by “delegation”, you mean “throwing a task over a wall and having it done according to your clear orders,” just forget about that. Humans don’t work that way. (Even Napoleon couldn’t get that to happen, ever, and especially on the road to Moscow, but that’s a tale for another day.)
The only thing that’s remotely like delegation-that-works is “management by walking around.” That is, constantly checking in on how tasks are going and spending just enough time to be sure you’d notice if something quite different is happening than what you need to happen. Verbal reassurances won’t do. as Since talk is cheap and our human capacity for self-delusion almost infinite, you have to visit subordinate people, teams and companies frequently. Meetings in your office won’t do, unless they can, say, bring all current work, blueprints, whatever, with them. Even then, if you visit them you may find important evidence lying in plain sight, so visit them (or at least their files and conversations on Slack.) Drop in on them, virtually or not, when and where things are happening. To put it another way, delegation is impossible but noticing isn’t.
That having been said, outsourcing is a particularly thorny kind of attempted delegation that illustrates a lot of common business and inter-team communication problems. In this case the stealth console game Starcraft Ghost was outsourced by Blizzard to a very much smaller game company with console experience: Nihilistic. Blizzard had a supervisor/representative/liason/producer (roles should perhaps have been clearer) involved, but weren’t able to have one such individual last through the whole project. For their part Nihilistic either weren’t given ultimate authority over the game or didn’t understand that they had ultimate authority/responsibility. So far this is troubling, but it also seems clear that Blizzard’s representatives weren’t told that they had to be very judicious in saying anything at all, to be fully up to date on progress before yapping, and to very, very clearly distinguish between suggestions and orders. The producers may not even have been properly and quickly reporting the suggestions/orders they were giving Nihilistic to their ultimate bosses as Blizzard, much less reporting any push-back. The result was that effectively nobody was in charge. Those suggestions by Blizzard’s representatives (that weren’t necessarily being reported to Blizzard at the highest level) became wrenching, game-delaying course corrections; and over a few years Nihilistic was slow-boiled, like a frog in a frying pan with the heat being turned up very slowly, as — suggestion by suggestion — it became impossible for Nihilistic to build the game they had originally proposed to Blizzard, and not long after that, to finish any game at all before that generation of console became obsolete.
Maybe the biggest part of being a boss is keeping your mouth shut, or holding your tongue even if it takes both hands to do that. You only have so much capital as a boss, if only because humans such as your subordinates only have so much attention, and can’t, logically, have a very large number of priorities. Every time your mouth opens, your capital diminishes. If your employees are about to find out what you want to tell them, shut up and let them bump into proof or the facts, and learn. If your way is better, but not that much better, shut up (you might be wrong after all.) Etc. The Blizzard reps Nihilistic dealt with (except one) were certainly perceived as being very loquacious.
Even more critically, anytime there’s a difference in authority, it should be accompanied by a very clear distinction between the boss’s 1) questions 2) suggestions 3) orders. With outsourcing, this is only more critical because more money is at stake and ties are looser.
I should say at the outset that there are good bosses who use questions or general remarks as a polite way of giving orders. There’s nothing wrong with that, so long as they are extremely consistent in how they do this; so that whenever they are instead genuinely seeking information and requesting an honest answer; that’s absolutely clear to their subordinates. There are lots of ways to mark the difference (as an academic linguist would say) — including tone of voice — but the difference has to be unmistakable. It’s probably best to mark real question by using either the word “question” or the word “if” in your first few words. For example: “I have a question, though…” or “What would happen if…”
Worse still, there are also poor bosses who think that asking questions of their subordinates drains them of power, and even truly catastrophic bosses who deliberately sow confusion and ambiguity so that their subordinates are repeatedly forced to ask questions, reinforcing their image of authority. Actually, those bosses are refusing to do their job supervising. May all such bosses end up in Hell without passing GO.
Suggestions and orders also have to be distinguished, and just as carefully: if the subordinate isn’t clear which is which, every time, without having to think about it, then the supervisor hasn’t done their job, much less done it well. “Theoretically clear” or “semantically clear” doesn’t cut it if you’re a boss. If you blur this distinction — a practice which is highly abusive but can seem empowering to bosses — you immediately take away all ownership from the subordinate (person or firm.)
