There will be caveats: Warming cautiously to OKR

From the Agendashift Slack a few days ago:

Early morning crazy thoughts spoken out loud:
Wholehearted: bringing OKRs to life with Agendashift
A workshop based on and expanded from edited highlights of the core Agendashift workshop and the outside-in strategy review

Why “crazy thoughts”? The background: we’ve been discussing Objectives and Key Results (OKR) [1] in multiple corners of the Agendashift Slack in recent weeks (channels #wholehearted-x, #bookclub, and #strategy) and I didn’t hide my nervousness.  Isn’t OKR just Management by Objectives (MBO) rebranded, with all the dysfunction [2] that goes with it?

To cut a long story short (two books later), it’s clear now that Agendashift  – outcome-oriented change and continuous transformation [3, 4, 5] – and OKR are a great fit – so great that they almost seem made for each other. As recently as this week, Karl Scotland blogged about the combination (more accurately he blogged about Agendashift and 4DX, but the distinction isn’t that important):

I do still have reservations. It was – shall we say – an interesting experience reading (or listening to) Doerr [6] and Wodtke [7]. Good books both, but page after page (hour after hour) my frustration would grow before my concerns would finally be acknowledged. Given the acknowledged risks, it seems clear to me now that OKR has something in common with Agile process frameworks: how you approach the framework matters very much more than the choice of framework itself. Depending on your point of view you may find that thought helpfully subversive, heretical, or commonsense; as for me, I’ve staked my career on it.

So to my caveats. Coming from where I’m coming from, they’re significant enough that they should treated not as footnotes but up front as design principles. OKRs must be:

  1. Respectful of diversity and autonomy at individual and team level
  2. Agreed through meaningful participation
  3. Executed knowing the difference between implementation and experimentation

Caveat 1: Respectful of diversity and autonomy at individual and team level

An objection sometimes laid at the door of OKR is that it’s all about alignment, and that the goal of alignment is to bring about some kind of monoculture. I reject this as a strawman argument; the goal of OKR is to provide enough direction that the organisation isn’t destroying its ability to get things done because its different parts keep pulling in opposing directions. For most organisations, too much alignment would be a nice problem to have, and address that very common issue well, great things can happen. In practice, key results (the KR part of OKR) aren’t long lasting (they work in timeframes ranging from days to months), and even many objectives (the O part) don’t last for more than a quarter; good luck creating a monoculture that quickly!

That argument dismissed, it’s worth remembering that OKR is a tool for strategy deployment [8], not operations management, and it’s explicit that existing operations must continue to perform well even as they undergo change. Resilient operations in an unpredictable world depends on diversity (you need to be ready to respond in different ways to respond as both conditions and internal designs change), and only a fool would seek to destroy options for the sake of consistency. Technically, we’re in the world of Ashby’s law of requisite variety [9]; colloquially, power is where the options are. If you can, why not create that power everywhere?

But that’s just the technical argument. Take away from people and teams their ability to create and exercise options and you destroy their autonomy. With that you destroy their engagement – and then it’s game over if what you need is their energy and creativity. So how then is strategy deployment meant to work?

Caveat 2: Agreed through meaningful participation

The textbook answer to this conundrum is that OKR works both top down and bottom up. Some objectives come from on high, with lower levels defining their own objectives and key results to suit. Others bubble up, high level objectives somehow summarising (blessing?) what needs to happen lower down.

I’ve long since abandoned this “Top down vs bottom up? It’s both!” thing. It’s a cop out that does little to help the inexperienced manager and may put even the experienced manager in a bind; small wonder that middle managers can be a miserable bunch (I’ve been one, so I know). Middle out is no help either; as with the iron triangle, it’s time to recognise that these metaphors make little sense in open-ended and high feedback contexts. Also, they are hierarchical in a way that’s quite unnecessary, and clinging to them just gets in the way.

My answer – and it comes from an area where Agendashift excels – lies in participation: facilitating challenging and meaningful conversations about obstacles and outcomes (and progress thereon), making sure that they take place frequently both within and between strategy, development, and delivery, and have diversity of representation in terms of both functional responsibility and seniority. In place of top-down imposition, authentic agreement on outcomes becomes the basis for change. Where in the past innovation and intelligence would become increasingly diluted and distorted as news passed up the chain, now we create frequent opportunity for rapid and informed responses.

Of my two most recent books, authentic agreement on outcomes is a key theme of Agendashift. My latest book, Right to Left [10] explores the implications for organisation design and leadership in much greater depth, in the final two chapters most especially.

