Injecting Customer Obsession into a foreign culture

Driving culture change can take many steps, and sometimes a detour or two

Carlos Arguelles
10 min readApr 28, 2024

What do you do when you care deeply about the Customer, but you’re in a culture where that isn’t necessarily fostered?

I spent 11 years working at Amazon (2009–2020), then 4 years at Google (2020–2024). I met and worked with individuals in both companies that cared deeply about customers, but the two companies had entirely different cultures around this.

The Tale of Two Companies

Let’s start with Amazon. One of the key aspects of the Amazon culture is that it is Customer Obsessed. This is truly, genuinely, embedded into the DNA of the company and always has. When you read Amazon’s Leadership Principles, it’s no coincidence that Customer Obsession is listed as the first principle:

Customer Obsession
“Leaders start with the customer and work backwards. They work vigorously to earn and keep customer trust. Although leaders pay attention to competitors, they obsess over customers.”

But articulating that “X is important” in a bunch of corporate websites is only going to get you so far. You need processes to channel energy towards X and continuously reinforce to each employee that “X is important.” When employees see that X is not just words, but also actions, they buy into it. Amazon has a number of processes to ensure Customer Obsession is ever present in everyday life.

First, the company probes for signals of Customer Obsession in potential hires during every single interview loop. You literally cannot get hired unless your interviewer has data that proves you think and care deeply about customers.

After hiring, a powerful internal process is the PRFAQ (press release and frequently asked questions). I’ve written about this extensively in two previous blogs of mine, so I’m not going to rehash that, I’ll just point you to them:

In 2020, I was ready for a new adventure and left Amazon and joined Google. My four years at Google were rewarding. I grew a lot and learned a lot. There’s a lot to like about Google and googlers.

However, it was clear to me from my first day at Google that Customer Obsession was not embedded into the DNA of the company. It’s not that googlers don’t care about their customers. They do, and I can attest that they care deeply. But the culture itself does not have processes like Amazon’s to channel that desire into an actual outcome.

Sure, there’s signs like this one (I snapped this selfie at one of the buildings in the Google Kirkland campus):

My somewhat suspicious look in this selfie says a lot. It just feels like a pretty sign added by a decorator, without any real substance or meat behind it. Again: actions speak louder than words.

[Side note: I detest the word “user” instead of “customer.” Those words have entirely different connotations. Jack Dorsey, co-founder and former CEO of Twitter and Block, wrote an excellent blog that dissects the differences.]

1. The Problem with Design Docs

In early 2022, a little problem started bothering me more and more. Time after time, even within just my little sphere of influence, I sat in meetings at Google when we were reviewing a proposal for a feature or a product. A smart engineer would show up with a well thought-out, high quality 30-page technical design that went deep into the weeds of how the feature would scale to millions, how extensible it would be, the APIs, the specific payload for the APIs, dependencies, resilience, etc. They clearly had given the problem space a lot of deep thought and had worked hard to produce a high quality technical document.

Yet I always found myself unsatisfied because my first question was always:

Why Are You Building This?

And despite the design docs being fantastic from a strictly technical perspective, they rarely ever answered questions like:

Who is the Customer?

What is their experience today?

What will their experience be tomorrow?

How will their lives be better?

How will you measure and prove this?

Is it worth it?

This bothered me deeply. How could anybody have spent this much time and effort without tackling the fundamental questions first?

At every single design doc review, I sounded like a broken record, asking that question over and over again, to the point that it became a running joke. I wanted others to be asking this question too.

2. The Problem with Customer Announcements

A separate problem started bugging me as well. Our product was internal infrastructure, so our customers were Google engineers inside the company. Googlers could sign up to be on our announcement email group, to receive updates on features added or product changes. This group was large (about ten thousand people) so we had a good opportunity to reach a wide customer base. Yet when we sent announcements, they were ineffective. They tended to follow the format “I did X,” where X was purely articulated in technical terms, and it wasn’t always entirely clear why X was actually important to the business. Putting on my customer hat: I’m your customer, I’ve signed up to receive updates on product changes, and I’m getting an email that you added X to the product I use. Why should I care that you did X? How is X going to make my life better? I had a suspicion that most people were ignoring these emails entirely. There was a missed opportunity here to reach and engage our customers.

Putting Two and Two Together

Eventually, I realized these two problems had the same root cause. We needed better mechanisms to channel Customer Obsession.

I brainstormed on this. There was a current state, and a desired state. How could I get from A to B?

The current state was engineers in my org begrudgingly wrote an ineffective customer announcement right before launch:

My desired state was for my engineers to eagerly write a high-quality customer announcement before they wrote a single line of code.

I knew exactly what to do here, from my time at Amazon. Excitedly, I put together a little do-it-yourself course on what it means to write a PRFAQ and Work Backwards from Customers. I sent it around to everybody I knew. I added a little telemetry to see who was taking the course. “If you build it, they will come!” I thought excitedly. But it had only modest success.

