Comment by daveungerer | Hacker News Reader

Comment by daveungerer | Hacker News Reader

Parent Full thread

Nowhere in the book is it pitched as a tool you roll out to a bunch of MBAs and hope that good things happen - it's never about out-of-touch managers demanding results. If someone cascades objectives without also cascading planning and estimation, are they not simply a bad manager?

What the book DOES say is that you should not create such a fear of failure that no-one will set stretch goals, and that OKR results should not replace individual performance evaluations. In short, OKRs are what you seem to think they are, plus the culture that makes it possible. If an organisation doesn't get that right, are they doing OKRs?

If your organisation does not allow developers to push back against impossible objectives, do you not perhaps have bigger problems that OKRs (or any other system) can't fix for you? Why would you blame OKRs - would any other system not also be painful when a company has become so dysfunctional? Also, did you know that all objectives don't need to cascade - some can be set bottom-up?

You're going to call No True Scotsman fallacy, and I'm going to call Straw Man, so I'm done with this thread. Perhaps I just fell for John Doerr's elaborately constructed fantasy that he uses to sell books. But right now it seems like a useful tool to keep a company on track, and I'll keep adapting it as the company grows.

I'll add to this thread that I was largely against OKRs, having seen them implemented in the style of "we need a quantitative measurement for success so let's make something up hastily", only to fall prey to Goodhart's Law[1].

But John Doerr's book introduced me to the bridge of concepts that I was not seeing: in a healthy setup, we are first and foremost focused on a qualitative objectives and THEN we attempt to model that fuzzy feeling with a quantitative measure (the "key result") that should reflect success. It takes several iterations in order to come up with a matching measurement, and even then we need to constantly re-evaluate whether the measurement is appropriate or if is devolving into a numbers game devoid of true objective.

In other words the full acronym is OAMBKY, or "Objectives, AS MEASURED BY Key Results". But that doesn't roll off the tongue quite as well.

So in that light, OKRs are a useful tool IF AND ONLY IF leadership—as well as the whole team—are focused on the philosophical, qualitative goal and are all aware that the measurement is only an imperfect proxy that is contantly re-evaluated to help us better assess the goal; not a goal unto itself. But that takes real leadership to drive that message (as well as avoiding setting up misguided incentives).

[1] When a measure becomes a target, it ceases to be a good measure.

> If an organisation doesn't get that right, are they doing OKRs?

I mean, this is kind of what the article is saying; just along a different dimension.

> Perhaps I just fell for John Doerr's elaborately constructed fantasy that he uses to sell books.

A lottery winner can write a book about their excellent saving habits and how they helped a lot. If you follow their advice, you may do quite well, and gain some wealth!

But the advice can also be largely unrelated to why they are wealthy; and people hoping to be multi-millionaires as a result of those habits might be disappointed. This does not mean that the lottery-winner's advice about being frugal and saving money is inherently bad. Go ahead and save, by all means.

Reply on news.ycombinator.com