Caveat 3: Executed knowing the difference between implementation and experimentation

A common lightbulb moment for participants in Agendashift workshops comes when we organise outcomes using the Cynefin Four Points Contextualisation exercise. We dare not speak its name up front – it rather spoils the surprise – so we go by the pseudonym “Option approach mapping” initially [11]:

mapping

The key insight is that not all outcomes are alike. Easily recognised (and all shades and combinations in between these extremes):

  • Some are uncontroversial and don’t need digging into, regardless of whether they’re to be done right away or kept for another day
  • Some you’re confident can be achieved reliably, but first they will need to be broken down by someone who knows what they’re doing
  • Others can be approached in different ways, but no single approach (or combination thereof) is guaranteed to deliver the outcome in its entirety; consequently we’re in the land of iteration and experimentation
  • Sometimes, where to start and even who to ask is beyond current knowledge

Reading/listening to OKR’s fascination with stretch goals, you’d be forgiven for thinking that the only thing in question with each one is whether we can do it in the time we’ve set for ourselves. If that’s your only hypothesis, much opportunity for learning – about customer needs as much as organisation capability or technical possibility – will be missed. Moreover, choosing a sequential approach when an iterative one is needed (or vice versa) is a costly mistake to make – costly not only in time and money but in reputations too. The books do get there in the end, but honestly, I feel they could do a lot better. For some balance on issues of complexity, I’d suggest pairing Doerr with McChrystal [12], and Wodtke (which seems to be aimed at the startup community) with a good Lean Startup book, of which my favourite is Maurya’s [13].

So what are we left with?

Ten years ago I saw my employer, UBS, nearly destroyed by the scandalously ill-chosen, ill-managed, and under-informed pursuit of the wrong goals (the recommendations of a benchmarking exercise conducted by a big name consultancy), so I speak from the heart here. But I’m not warning you against OKR – in all honesty I’m really warming to it.

My caveats take nothing away, because I don’t think I’ve said anything contrary to the literature, albeit that it takes a long time getting round to it. So a few pointers:

  • Find groups of people – let’s call them circles –  who share (or should share) some common objectives. Give them the opportunity to explore thoroughly their landscape of obstacles and outcomes, decide what’s important, and set some priorities. Agendashift is the manual on that! Expect them to track progress and revisit both their understanding and (accordingly) their plans on appropriate cadences.
  • Look for overlaps between circles, and where they don’t (a single manager isn’t enough), delegate people into the intersections. Not only will the conversations here be a lot more interesting and challenging, but we’re very obviously creating opportunities for both alignment and mutual accountability. A wider organisation listening not just for progress but for learning here will be sending a powerful message (not to mention learning itself – if it’s listening).
  • Metrics can be great, but don’t reduce it all to numbers. I’d argue that the “Measure what matters” in the title of Doerr’s book is a little misleading – certainly it deterred me for a while! Moreover, and as Doerr rightly emphasises, it would be a catastrophic mistake to connect OKRs with individual compensation (Drucker’s plausible but ultimately disastrous error with MBO).

If you’ve read Right to Left, you’ll know where the above comes from. If you haven’t, put it on your list. Doerr, Wodtke, and my recommended pairings too! Mercifully, mine isn’t too long, so you might want to start there 🙂

[1] Objectives and key results (OKR) (wikipedia.org)
[2] Management by Objectives, Arguments against (wikipedia.org)
[3] About Agendashift™ (agendashift.com)
[4] Agendashift: Outcome-oriented change and continuous transformation (2018)
[5] Agendashift partner programme
[6] Measure What Matters: OKRs – The Simple Idea that Drives 10x Growth, John Doerr (2018)
[7] Radical Focus: Achieving Your Most Important Goals with Objectives and Key Results, Christine Wodtke (2016)
[8] What is Strategy Deployment (availagility.co.uk)
[9] Variety (Cybernetics) (wikipedia.org)
[10] Right to Left: The digital leader’s guide to Lean and Agile (2019)
[11] Agendashift in 12 icons
[12] Team of Teams: New Rules of Engagement for a Complex World, General Stanley McChrystal et al (2015)
[13] Running Lean: Iterate from Plan A to a Plan That Works, Ash Maurya (2012)


What if we put agreement on outcomes ahead of solutions?

Agendashift™: Serving the transforming organisation
Agendashift  Academy: Leading with OutcomesHome | Store

Links: Home | Subscribe | Become an Agendashift partner Events | Contact | Mike
Resources: Tools & Materials | Media | Books | Assessments
Blog: Monthly roundups | Classic posts
Community: Slack | LinkedIn group | Twitter

Leave a comment