Not for the first time, I'm engaged in a wider team effort to form up some Objectives and Key Results for the new year. It's the first week of February but as these things often go, it's the first opportunity most of us have had between the New Zealand summer break, folks returning from long annual leave, and generally getting through a lot of the other administrative tasks required at this time of the business year.
OKRs are notoriously difficult to do well. The idea is that you start with a company vision, distill an organisational strategy for the next 1 - n years (often around 3 to 5 years), and over each year, break down the objectives you'll need to meet in order to deliver on that strategy. There should be a clear line-of-sight that threads each of these tiers together in a congruent, coherent way otherwise people will be confused, distracted or disengaged. The quality of an organisation's strategy can be highly variable too! Whether you have a whole-of-company unifying strategy or your company breaks out large departmental ones (e.g. you have different organisational departments delivering on different markets), setting out the game board for the year ahead can quickly get muddled and unfocused if not directed well.
I'll often see things like:
- The list of things the practice will deliver : strategy is knowing why you're trying to make the things you make. Not just list some stuff because it's next on your Jira backlog
- Setting deliverable thresholds to meet : welcome to secular transition! If you reach your target, you can pack up, give everyone the rest of the year off secure in the knowledge that they won at businessing
- Short-term strategy : this is just furiously laying down the tracks from the front of the train as it hurtles through the year
There's a great quote from Alex Russell back when he used to be on the dumpster fire that is twitter:
You can tell that something somebody says is "our strategy" is, in fact, bullshit if you could reverse the order of the actions and outcomes and have it make just as much sense.
The challenge with implementing an effective strategy is making sure that all supporting departments are adequately included and have clear paths for feeding into the outcomes the company is seeking to achieve. Very commonly, you'll see your organisation produce a strategy that is heavy on marketing slogans but little practical substance, or highly product-focused with little clarity on the engineering contribution.
If you are wanting your teams to feel that they're effectively delivering high-impact work, that impact needs to be quantifiable. DORA metrics aren't the right tool for this. Impact isn't about your mean-time-to-recovery. It's about delivering changes that resonate with people outside the engineering practice: customers, sales and success teams, customer support... the people directly affected by what you're delivering. Examples I've seen elsewhere include:
- Number of new billable features released to production
- Number of projects shipped to production per team per quarter
- Number of technical investments per team per quarter
- Percentage of time on maintenance / remedial work
These measures have a direct relationship with a tangible effect on either the customers using those features or the proportion of available time engineers are being required to do "fix/glue" work instead.
Objectives and Key Results
Regardless of where you fall on the strategy side, any targets or measures you put in place need to be congruent! There has to be obvious, demonstrable evidence that a shift in the measure has a direct causative relationship with the objective you're seeking to influence. A strategic goal like "we deliver high-impact work to our customers" and setting an objective of "ship at least two customer-facing changes per week" has so many hidden layers of abstraction it might as well be written by ChatGPT or a random number generator. This is a vanity metric that feels great if you achieve it, but offers practically no value beyond the superficial act of achievement.
Effective OKRs tie directly to metrics that
- Provide context
- Have clear intent
- Guide action and learning
If you can't explain how your measures are related to your goals, you can't communicate a coherent plan to your team and they in turn will lose confidence in their ability to execute on that plan. Abstract OKRs introduce too many wild guesses and assumptions in the middle somewhere. Involve your people and bring them on the journey with you.