Sheldon's Software

Google changed its performance review system, and chaos was everywhere. The new system was called GRAD (Googler Reviews and Development), and after it was announced, there was a boatload of questions. ICs asked their managers, managers asked their directors, directors asked their VPs, VPs asked the CEO, and the CEO deferred to the GRAD team, whose documentation was incomplete. Trying to change course for a company of over 100,000 people is no easy task.

Despite the uncertainty, I looked at the situation, cut through the chaos, and received the highest performance rating at Google and a 14% raise. Only one other engineer and I in my director’s organization received this rating. My secret was this one weird trick (and my manager loved it).

But before I can tell you what I did, we have to go back.

The Rubric

A year ago, when GRAD was announced, I examined the rubric for my role, L5 software engineer. The rubric had attributes relating to contribution, challenge, leadership, and teamwork.

As I planned projects over the next year, I kept this rubric in the back of my mind. How will each of my projects demonstrate each aspect of the rubric? How can I fulfill my responsibilities and go the extra mile?

And every week, I had a 1:1 with my manager where we discussed exactly this.

Using 1:1s to my advantage

The first item on our recurring 1:1 agenda was always GRAD. I shamelessly asked two questions:

• What is my expected GRAD rating?
• What can I do to get to the next rating? (Or at least, to increase my chances?)

Then we talked about the same strategic 1:1 topics that I used leading up to promo.

Some people are scared to talk about performance reviews. Pardon my language, but that’s a load of baloney! Your performance rating should never be a surprise. You should be in the driver’s seat, fully knowing how you’re performing against your manager’s expectations.

So, I formally wrote my expectations and agreed on them with my manager. We reviewed my projects and progress against the expectations periodically. Occasionally, we would look at the rubric together and rate myself against those unbiased, unfeeling sentences.

Then, after our 1:1, I doubled down on the things I was doing well. I found ways to compensate for my weaknesses. I worked on projects that would fill the gaps in my performance evaluation. I put thought into what projects I would work on, not just the work itself.

And at the end of each week, I wrote a few important things down.

Weekly Snippets

Every Friday, I would add my most notable pieces of my work to an internal website known as Snippets (though a regular Google Doc would work just fine). I linked to design docs I wrote, large bugs I closed, and major presentations I gave. I put in costing worksheets I crafted, production incidents I handled, and projects that I launched.­

It barely took me 5 minutes a week. And at the end of the year, I used my snippets to easily summarize my work. Nothing fell through the cracks.

But weekly snippets weren’t the one weird trick that gave me the top rating. There was one more thing I had to do.

Measuring Impact

At Google, there’s a feeling that people don’t get rewarded for doing necessary work that’s difficult to measure, like maintenance or polish. The company has so many products because people want to get promoted, and launching a new app is a quick way to show something shiny to the promo committee.

Sometimes this culture irritates me because I know there are changes that ought to be made to improve the user experience, but I know they won’t move top-line metrics. So those issues sit in the backlog forever, waiting for the next intern or fixit to be hopefully picked up like a forgotten puppy at an animal shelter. GRAD is supposed to fix some of this, but institutional memory will be hard to change.

So, Googlers have to measure everything. If you launch a feature and don’t have any metrics, there’s no point to launching.

Anyways, I intentionally measured my work’s impact wherever I could. I advocated to improve our logging and to automate metric analysis. And when my teammates proposed features, I asked questions like:

• What’s the problem we’re trying to solve?
• What’s the impact of making this change?
• How will you know if you succeeded?

Occasionally, these questions saved us time because we didn’t implement features we didn’t need. One of my teammates in particular really appreciated these questions because it made him think about the big picture.

Now that I had a list of projects and the impact each one made, I was finally ready to open the GRAD website.

The One Weird Trick

Here’s my secret: I wrote my own performance review, and I sent it to my manager and his manager for their feedback ahead of the deadline.

In GRAD, your manager is supposed to write your performance review. I didn’t want to rely on my manager’s memory–there were simply too many things he could miss, or simply not understand. He has many people to manage and I’m just one member of his team.

So, as the deadline for performance reviews approached, I took another look at the L5 SWE rubric. I copied each attribute from the rubric into a blank Google doc. Then, I went line by line, selecting evidence from my weekly snippets that best demonstrated each aspect of the rubric. For instance, if the rubric said, “manages projects spanning multiple quarters,” I would link to a large, complex project that I led that spanned at least two quarters. I repeated this process down the line, adding links to design docs, project proposals, costing worksheets…­ in all, it took me about an hour.

Once I had my informal self-assessment done, I sent it to my manager and his manager. I asked if there’s any other evidence they would need to bolster my case during performance reviews. My manager basically said “This is helpful, thank you.” (It also helps that I have a good relationship with my manager and my skip-level manager.)

A few weeks later, my efforts paid off. My rating was “Transformative”–the top rating, only assigned to 6% of employees. This translated into a bigger than usual annual bonus and an above-average raise.

After ratings were released, I read what my manager wrote in the official GRAD review. At the very bottom of the report, there was a line that said “After talking with Sheldon, I realize that he deserves a transformative rating due to…” This is evidence that I was originally slotted for a lower rating, but my self-assessment convinced my manager and the committee to accept the highest rating. My skip-level manager also admitted the doc helped push me over the threshold.

In the official review, I was praised for improving reliability, collaborating with other teams, mentoring Nooglers, and directing projects. I was genuinely proud of accomplishing so much with just a core team of myself and two people–I put in a lot of effort into my day-to-day work. But what surprised me most was that my efforts around diversity and inclusion were mentioned positively. I barely did anything except go to a few DEI meetings and announce events at my team’s weekly meeting. I just wanted to learn and to foster a good environment for my team. Part of GRAD includes intentionally recognizing DEI efforts, and that’s one positive change over the old system.

While all this chaos was going on, I was not attached to the outcome. Given the uncertainty around the new review system, no one was sure how ratings would be assigned, not even managers. People were especially unsure about the boundary between each rating. Last year, I got a “Strongly Exceeds Expectations”, the 2nd highest rating. This year, I was doing even more far-reaching work, so I should have maintained or increased my rating. Therefore, the highest GRAD reading was a stretch goal. The 2nd highest might have fit. The 3rd rating would’ve been an insult.

It is, however, better to ask for the highest rating you can reasonably get, and have your manager find evidence why the rating should be lower, than to argue your case upward. I guess a performance review is like a negotiation. A very long, drawn out, convoluted negotiation, but a negotiation nevertheless.

Why I wrote this article

I didn’t write this article to brag. I wanted to demystify the system, show you how to navigate performance reviews, and improve your chances of a positive outcome. And if you really don’t care about performance reviews, that’s fine. You have to see the game being played around you before you can choose to play.

Personally, I’m okay with spending time to gather evidence for a raise. A raise will benefit me for the rest of my life. Besides, this was all done during business hours as part of the review process, so I don’t see it as time wasted. Call me cynical: it’s called a performance review and not a work review for a reason.

On the other hand, some people manage to get great performance ratings seemingly without doing anything extra. And there are lots of other things to focus on in your career or in your life outside performance ratings and your salary.

With last year’s GRAD behind us, it was time to write expectations for the next year. Out of curiosity, I graded myself against the next level SWE rubric, and found that I could be hypothetically slotted into the middle rating at L6. Currently, I’m expanding my responsibilities and influence to get to the next level. I’m not in a rush, but I think I can be promoted within the next two years.