Did I just say “abusive?” Yup, and here’s why, precisely: melding these two is a way to deliberately create a Gregory Bateson “Double Bind”: that is, a maddening dilemma in which no matter what the subordinate does, that subordinate can be labeled as wrong, disobedient, or even fired if if the results suck. Even and especially if they weren’t responsible. The effect of blurring orders and suggestions together is that it turns every utterance into an order — but one the boss can pretend was just a suggestion, or not even a suggestions, whenever they please, leaving their subordinates holding the bag for decisions they didn’t make. That’s abuse. In the short run, this is a strategy that can seem very appealing to psychopathic bosses, so it does happen a lot in business. But if someone farther up the food chain has a clue, it’s a doomed strategy for the psychopathic boss too, eventually. When you hear business leaders insist that “leaders have to have integrity for a business to succeed”, this is one of the principle nasty and common underhanded methods they are advising against. Real leaders make their orders clearly orders, and their suggestions clearly suggestions.
To summarize the logic of the situation: if an utterance is a suggestion, then the employee bears at least some responsibility for their decision to follow it or not. If it’s an order, the boss bears all the responsibility for that decision (as opposed to the detailed execution of that decision) however it turns out. Cheat on this, and you not only won’t have your subordinates’ loyalty, you won’t deserve it.
One more difference between orders and suggestions. If you’ve just given a subordinate an order, in most cases your boss has to be informed about that, explicitly and usually in writing; unless it’s a logically inescapable part of the orders you’ve already gotten from them. If not, you’re making a decision for your boss, and then not even keeping them in the loop about it. My impression, based on this report, is that a lot of this sort of miscommunication and missing communication seems to have been at the root of the Starcraft Ghost debacle. If the reporting is accurate, most of Blizzard’s reps to Nihilistic, seem to have felt that they had no ultimate responsibility for the results of their “suggestions” — suggestions which ultimately doomed the game — no duty to report the details of those suggestions up the food chain, and no duty to report the response to them, either. Nonetheless, this series of reps/producers seem to have felt that their suggestions should always be adopted — at a minimum, they weren’t shy of presently those “suggestions” very strongly.
If that’s what happened, no wonder nothing good can came of it: the middleman (sometimes just a “consultant”) who has the least at stake was making the most important decisions without either effective pushback or supervision. The smaller company was damned whatever they did (that’s the double-bind in action.) If they had strongly protested against such orders they could either have been summarily turfed and had to go looking for a paycheck quickly; or instead told there never was such an order, and that they were the ones responsible wasting time and money fulfilling what was just a stray thought. The impression one gets from disgruntled Nihilistic staff is that Blizzard’s producers were absolutely in charge unless and until something went wrong. Then Nihilistic was responsible. If so, the truth was that nobody was in charge.
The tragedy was that in the case of Starcraft Ghost, Nihilistic’s very competence at quickly turning every new suggestion/orders into impressive demos encouraged Blizzard to keep those “suggestions” flowing, even though the smaller shop just didn’t have the bandwidth to keep up with them, and didn’t agree with them.
A few extra notes:
I might also add that sometimes it’s useful to give vague suggestions, still: to avoid creating an abusive double-bind, then the rule has to be that vague orders can be fulfilled minimally; and more specific orders can then be given to shape the result further. However, as Grant found out at Shiloh, vague orders are almost never a good idea. After Shiloh Grant wrote far more extensive orders and spent a lot more time disambiguating, specifying which road, and which bridge, much more precisely.
Suggestions can be ignored, at the discretion of the smaller firm’s head. No justification is needed, explanations may be useful to avoid escalation to orders, but no justification is required (sometimes intuition derived from long experience is highly reliable, but doesn’t yield reasoning.)
It’s not a bad idea for the smaller firm to have a representative who spends time at the larger manufacturer consulting with peeps that outrank the big firm’s producer/liason. This ensures communications is in fact two-way. Not to have this in place AND not to have producer continuity, was very dicey.
Re entertainment, and refining games (or movie scripts.) Adjusting to real problems is necessary, but simply adding new features and cool stuff, isn’t. Scope creep is a horror; but making adjustments because customers are bored or confused isn’t — that’s strictly necessary. Features can be added infinitely, to any game, and new twists can be added to any plot: the real question is whether players are bored enough to notice that your game doesn’t have cool feature X or Y. To quote a screenwriter: story logic that contradicts itself and similar problems only matter if your readers/viewers are so disengaged they can trouble themselves about such details. But if they’re that bored, then story logic is probably the least of your troubles, you should be making bigger adjustments then that. Whereas if they aren’t bored, that you the writer notice a logical problem is immaterial: so leave the logic alone and submit your script, already.
PS — the book on Napoleon’s real leadership record is:
History of the Expedition to Russia by comte de Philippe-Paul Ségur
http://www.gutenberg.org/ebooks/18113
I highly recommend it, reading it will keep your expectations of what leaders can do realistic, which is a very good thing.