Maybe my initial approach lacked personal engagement. I believed in what I was proposing, so unfazed, I started showing up at team meetings in various parts of my organization, relentlessly doing the presentation over and over again to smaller groups and trying to engage people face-to-face. Engineers were receptive to my ideas, and expressed support for doing this.

Yet nobody actually did it.

I felt deflated.

I learned two lessons from my failure:

  • By default, culture tends to reject ideas that are foreign to it. Injecting new ideas requires a more nuanced approach than you saying: “Look, here’s an amazing idea!” expecting that suddenly everybody will adopt it.
  • Rejection comes in many forms. Sometimes it is very direct: somebody saying: “No, I don’t like your idea!” straight to your face. But some other times, rejection is indirect. Nobody tells you they aren’t going to do it, and it may even appear that they like your idea, but then they will simply ignore it or forget about it and nothing happens.

A more nuanced approach to Culture Change

Maybe my proposal was just too foreign. Could I start with a smaller proposal then grow it from there? Sometimes change is revolutionary, but culture change is often evolutionary.

Maybe I could tackle fixing the quality of the announcements first, being very tactical about getting engineers around me to care about articulating things in terms of value to customers. Then, once that was successful and people understood the benefit, start shifting it to earlier and earlier in the process. That was a strategic culture shift with tactical steps to get there.

I now understood that my initial push had felt like “here’s yet another bureaucratic process that leadership is trying to force us to follow.” Googlers hate and reject bureaucracy (with good reason).

When you want anybody to truly buy into anything, it helps to phrase it in terms of value to them.

I suspected that since our current customer announcements weren’t written very well, most people ignored them, and the features had poor adoption. My theory was that if we wrote better customer announcements, more people would read them and use whatever it is we had worked hard on to release, therefore having more impact.

I put telemetry in the customer announcements we sent, to track how many people actually read them. The numbers were terrible. We were blasting these email announcements to ten thousand people, and about a hundred (1%) actually even opened the email.

That gave me the piece of data I needed to phrase my proposal in terms of value to the engineers in my team sending the announcements.

“If you write better customer announcements,” I pitched, “more people will read them. You have nothing to lose, right now only 1% of your audience reads them. Give me an hour of your time, I’ll help you write it in a way that appeals to them. More people will read it, then more people will use the thing you worked so hard on! Your work will have higher impact and that translates directly to a better performance review and more money for you!”

The email group was moderated, so I took the drastic step of injecting myself as a moderator. Customer announcements couldn’t be sent without my approval. This was forceful and it annoyed some of my peers, but I needed to take a stand in what I believed. Most understood that I was genuinely trying to help them.

This was a high touch engagement for me. I painstakingly read dozens of drafts of customer announcements, and iterated with individual engineers in my team to improve them. It took more of my time than I wanted to, but I needed to vote with my feet: I wasn’t just preaching that my engineers should follow a bureaucratic process from my ivory tower, I was willing to roll up my sleeves and work right there with them. And slowly, more and more engineers understood the benefit of writing better announcements, clearly articulated in terms of the value to their customers.

As I was doing this, I met others that also cared deeply about Customer Obsession, so I enlisted them to be my partners and also be feedback providers on the drafts and approvers of the announcements. They became some of my closest friends at Google.

This helped me create a process that was scalable and sustainable, which is critical to Culture Change. I was so pleased to see that when I stepped away for a while, it kept going. I had built enough momentum, and found a critical mass of buy-in, so that the gears of culture change had begun.

I also needed to prove success with data, so I tracked how many customers read the emails. When we started the process, about 100 people were reading our announcements. Within six months, they were getting 1000 reads, so we had improved reach by 10x. With more reads came more customers excited about feature launches, and more customers using these features, therefore more impact for the hard work of engineers in my org.

I had achieved this:

Now it was the time to start shifting left to get here:

This turned out to be slightly easier than I had anticipated. People not only had gotten used to writing great announcements, they had understood the benefit of doing that.

And they were teaching others. That is another critical aspect of culture change. Others teach others, who teach others, who teach others.

To shift left, I leveraged a working group of Senior Staff and Principal Engineers in my org. We tackled creating a Design Review Template to have better consistency in artifacts created in our space. A template was the perfect place to inject questions around Why Are You Building This? Who is the Customer? What is their experience today? What will their experience be tomorrow? How will their lives be better? How will you measure and prove this? Is it worth it?

I left Google in April of 2024 (to come back to Amazon). In the end, I didn’t change the culture of the entire company. I probably didn’t even definitively change the culture of my entire org. That takes time and investment and persistence. But I did plant the seeds, and set the wheels in motion, so that people think deeper about the Customer, and teach others around them to do so as well. I do believe I left a mark, and I’m proud. Culture change always starts with just one person willing to challenge the status quo.



Carlos Arguelles

Hi! I'm a Senior Principal Engineer (L8) at Amazon. In the last 26 years, I've worked at Google and Microsoft as well.