Two weeks ago, the Medium team paused our regular work to focus on shipping small improvements for writers and readers. Inspired by a conversation with an editor, we called this Hot Broccoli Week.

The idea began earlier this year when the editor shared feedback that some of our recent changes had felt unappetizing, like cold broccoli. This feedback galvanized our team into thinking differently: What if we we shipped small, user-facing improvements that had outsized, positive impact on how they used our product? We decided to turn the metaphor on its head — and Hot Broccoli was born.

While we normally focus on shipping changes as part of major product initiatives with big impact for readers and writers, Hot Broccoli Week was all about improving the quality of the Medium platform in small but meaningful ways. To make it a success, we invited team members from across the company to submit their candidates to a shared Hot Brocc-log, a backlog of improvements and fixes we could make across platforms. These suggestions came directly from writers and users, based on feedback and research. The idea of “hot broccoli” also provided a fun, cruciferous theme for the week, which featured custom broccoli emojis, a broccoli profile pic competition, broccoli-themed prizes, and more (to say nothing of countless broccoli puns throughout the week).

The work paid off. Our small team completed 30+ user-facing changes in the product, as well as many more internal optimizations to our platform, tooling, and processes. We focused on what we’d heard the most feedback about, as well as on what we could reasonably complete in a week. We cleaned up old bugs, improved our site search, and made Medium look and feel nicer to use across the web and mobile apps. Here is a detailed look at what we accomplished:

• We undertook a large-scale effort to de-index spammy stories and accounts from Medium’s on-site search, and we ultimately removed 100 million spam records from search results. As a result, search results are more accurate and you’ll find what you’re looking for more quickly.

• We made improvements to make search results more consistent. This means you’ll now see the same results for the same search query across surfaces where results appear, such as in the search dropdown and on results pages.
• We revisited our ranking and sorting factors, tweaking them to better surface relevant content for search terms. For example, we now surface publications that match the query in the dropdown.

• We added new hot keys to open and close the search bar. You can now press Cmd+K to open the search menu, as well as the esc key to close it. This is a common pattern in many desktop apps and makes it extra easy to start searching.
• We fixed a bug with the search dropdown menu that prevented it from correctly closing after pressing “Enter” and caused it obscure the results. The dropdown now disappears and the search results are fully visible.
• We changed up old, outdated brand art across the site. Goodbye, old ASCII art of Medium past! You’ll now see our new and improved brand imagery whenever you do things like share medium.com.
• We rolled out new default preview images for stories published without images. These will appear in story previews throughout the platform.

• We improved features in our internal admin tools, like making it possible to search our admin tool via user’s email address. This simple change will allow Medium’s Support team to create shortcuts to access user accounts directly, which means we’ll be able to respond to your Support tickets even faster.
• We designed and built an onboarding survey for new users, which will allow users to optionally share more information about themselves with Medium as they join our community. This means we can more accurately identify new features that you want.
• We made it easier to sign in to Medium accounts on web. When logging in, we now send both a login code and a magic link by default. Instead of clicking the magic link, you can simply enter the six-digit code sent to your email address, and you’re in!

• We fixed a bug in the iOS app that caused tooltips to appear in triplicate. Now, tooltips should appear once and correctly close when clicking “Okay, got it.”
• We fixed a bug in which our iOS app UI appeared overlaid on the iOS UI. This ensures that the iOS UI features like the clock, Wi-Fi connection, and device battery life always remain visible while using the app.
• We improved the behavior of the apps after editing a published story. Now, after updating a published story, the story page will refresh to reflect the newly published changes.
• We fixed buggy behavior in the implementation of “sections” in the Android editor. It’s now much easier for writers to add and remove sections between other media and content types.
• We fixed issues with deep linking in both the iOS and Android apps, making it easier to access screens like reading history and stats within the app.
• We made it easier to find recently edited drafts in the Drafts screen of the iOS app.
• We updated the UI for “empty” states across the iOS app, making it clearer and easier to understand when a list or tab in your user profile or library has no content in it. We also updated the information on those screens to clarify what actions you can take to fill them with stories.

• We fixed a bug in the Android app in which users would be logged out after switching between light and dark mode. Now, you’ll stay signed in after switching themes.
• We made it easier to clap for stories in the Android app by increasing the size of the tap targets for clapping.
• We improved the the publishing flow in the iOS Editor to display an error message if you try to publish a story without a title or body copy.
• We improved the UI of the “Follow” button in the “Who to Follow” section of the iOS app, making it easier to tap and follow new Medium writers.
• We fixed some unclear or outdated copy across the site, making it easier to understand Medium emails and in-product messaging like tooltips.

