This is a conversation between Adam Pisoni, Stowe Boyd and me relating to a guest post I made to Brian Solis' blog, Impatience is a Virtue – What's Next for Social Business.
The conversation played out on email, which is ironic given that all three of us advocate "working out loud" unless confidentiality precludes it. I take the blame for emailing in the first place and hope to make up for the transgression by publishing it now. I have removed those conversational niceties that pepper emails, inserted some helpful hyperlinks and comments in square brackets by way of explaining some of the terms used and topics raised, and tweaked a few things to improve readability here.
Honestly, one of the most enlightening aspects of finally working within a real, big enterprise [Microsoft acquired Yammer in 2012] is the affect of performance management and budgets. Yammer loved to yell at our large customers to just change how they worked. That anyone at any level could affect change. But what you see inside large companies is that really good people will do all the wrong things either because they eventually feel pressured to optimize for what they are incentivised to do, or because their scope of power is too narrow to affect any change. This happens with budgets all the time. Two people in different parts of the org may have an idea that could make the company lots of money, but since the budgets were set up a year in advance, they can't shift the money between them.
The performance systems are far more insidious though and a wholesale rethinking of how they work is required. You see, whether you stack rank, or have some other system, all performance systems are run through the org chart. That is to say, your boss, and maybe his immediate peers, are up judging your performance as it relates to your commitments made too long ago to matter. As such, your boss can really only judge you on what she sees, and furthermore, has no real interest in encouraging you to do things for others that don't affect your boss because your boss has to show progress to her boss as well. I see this every day when people genuinely want to do something so simple as working cross-functionally with another team. Suddenly, the boss wonders what they're doing and how it will affect her part of the org.
I've had this discussion with lots of people and I haven't yet seen good models to get around this problem. I have lots of examples of where it's broken though. I think I've told you guys about the really innovating development model Yammer invented. It really is a revolutionary change to how we organize and build software, but it basically doesn't work in a hierarchical performance model because in our methodology an engineer's boss doesn't tell them what to work on. Longer story, but I agree with your premise [about co-opting performance management presented in the post] and think people need to start showing a big company can create a non-hierarchical performance system that successfully moves out bad people and recognizes good ones.
Adam - I'd like to hear more about the method.
Philip - Keep me posted. We can discuss the 'shared purpose and common values' next time we're talking in real time, too, because I believe fast-and-loose require dissensus ["active and directed dissent is a better way to counter the cognitive biases of groups and individuals, and to sidestep groupthink"] and reduced communications! Just like Adam's method, in which an engineering manager doesn't tell his direct report what to work on, laissez-faire management is about getting the context right so that people can work together or alone, based on what makes sense, and they won't spend a great deal of time justifying what they decided to do today, yesterday, and last week.
Adam, I know we share heartfelt dismay at the state of budgeting today. It's so incredibly awful that it's hard to think of one redeeming characteristic. And perhaps my "tapping the monster's own strength" [a reference to a jujitsu style tactic to change the nature of firms] demands a two-pronged attack, performance management and budgeting, although I'm confident we can co-opt the former more easily than the latter.
I hear your criticism that performance management is run through the org chart. That's not how it should be. It should be run through the organizational mission, objectives and strategy – what Kaplan and Norton refer to as strategy maps – and the Balanced Scorecard Hall-Of-Famers are therefore better at this than those who have no other guide for performance management than operational P&Ls and cashflow statements. We filter our prospective customers on the maturity of their existing business performance management.
Your phrase "your boss can really only judge you on what she sees" sits at the heart of my Trojan horse tactic. If we can wield today's big data / analytics capabilities to track, analyze and understand the flows of influence – many are intent on doing so out there in the social Web but not so much closer to home just yet – then we encourage what I like to call a sensory environment. It's not so much performance management in the old simple counting sense, but in the full biomimetic sense. Gone are the days when my boss presents my performance data to me, with her interpretation. She knows and I know and everyone already knows the nature and quality and impact of my and my team's contributions because we've applied information technologies to sensitize ourselves to it. And the organism responds to the nature of its sensory feedback, even monsters.
For example, one of my clients has fully accepted that its 2014 Plan must be supported by social and that the development of the 2015 Plan must be driven by social based on what 'it' will tell us through 2014. Early days I know, not least the idea of maintaining 12-month plans, but when I hinted that the 2015 Plan could be the last one, I got some peculiar looks. [A plan describes how an organization might execute its strategy, but strategies must be informed and bound by events, and not time bound. What has the orbital period of the Earth around the Sun – a year – got to do with business exactly?]
Stowe, I'm interested in understanding how "loose" you see "loose" being. At some point, the definition of an organization must refer to some common fabric. I frequently compare organizations with my home city of London. Now easily a thousand years old and still going strong, how many of my prospects / clients are so confident in their organization's longevity? And if they cannot be confident, what's the difference between the city and the company? Perhaps a certain looseness.
I'm not sure if you've had time to read Attenzi - a social business story, but I'll skip straight to a question I pose in the final chapter ...
