Case study · Google (via Accenture) · 2022–25
Editorial standards for a global writing team
Inconsistent, uncoordinated writing was making editorial review a bottleneck. At worst, it risked embargoed launch documentation going live early. Here is how I built a system to fix it without new tools, more headcount, or slower deadlines.
A note on confidentiality. This work was done at Google via Accenture and is covered by an NDA. Product names, internal metrics, and source material are omitted; any figures here are relative, not absolute. What follows is how I approached the problem, which is mine to share, not the confidential work itself. The sample artifact below is rebuilt for this portfolio.
The problem
As a lead technical writer on a help-center content team, I sat at the end of an editorial pipeline that was quietly breaking. Writers spread across India and the Philippines produced help documentation for product and feature launches, and by the time it reached me for review, three problems compounded. The work was inconsistent from writer to writer, so review became a bottleneck: nearly every piece needed heavy correction before it could ship. Newer and global writers repeated the same mistakes, because there was no shared reference for the decisions we kept making. And work was distributed blind. Some writers were drowning while others had bandwidth, and no one had visibility into who was carrying what.
The people who felt this weren't abstract. Content strategists were staring down fixed launch deadlines. QA kept bouncing work back for the same recurring issues. And I was the last line of review, which meant the bottleneck was me.
The worst-case consequence wasn't a typo. It was an embargo break. Launch documentation that went live before its product was announced would be a confidentiality failure, not a quality one. That risk is what turned "our editing is inconsistent" into something worth building a real system around.
The constraint
I couldn't solve this the easy way, because almost every easy lever was gone. The team was global and asynchronous: half in India, half in the Philippines, on shifts that rotated between days and nights week to week, with only a small overlap. Gathering everyone in a room, or reviewing in real time across the working day, was never an option.
I was also inside Google's stack: their CMS, Workspace, their style guide, their process. I couldn't buy or build new tooling to fix this. Whatever I made had to live in the tools we already had. And I couldn't move the deadlines, because product owned them. That would have been survivable, except the priority system had collapsed. Everything was supposed to be ranked P3 to P1, but over time every request drifted to P1, until "priority" told you nothing at all.
The hardest constraint was the one that sounds smallest: the style guide itself. It was non-negotiable, but its rules sometimes competed with each other, and resolving those conflicts took judgment, the kind you can't hand someone in a rulebook. A junior writer treats the style guide as a list of rules. The real skill is knowing which rule wins when two of them collide, and that was the thing I couldn't simply write down.
A junior writer treats the style guide as a list of rules. The real skill is knowing which rule wins when two of them collide, and that's the thing you can't just write down.
The decision
The obvious move was to become the fix myself: review everything, coach everyone, be the quality gate. I rejected it, for two reasons. It wasn't sustainable. I was already the bottleneck, and stacking coaching on top of review would have made me slower, not the team better. And it kept the team permanently dependent on me: nothing would improve structurally, and the moment I stepped away, quality would collapse. It also wasted the judgment my senior writers already had. I took the longer view, because I wanted those seniors to become leads with their own teams one day.
So I built for that instead:
A living checklist for the recurring, teachable mistakes: a shared reference for the decisions we kept re-making, so the same corrections didn't have to happen twice.
One standing meeting at the single hour both shifts could make, where the competing-rule cases a checklist can't hold got discussed in the open. Codify what's codifiable; discuss what isn't.
Bi-weekly office hours where senior writers coached juniors. Near-peer teaching worked better for the juniors, and it was a rehearsal for the lead roles I wanted the seniors to grow into.
Push-back on priority, using deadline and volume as evidence so the labels meant something again. When too much landed at once, I farmed high-priority work out to other product teams rather than let it crush us.
When something went wrong, we ran a post-mortem and fed what we learned back into the checklist, so the reference got sharper every time it failed.
- Reading level: confirm grade 7 or lower before staging.
- Voice: second person, active, present tense.
- Jargon: replace product-internal terms with a user-facing equivalent on first use.
- Sentence length: flag anything over ~25 words for a rewrite pass.
- Embargo: no launch-dated article moves to "ready to publish" without a confirmed go-live date.
- Links: verify every cross-reference resolves to a live help article.
What happened
The measurable things moved in the right direction: QA sent back fewer pieces, and review time dropped. But the structural win mattered more. Nearly every senior writer I coached had become a team lead by the time I left. That's the outcome I'm proudest of, because it means the system outlived me.
What I won't claim is that I solved the bottleneck. I didn't. Work volume kept climbing, and throughput never fully caught up.
If I ran it again, I'd push management to promote my trained seniors into lead roles faster. I waited too long to make that case, and it kept me in the bottleneck longer than it needed to. Promoting them sooner would have freed me for the higher-leverage work: training leads, and building AI-assisted workflows to QA articles more efficiently, which is exactly the direction I'm teaching myself now.
“Through her deep understanding of our workflows and style guide, she helped develop the specialized team's workflow… she has shown a commitment to the growth of the team members by mentoring and supporting junior and senior writers.”
Jimmy Morales Jr., managed Eugene directly at Google