These are ultimately just some of the tasty changes we released in the span of a week! There were even more changes we made internally to our tools and processes to help us better serve Medium users. While we are proud of what we accomplished during Hot Broccoli Week, there’s always more on our list to be done.

To everyone who has shared feedback with us about ways we can make Medium better: Thank you! We’re listening, and we appreciate how much our community of readers and writers care about Medium. We’ll continue to make improvements to Medium in our pursuit of building the best place for reading and writing on the internet.

Florets and fixes: Updates from Medium’s internal “Hot Broccoli” Week was originally published in The Medium Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.

Almost 10 years ago exactly, I quit my job to enroll in a coding bootcamp and become a web developer. At the time, it felt the scariest, riskiest, craziest thing I’d ever done — which is saying a lot for someone who, until then, had aspired to be a professional news writer.

To document my experience learning to code, I decided to keep a blog on Medium. The day before my class began, I wrote my first post.

The Way We Tell Stories

Ever since I was a kid, I had wanted to be a writer, then a journalist, and then a news reporter specifically. But that was 2006–2012, and neither my high school nor university were particularly ahead of the technological curve. The schools’ journalism teachers, curricula, and student newspapers all prioritized the print side of news, instilling in me love for and devotion to a dying medium.

So, imagine the dissonance; I was split in two. On the one hand, I was wracked with guilt over “selling out,” as if I were betraying newspapers and personally hastening their downfall by embracing the Internet. I worried that I would have to take a job building websites for some random company in an industry that I didn’t care about — or worse, that I wouldn’t be able to get a job at all.

But beneath my spiraling thoughts, I also knew something that felt true. At the time, I wrote:

“Is print the best medium for stories? I’ve thought a lot about the future of media, and my answer is no. There are many ways to tell stories more effectively, more engagingly, and more beautifully than a static newspaper article can. Online, there are endless possibilities for storytelling. I think we just have to build the right medium.”

It was mostly hope, a belief that the internet could be a better, more creative place for storytelling than what had existed before. I didn’t know exactly how I would fit into that vision — just that I wanted to be part of building it.

Looking back, the silliest part about how bad I felt then is that I landed a job as a software engineer at The Washington Post a few weeks later. For years, I couldn’t believe it; building tools for journalists and better news products for readers felt like the pinnacle of my career.

A year ago, though, I joined Medium, a team and company whose mission aligns in many ways with the hopes I put into words on this very platform back then. For the first time in my career, I have been building a product I personally love to use, and that feels good and meaningful. It’s hard to describe what it was like to re-read my own words in that blog post when I stumbled across it late last year. It felt like I had come full circle, back to the platform where my software engineering journey had started; it felt like I was home.

Like anything meaningful, though, building Medium has come with its share of startup ups and downs. Most days, my work feels energizing, and I can see a direct line between my daily efforts and that bigger vision I first imagined. But there have been stretches where that connection felt harder to see. A few months ago, as we navigated multiple major organizational transitions that created lots of uncertainty, I nearly lost sight of the vision.

What I’ve learned this year is that staying connected to a vision (any vision!) requires more than just enthusiasm. It requires commitment. For me, that’s not commitment to a company or a title, but to the deeper love and purpose that called me to my work in the first place.

Choosing to stay committed — to myself, to Medium, to my vision of building a medium for stories online — has made this work richer. It has also helped me realize that finding meaning in my work isn’t a destination to arrive at. It’s something I practice returning to, again and again.

Over the past year, when my enthusiasm for my job wavered, I found myself returning to that original post, to remind myself of my own belief that I can help build a better platform for stories to live and breathe online. I am here because I believe thoughtful, creative, human storytelling matters. I am here because I believe the internet can be a place where we find community in each other’s words. I am here because building something meaningful is hard and worth the effort.

I wanted to wake up this morning on my one-year anniversary at Medium feeling bright and energized and connected to meaningful work. Today is just a Tuesday, so I practice returning to what I still believe:

Online, there are endless possibilities for storytelling. We just have to build the right Medium.

Unlike most people during April 2020, I was still enjoying social distancing in my apartment outside Washington, D.C. Then, I found out that your cancer had come back.

I realized that I might never see you again. It was the beginning of the pandemic, and the thought of making it to Lake Havasu City by plane from D.C. was laughable. Still, it broke my heart to think I wouldn’t get to follow through on a promise I’d made many years before, that we’d share a beer together one day. For a few months there, I thought we never would.

