[rroblak]

Google is, fundamentally, an engineering company. Despite their size and breadth, they still don't understand customer support. Their approach is to use software to solve problems, and they insist on doing so even when it's clear that software isn't up to the task.

Unfortunately, customer support is a hard problem. Despite all of the advances in NLP, I still abhor automated customer support systems when I have a complex issue. Just let me talk to a human.

Google long ago ran the numbers on providing human customer support and realized it's not the sort of ultra-scalable business function that they like to invest in. Rather, they'd like to believe that they can build software systems that don't require human customer support. As an end user, this feels like too much hubris and not enough empathy. It may work from the perspective of a product manager looking at percentages on a dashboard, but it sucks as someone in the real world trying to get something done with one of their products that's not functioning as it should.

I use the full suite of Google Products, including Project/Google Fi. This article describes one of my nightmares— getting locked out of my Google account. I'm fortunate that I have good friends that work at Google that could help out in such a worst-case scenario. This blogger is fortunate, too. Undoubtedly, some Googler will read this post and help them out.

But the average person isn't so lucky. If you're Jane or Joe Schmoe in Middle America, you're going to be screwed when your Google account goes haywire. I've had friends whose Google accounts have gotten into weird states that prevented them from using Google services for no obvious reason. I suspect this is due to an unfortunate consequence of Conway's Law [1] at work in Google's identity implementation.

[1] https://en.wikipedia.org/wiki/Conway%27s_law

[wpietri]

I don't think engineering is the problem here. As an example, Toyota, a very engineering-driven company, is also famous for customer focus.

I think the problem is that Google is mostly about selling users' eyeballs to their real customers, advertisers. That's not a business of making individual users happy; it's essentially statistical in nature. With a search engine, if something works for 80 or 90% of people, that's great. If it's bad for the rest, well, tough luck for them. It's very hard to go from that to seeing each individual as valuable and important.

[ehsankia]

it costs $xxK to buy a car. It costs 0$ to use Gmail, Drive, Youtube, etc. These two are not gonna have the same level of customer support. Fi is a paid service, and I do expect it to have a better support (which in my own experience, they do), but to compare Google as a whole to Toyota doesn't seem fair.

[walterbell]

> That's not a business of making individual users happy; it's essentially statistical in nature.

Good point. Could telecom service (Project Fi) for individuals be moved to a different division of Alphabet?

[hogu]

I have friends at Google that are L5. When my AdWords account was suspended (long story but if you Google AdWords banned my hackernoom article explains) one of them tried to file a ticket on my behalf. Went nowhere. As far as I can tell internal actions like that go the same route as tickets that I file as a normal person. So don't let having googler friends give you a false sense of security.

[kxrm]

You nailed the issue for me, the problem is that these software approaches to customer service always assume that the service is 100% not the problem and that the customer is the one causing the problem.

[r00fus]

> Despite their size and breadth, they still don't understand customer support.

Cue Larry Page's view on customer support circa 2000, and it still makes sense. Leaders fundamentally don't change views like that, and it impacts the organization - look at Zuckerberg's formative views on privacy.

[hn_throwaway_99]

I didn't previously know what Larry Page's view on customer support circa 2000 was:

But while it's easy to scoff at Page's quirks—his odd obsessions, his unrealistic expectations, his impatience for a future dangling out of immediate reach—sometimes his seemingly crazy ideas wind up creating breakthrough innovations, and skeptical Googlers wind up admitting Page was right, after all. That was the reaction in 2003 when Denise Griffin, the person in charge of Google's small customer-support team, asked Page for a larger staff. Instead, he told her that the whole idea of customer support was ridiculous. Rather than assuming the unscalable task of answering users one by one, Page said, Google should enable users to answer one another's questions. The idea ran so counter to accepted practice that Griffin felt like she was about to lose her mind. But Google implemented Page's suggestion, creating a system called Google Forums, which let users share knowledge and answer one another's customer-support questions. It worked, and thereafter Griffin cited it as evidence of Page's instinctive brilliance.

https://www.wired.com/2011/03/mf-larrypage/

[PavlovsCat]

Good engineering, just like good customer service, is super easy if you put integrity first and don't compromise it. Don't grow beyond your capabilities to handle your stuff with integrity, done.