Apex Legends
These days, Atlassian does two formal performance evaluations each financial year; one halfway through (i.e. January-February) and one right at the end (i.e. July-August).
The entire process is known internally as APEX, formerly OneCycle. Personally, I liked the unofficial name of Bi-Cycle, but you can't win them all I suppose.
Anyway, it's February right now, and while we're not quite done with the whole shebang, we're close enough that some reflections are in order.
Starting right from the beginning.
First Blood!
The first step in the APEX process is for everyone to do a self-assessment.
The intent of this is to give the person being evaluated a chance to use their own words to summarise their activity over the evaluation period, while also providing information to their manager and filling in any gaps.
This was always an informal part of the performance evaluation process, but it's now a required activity for every single person in the company. Well, except for the founders I suppose. I don't know what they have to do.
The self-assessment can be broken down into a few major elements:
- A reflection on your goals
- An overall summary of the value you delivered
- A more detailed per-pillar breakdown of the value you delivered
- Any ideas about what you want to focus on moving forward
The goals reflection is straightforward. Everyone is expected to set goals at the start of the evaluation period and reflect on them throughout. This section is mostly just a summary of the reflections that have already happened over the period, along with a critical evaluation of whether or not the activity or progress on the goal was below, at or above expectations.
The overall summary is an opportunity to tell a short story describing the major things that you delivered across any pillar during the evaluation period. You don't want to try and list every little thing here, and in fact, you can't, because it has a pretty strict word limit. It's meant to frame the overall situation.
The per-pillar breakdown is where the overall summary is expanded, and you have a lot of freedom to describe the various things that you did over the quarter that aligned with each of the performance pillars.
To refresh your memory, those pillars are:
- Project Impact (Team Impact for managers)
- Craft Excellence (People Management for managers)
- Direction (Collaboration for managers)
- Organisational Impact
If you want more detailed information about each of those pillars, I suggest you read this blog post from November last year. It's filled with all sorts of goodness and is way too long to repeat here.
Finally, the ideas for moving forward is an opportunity to self-identify growth areas or things that you would like to focus on, in order to help your manager plan and help you grow. It doesn't really factor into the evaluation, but it's useful all the same.
Personally, I found the self-assessment to not be that bad. I keep good records, so summarising all of the various things that I'd been up to was relatively straight forward. There is a bit of implicit pressure to present yourself in a positive light, but I did my best to resist that in the spirit of Open Company, No Bullshit.
I know it wasn't as straightforward for other people to fill out though, especially those who don't write as much as I do. In fact, this time around, I spent a decent amount of time reviewing self-assessment drafts for people and coaching them on how to write more gooder.
I'm of the opinion that the self-assessment is a good thing, as it makes each person an active participant in their evaluation instead of an observer. Like most things, it will get easier with time and repetition.
Ring Movement In Progress
The second step in the APEX process is for the Engineering Managers to write up calibration notes for everyone person that is being evaluated.
The intent of these notes is to objectively summarise what value the person delivered and how they went about delivering it, such that the information could be read and understood by another manager somewhere else in the business.
The calibration notes can be broken down into the following components:
- An overall rating
- An overall summary
- A per-pillar rating
- A more detailed per-pillar summarisation
The overall rating is the suggested rating for the person in question, based on what the manager currently knows. It is one of:
- Does Not Meet Expectations (DN)
- Meets Most Expectations (MM)
- Meets Expectations (ME)
- Exceeds Expectations (EE)
- Greatly Exceeds Expectations (GE)
Again, for more information about what those ratings actually mean, I suggest you go and read this blog post from late last year.
The overall summary is slightly different to the one in the self-assessment. In the calibration notes, it's meant as more of a framing device to help another manager understand the overall rating. It typically takes the form of a few dot points, each one prepended with a -, = or + to indicate if it was below, at or above expectations for the role and level.
The per-pillar rating is an indicative rating for the pillar in question. There is no strict formula that is used to calculate an overall rating from the aggregate pillar ratings, and some pillars are more heavily weighted than others, so this is primarily intended to provide context.
The per-pillar breakdown is where all the detail lies. It summarises the major contributions that the person made by pillar and the information therein directly drives the per-pillar rating. Similar to the overall summary, each dot point here is prepended by a -, = or + to indicate whether the item indicated below, at or above expectations for the role and level.
The writing of the calibration notes was where the bulk of my effort went, because it was where the majority of the actual performance evaluation took place. It involved drawing together a huge amount of information, picking the bits that best represented the person in question (for good or ill), making judgements about those items and then summarising them appropriately.
Not only that, but each item in the notes needed to be written in such a way that a person with very little context about the situation could understand it and judge whether it was an appropriate indicator of performance (or lack thereof).
I had ten sets of calibration notes to write, and I would say that each one, including research, took somewhere between three and five hours to complete.
It was literally all I thought about for about two weeks, and I was completely and utterly exhausted at the end.
However, that doesn't make it bad. The process itself was well designed, there was plenty of guidance about how to write good notes and we had access to good tooling to make the whole process simple and straightforward.
Similar to the self-assessment, I think the existence of calibration notes is a good thing, especially when it comes to the next step in the APEX process.
Ring Closing
Speaking of which, that next step is calibration.
The intent of calibration is to ensure that we evaluate people consistently across the business. If someone in one team gets a Meets Expectations, then other people who get the same rating should be approximately equivalent in terms of value delivered and behaviours exhibited.
The actual calibration process involves a bunch of managers sitting together in a meeting and working through the notes for groups of similar people (by role, level, etc). They absorb the information, ask questions for clarification and generally judge whether or not the evidence provided is enough to justify the rating given.
That's pretty much all it is. A lot of reading, a lot of questions and a lot of hard decision making when things don't quite line up.
This isn't the first time I've done a calibration activity, but it was by far the smoothest it has ever been, and I put that down to two things.
The first was the existence of calibration notes. In prior calibration rounds, we had notes, but they were written to be consumed by the person, not other managers. Changing the intended audience (and standardising the format) made the raw information much easier to consume, and more importantly, compare.
The second the existence of good tooling. In the past we've used a weird combo of Google Sheets (lists), Confluence (notes) and Workday (ratings), but this time we had a purpose-built internal tool. It meant that we were able to do everything in a single place, dramatically improving our effectiveness. From my perspective, the time and money invested in that tool was definitely worth it.
We Have Our Apex Champions!
Calibration isn't the last step in the APEX process.
The last step is taking the calibration notes and the agreed upon rating and summarising all of the information such that it can be delivered to the person who was being evaluated.
That hasn't happened yet though, so I can't write about it.
I don't foresee it being particularly challenging, as the raw information is all easily available and all of the actual decisions (aka ratings) have already been made and approved.
As for the experience of executing the APEX process itself, it was exhausting from start to finish, but it was definitely an improvement over the last time I did it.
I'm sure we'll be even better at it next time.
Well, other people will anyway...