So that’s how, while all of my friends stayed shut up at home, I ended up on the road, having the adventure of a lifetime. It’s because I promised you that we’d share a beer.

Between April and June, I spent hours researching ways to get across the country. I got quotes for custom vans, researched how to get wifi on the road, traded in our sensible sedan for an impractical-but-beautiful diesel truck, and finally found the little Basecamp trailer — the lil squirt! , as you called it— that took us all the way from D.C. to you.

It was mid-August when we finally arrived, but I’ve never felt more proud than when we parked our trailer under the RV port. The previous occupant of the RV port at your house was your Pace Arrow, the motorhome you drove from Arizona to Washington to visit us every summer—the motorhome where my best childhood memories were made.

And there I was as an adult, parking my trailer where that motorhome once stood.

Everything came full circle in a way that I had never anticipated. Instead of you coming to visit us for the summer, I came to visit you.

Instead of eating your delicious cooking, I cooked for you. You hadn’t been eating well lately but together we ate pork chops and potatoes and donuts and baked chicken wings and copious amounts of vanilla ice cream.

Instead of you fixing things around our house, I fixed things around yours; I saved the old bracket that I replaced in the cabinet shelf; I sorted out your new cable system so that we could watch your favorite channels, and we saw some new war movies together, too (they were not as good as the old ones, I will admit).

It was a great trip, but we still didn’t get that beer. You were on a fentanyl patch, which scared me, because I know what fentanyl provides: powerful pain relief for those who truly need it. I cried myself to sleep the night you fell, hit your head and tore the thin, delicate skin on your arm, and said it was nothing.

So, we made a second deal, that after our trip to visit the rest of the family in Seattle, we would come back for another shot at it later in the year, on our way back to D.C.

Like a Steffan, you wanted details. “When?” you asked.

Like a Steffan, I gave you a date. “October,” I said — and like a Steffan, I kept my word. On October 31, we rolled back in Havasu, towing a much larger fifth-wheel trailer (this one didn’t fit under the RV port!), and thirsty for a beer.

You had changed some things as well. In the months we’d been gone, you stayed motivated for our return. You started eating more. You had gotten off the fentanyl patch — and you were ready to have that beer, too.

There’s a line in one of my favorite songs, “No other magic could ever compare,” and that’s true of my memory of that day. We drove the half-mile down the road to Mudshark, we laughed with your favorite waitress, we ordered your favorite beer, and we raised our glasses to a beer long awaited.

Pat caught it on camera, the moment we both took that sip — and no other magic could ever compare.

I wanted a beer with you and I got it. Somehow, in a life filled with precious memories full of people who love me, this memory stands apart. I think it’s because, in between our stops in Havasu, I learned so much about marriage, long-lasting love, what it means to be family, and how I can lead in my relationships going forward.

I got the beer, but in the process I got so much more. Thank you for the beer — and for everything else. I love you.

A few months ago, we launched Featured stories for publications as a way for publication editors to directly influence story recommendations for their followers. Since then, our team has released multiple improvements, making Featured stories visible in stats and post page labels.

Now, we’re excited to give Featured stories even wider distribution and more prominent placement on each user’s logged-in Homepage — great news for both writers and readers. Over the next few weeks, we’ll be rolling out new Featured story push notifications and new “Featured” story Homepage feeds for publication followers.

For writers, this means even more people seeing your Featured stories. For readers, it will be easier than ever to see Featured stories from all of the publications you follow.

In addition to being recommended more highly in Digest emails, the For You feed, and publication story pages, Featured stories now will be pushed directly to a publication’s followers via in-app notifications. Followers with the Medium app installed on iOS or Android also can receive Featured story push notifications whenever a publication they follow features a story.

Featured stories also will appear in the new Featured tab on the homepage for logged-in users. Unlike many of Medium’s other feeds, which include content recommendations from the Medium algorithm and our in-house curation team, the Featured feed highlights all Featured stories from publications that a user follows in the order they were featured. This means readers will never miss out on seeing a publication’s Featured story.

The Featured feed is another way that Medium is investing in both publications and human curation. Publication editors play an important role in understanding what kind of content is uniquely valuable to their readers. By giving publication editors more control over the content that followers of the publication see on Medium, we are excited to help even more stories find the right audience. To learn more about Featured stories, check out the help center.