Do you help all the individuals associated with your organization (employees, customers, partners, suppliers, shareholders, etc.) build worthwhile relationships with each other and others, coalescing by need and desire, knowledge and capability and shared values, to create shared value?
The ramifications of the answer are such as to redefine some of the well-known phrases invoked in the question. Again, it references this idea of sharing something if only because if they don't then why consider yourself part of the same organization?
But returning a moment back to the next year or five, I wonder if you agree with my intent to tap the monster's own strengths? Because right now, "dissensus and reduced communications" (by which I read "explicit communications") is perhaps so far removed from today's organization as to fire up its antibodies.
The missing piece is explaining tactically how you can have a system of checks and balances without the org chart. For example, I believe companies should constantly be trying to raise the bar for talent. That means helping people improve as well as moving out people who do not want to or do not have the capabilities to improve. We deal with that today with the performance management process run through the org chart. Your boss has the authority to fire you and is supposed to be the judge of your contribution.
Say a company recognizes how that system is broken. They want to work more like a network. What do you tell them to do? Stop all performance discussions with employees? Stop giving raises?
You suggest using "influence" as proxy metric for contribution. I'm assuming you think one could do network analysis to determine one's influence and reward accordingly. Theoretically it could all be automated. No bosses needed. Of course, the problem with applying proxy metrics to humans is that they tend to optimize for the metric as opposed to what actually provides the most value. [People perform as they are measured, so defining such measurement demands great care.] For example, it would be easy for a person to completely optimize for the breadth of people they touch within the org, but manage to do so to the detriment of their existing team.
If we're going to help companies grab the next rung, we have to be more explicit about what that rung is.
As far as our development methodology goes, I have a more detailed write-up we're working on if you are interested, but here's a talk my VP of Eng gave on it [YouTube video embedded here].
I just wrote a piece about prosocial bonuses, which are based on circumvision instead of supervision, a distributed decentralized mechanism for aligning rewards with higher performance and better behavior. The same approach can be generalized to advancement, team formation, and so on.
That's a fairly specific set of recommendations.
Not specific enough, but I generally agree with having performance be a measure of the "customers" or "team members" rather than the boss. The problem is in creating the system that has that happen in a fair and balanced way. I do not believe we've found adequate proxy measures to do it in a more objective way.
Though just to state it again, I think starting with the performance systems is absolutely the right way to go.
I don't think it needs to be 'objective' – which generally means something uniform and managed. We shouldn't be afraid of the subjective will of the swarm. What is important is that those that are distributing the prosocial bonuses are happier and more engaged because of their involvement and realization of the third way itself [a term Stowe uses to describe a future way or working]: not individual, or team, but swarm-based reinforcement of cultural norms.
Do we have any examples of large enterprises who took their performance management system and turned it into something more like we are describing?
Just emerging from the conclusion of a client project. Apols for radio silence. Talking of performance metrics, this particular client once talked about "growing metrics of engagement" in the same breath as saying how important it was to drive "call center calls per hour". I think they've come a long way this year, but tapping back into this thread reminds me how much further there is to go.
I assisted the complete rewrite of WOMMA's Influencer Marketing guidebook earlier this year. I always define influence as a change in opinion or behaviour. The guidance proposes: "The ability to cause or contribute to a change in opinion or behavior."
The english language is ambiguous – it is something that happens and it is the potential one is attributed to make it happen (making the assumption that past performance is a good predictor).
I still defer to my favored definition because it's the very outcome that so-called influencer marketers seek. Any quantitative attribution of influence being 'possessed' by an individual can only be qualified by looking for this outcome, which is where Klout makes no sense (and to your point earlier Adam about gaming the proxy). This distinction has been recognized in the guidebook however: "There are two distinct states of influencer measurement that are relative to the point in time an influencer marketing program begins: 1. The potential to influence (before); 2. Actual, observed influence (during/after)."
Nevertheless, objectively identifying when any one individual or team has exerted influence isn't easy. For example, the result may be an absence... a decision not to do something. And of course not all influences are equal in sign or magnitude, however one might adjudge the value. While we can machine an increasing quantity of digital exhaust, and even design to kick off more, I don't believe this is yet sufficient to proxy an appreciation of influence flowing through an organization... but fortunately, we're pretty well evolved to make this assessment ourselves. It's part of what makes us social animals and helps our communities prosper. I'm not a trained sociologist or psychologist, but I am aware that these assessments, when elicited, can be colored by other considerations, not least emotions and non-work related connections.
Perhaps this noise is reduced when the elicitation is quick and in-context; real-time. Yammer's 'like' button for example.
When you ask about "checks and balances without the org chart" Adam, I think the answer lies in making the outcome of this assessment transparent, although this runs the risk of regressing to the mean – a potential downside of crowdsourcing.
Just pulled these out of my bookmarks fyi. The third and last is a little tangential to the topic here, but I know you'll appreciate the connection.
- From the Atlassian blog – Radical change to performance reviews wins Atlassian Prestigious Human Capital M-Prize
- From the HBR blog – Crowdsource your performance reviews
- From Wired – Slime mold grows network just like Tokyo rail system