Is Google Cloud experiencing challenges going upmarket?
Published (updated: ) in Cloud. Tags: Cloud.
Originally written for InfoWorld.
Google is known as an engineering- and automation-driven company. Despite its core product, search, being as mass-market as you can get, Google has historically done a good job at working with developers: quality APIs, good documentation, and true engineering skill.
This focus on engineering has hurt its reputation when it comes to support, customer service, and ability to work with enterprises. However, in recent years that has changed significantly, and it’s only in online techie forums (such as Hacker News), from people who have never interacted with paid Google support, that you see complaints about how it’s impossible to speak to anyone. As a customer of Google Cloud and G Suite — both paid products — I can personally attest to the responsiveness and usually high level of quality when receiving technical support.
Targeting the grassroots developers
That said, Google Cloud started very much targeted at grassroots developers. Products such as Google Compute Engine were very well designed, but Google had a very arrogant view that its way was the best way and if you didn’t like it, you could go somewhere else.
Indeed, this was, and still is, part of the selling point: operational simplicity vs flexibility. Google has significant experience of running large scale systems so they will make many of the choices for you, so you can focus elsewhere. Contrast this to AWS’s approach which is essentially to provide every possible option and configuration you can think of, and let you decide how to set things up.
AWS wants to replicate your data center as an API. Google wants to transition you to the optimal cloud-based architecture, even if that means changes to how you’ve done it historically.
The go-to-market approach has changed
With Diane Greene joining, the commercial go-to-market approach has changed. The product quality is still just as high and you can see Google continues to place a great level of importance on engineering the best cloud products. However, Google is listening to feedback, it has more consistent messaging and it is attempting to show that it has real customers across a range of demographics. It’s no longer just Snap as the only high-profile customer!
The Google Next Conference last week was a good example of this. Spotify, Evernote and the range of enterprise customers like HSBC that were just announced all make up a growing customer list. I was able to attend the pre-conference Community Summit and the messaging was very consistent, following through to the full conference that followed over the next 3 days. Google is positioning itself as the place for machine learning, serverless, and container based workflows. It is doing this through building the best open source frameworks (Kubernetes, TensorFlow) then supporting them with what it hopes is the best place to run them as managed service (Google Container Engine, Cloud ML). Such an approach relies on developers adopting those frameworks, then bringing them into work and having their teams deploy on Google Cloud because it is the best place to run them.
Unfortunately, I get the feeling that in trying to highlight its enterprise credentials, Google is in danger of alienating the community it needs to get into those accounts. This became apparent from the first day keynote at Google Next with the comments online and in the hallways at the conference itself. From a customer and developer perspective, there was almost nothing interesting announced (Video Intelligence API being the exception). The speakers were unremarkable and it was really designed to get the attention of people who don’t already use Google Cloud, and of course enterprises. Look back a year to the previous keynote and it doesn’t look like anything has changed much from what one analyst called, Google’s Scalability Day.
Making it up on the second day
Fortunately, the second day was a true technologist keynote. This is where Google was able to show the features to back its claims that it is able to service enterprise. Security functionality such as its Key Management Service and Identity Aware Proxy as well as pushing out releases for more bleeding edge platform services like Cloud Functions.
There’s no question that Google has the best technology, particularly when it is productizing things it uses in-house (Container Engine, Cloud Bigtable, Cloud Spanner, Cloud Load Balancing). But it has to do more than just make internal products available publicly. Google is well along that path and it’s no longer a question of whether Google is “enterprise ready” (it is). The challenge now is how it perfects the go-to-market approach. Pushing into those larger accounts without damaging its reputation with the customers who helped to get it to where it is now will be crucial for continuing to sell into the bigger accounts.
Google has the technology, and it has a view on where it stands out. It is pushing into the enterprise and closing good deals. It just has to be careful not to forget where it came from.