P.S. We’ve had the Featured feed turned on for Medium staff for about a month now. What started as a staff-only prototype made its way to production pretty quickly after we had the chance to experience it. Trust us when we say that it has become a go-to source of reading material for many of us!

Making Featured stories even more visible was originally published in The Medium Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.

About a year into my new career as a software engineer, my boss invited me to help launch a news website on site with a client in Argentina. It was our first major client to launch, and my first launch ever. I was terrified.

I felt both incredibly proud of the frontend I had built and also deeply unqualified to be in the room when traffic was cutover. Surely there was someone more qualified they should ask instead? Apparently not! So on that trip I made it my goal: I wouldn’t be the wrong person to have in the launch war room.

It was my very first launch from hell, the kind most software engineers only suffer in their nightmares. As soon as we began the cutover, everything from the platform to the admin applications to the user-facing website went to shit. Nothing worked as expected. 📉

The office we were given for the week was a quirky conference room—the kind that’s perfect for a quick team huddle, but bad for, say, an extended in-office stay. One wall consisted of internal windows facing a main walkway, broadcasting our every move (namely, our growing distress) to passersby. All remaining wall space was painted to match the color of the company’s brand: bright orange.

It was my very first launch from hell, the kind most software engineers only suffer in their nightmares.

The world’s orangest conference room felt cheery when we arrived in the daylight, but became menacingly bright under the fluorescent office lights in the wee morning working hours of a round-the-clock launch day. From 8 a.m. Tuesday until 5 a.m. the following morning, we fought to keep things stable and to fix enough frontend issues to build the client’s confidence in rolling forward. After 21 straight hours in the Orange Room, we went back to our hotel, ate breakfast in bleary-eyed silence, slept a few hours, and returned early that afternoon for another grueling evening.

Against this (literal, sweaty, orange) backdrop, I worked to prioritize and fix frontend-related issues that the newsroom was reporting. Traffic was going down; stories weren’t updating quickly or at all; readers were complaining on social media; the tools were timing out under increased load.

P0. P0. P0. At a dizzying pace, people reported bugs in features I had personally built, activating my worst fears of failure; maybe I wasn’t actually good enough to do the job, I thought. I kept getting pulled in multiple directions—an editor needed help; a data scientist wanted to review the analytics; the ad ops director needed reassurance. I worked as fast as I could to commit and deploy new versions of my code, until the fragile, flaky deployment tool stopped working altogether.

An anxiety attack forced me out onto the sidewalk outside for relief on the second day of non-stop fire fighting. Outside was scorching, bright, and humid as hell, but (and this was critical) not orange. I called my husband and said I wanted out, wanted to come home early. It was too much pressure, and I wasn’t good enough to keep going.

I was failing, and it felt terrible.

That’s fine, he said. But you’re the only frontend developer in the company who knows this codebase. Who else will they call to help firefight?

Literally anyone!, screamed my brain. And yet I knew: There was no one else, only me.

I returned to the Orange Room that day—and the next two days—armed with a double espresso and newfound resolve. I faced off against my biggest fears and the ever-growing list of bugs. I prioritized them and got to work. They were bugs I had introduced, yes, and ones I knew I was capable of fixing, because something had clicked.

My team didn’t need the world’s best software engineer in the room with them on launch day. They needed me, doing my best.

In the Orange Room I survived a nightmare launch and didn’t get fired or even “found out.” In fact, shortly afterward, I got promoted. I received the opportunity to lead more projects for even bigger clients around the world. I started to (ahem, this is a technical term) crush it.

I became the right person to have in every launch war room, because as it turned out, nightmare situations were kind of my specialty. I even kind of enjoyed them?? I suppose it was a specialty born of necessity, because we never did figure out how to launch a new client’s site without taking down our entire platform. We’d (usually) get it right eventually, but not without crashing everything a few times. We failed a lot—and I was actually pretty good at handling things when we did.

We failed a lot — and I was actually pretty good at handling things when we did.

In those moments, I didn’t magically transform myself into an ops engineer, capable of addressing the main issues facing our systems. I wish. But that’s not even what my team needed me to be.

My team needed someone technical on the ground with the client, who could communicate to them in non-technical terms what was happening in the system. Someone who could train editors on their new systems and troubleshoot usability issues. Someone who—in the midst of an absolute shit-stack of cutover chaos at 5 a.m.—would make an editor feel heard and make their individual concern seem like the most important one. Someone who could investigate why Google Analytics wasn’t working. Someone who could explain why it was taking so long for a story to update in our platform in plain terms.

At the time I didn’t understand why I actually enjoyed these situations—objectively terrible launches—so much. They sucked so badly, yet I still reminisce fondly about how much fun I had. Remember the Orange Room?!?!

In retrospect, I can see more clearly. Under pressure, I didn’t have time to worry or ruminate or ponder my existential fraudulence. In those moments, I got to be the best version of myself: not only a skilled software engineer, but an empathetic listener, a clear communicator, a kind teacher, and a resilient leader.

I didn’t have to be perfect; I just had to be me.

I loved it for a while. I thought I loved it. I was succeeding at my dream job, and wasn’t that enough?

The “best” version of myself was exhausted. My schedule, when I was in the office, consisted of wall-to-wall meetings, with little time to accomplish deep technical work. But I wasn’t actually in the office that often. My constant schedule of international travel—a novelty I enjoyed at first—drained me; I regularly traveled to South America and Europe in back-to-back weeks. Most of the time, I flew in premium economy and went straight from the airport to the office. My only time at home was a brief weekend pit stop to recover, physically and mentally.

I was succeeding at my dream job, and wasn’t that enough?

I was not actually doing well. I felt like I couldn’t stop working, and I left my depression completely unattended. I drank heavily to cope. Yet, I felt “lucky” to have found an engineering role that both gave me a sense of purpose and suited my skills so well. I was building websites for some of the biggest publishers in the world.

After he promoted me again, my boss suggested I watch out for “golden handcuffs.” I was being rewarded for my hard work, which was justified, he suggested, but I had been promoted to a level internally that he thought I couldn’t sustain or justify externally. I understood what he meant: If I ever wanted to look for a new job, I’d have to look several levels (and pay grades) down.

Still, I faced down challenge after challenge with the resilience of a champion boxer, throwing myself into scrappy, tough situations again and again. Of course it hurt to get punched in the face, but most of the time I’d emerge victorious.

Once, I was tasked with training a client’s team of 15+ software developers solo in Argentina. When I arrived, I learned that the team I needed to train only had access to Windows machines—and our local development environment was known to only run on Linux. My platform team back in the U.S. gave me the equivalent of a “yikes 🤷‍”, so I improvised 3 days of hands-on developer training that gave each person a chance to build a feature using the only Linux machine in the building: mine.

Another time, I went on a sales trip to Hong Kong to meet with the technical team from a prominent pro-democracy publisher. The engineers peppered me with questions about our CDN and infrastructure, insisting our platform wasn’t prepared to handle the kind of security concerns and attacks they attracted. I admitted I didn’t know the specific answers, but I talked them through how I had seen our team handle similar situations for other major publishers. We brainstormed a list of questions I could take back to my team, which I did — along with their approval to move forward with a purchase order for our services.

I finally cracked at the end of another hellish launch, one not dissimilar from the one in the Orange Room. Sleep-deprived and a little manic from 3 straight days of bug fixing, I continued committing code until the moment my flight took off to go home. Instead of breathing a sigh of relief when the flight attendant announced a problem with the on-board WiFi, I burst into irrational, exhausted tears. I knew I couldn’t keep going.

For a few months, I felt the golden handcuffs tightening around my wrists, binding me and my identity to my job. But a strange thing happened the minute I finally acknowledged that I needed to leave: I realized the handcuffs weren’t actually locked. In fact, there were no handcuffs at all. I was the one with a death grip on my job, hanging on for dear life, afraid of who I’d become if I let go of the role that had come to define me.

I realized the handcuffs weren’t actually locked. There were no handcuffs at all.

I hadn’t stopped working since the Orange Room had given me the chance to prove myself. If I wasn’t working, who even was I? What would happen if I let go? Would I fail? Would I land on my feet with a new job? What would that be like?

The answer dawned on me: It would be exactly like the Orange Room, exactly like every other challenging situation I had faced, navigated, and survived. Those situations didn’t scare me anymore, and there was no reason to be afraid of this one either. I knew I could handle anything, because I already had handled everything that happened to me before. The common denominator for my success wasn’t luck or even the job itself; it was me, showing up for myself every time.

My grip loosened.

In just four years, I had grown from a terrified junior developer into a competent technical architect who ran toward even the most difficult problems. The title was new, but I had always been that person. In the end, the title was the only thing I was holding onto, because it was the only thing I couldn’t take with me if I left. The skills I built, the domain expertise I learned, and the experience I had gained—those all were mine, and they’d stay with me if I left the job.

I wasn’t scared to let go, so I did. I’m happy to say I’ve had confidence in myself ever since.