sogrady (Stephen O’Grady): So all the monks are back from San Francisco: what were everyone’s general impressions of Google Next?
fintanr (Fintan Ryan): Overall positive steps, and the developer message around open source is strong, but still a lot of work to do.
sogrady: Which part of the open source messaging resonated with you?
fintanr: For me it was definitely the messaging around being community players, working heavily with existing communities, building new ones.
sogrady: The Community Day they held certainly helped in that respect.
rachel (Rachel Stephens): Yes, they did a great job with the Community Day.
monkchips (James Governor): I came away fairly positive overall. There is a lot of strength and depth at Google, and if it can just start story-telling more effectively, and do a lot of boring block and tackling, it’s going to be a truly formidable competitor in the cloud space
fintanr: +1 on story telling @monkchips
monkchips: The community day was great fun:
Unconference about OSS and stuff. #gcpnext17 rocking it old school
— Artie Coolfifty (@monkchips) March 7, 2017
fintanr: I am always a fan of the unconference format – it can be really productive with the right folks.
sogrady: Another +1 for story telling. Too often the show felt like a bunch of unrelated product announcements thrown out there with not much in the way of a larger narrative.
rachel: That makes it a full house for the story telling +1s.
sogrady: For my part, I thought the show got better as it went along. Day one, particularly with some of the teleprompter issues and so on, felt like something of a lost opportunity to me. It just didn’t feel smooth or well rehearsed.
By day three, it felt like it was hitting its stride.
monkchips: Ah yes @sogrady charitable as ever…
fintanr: The day three keynotes flowed well.
monkchips: For me day two was when Google hit its stride, with Brian Stevens on stage, they were dangerously close to crushing it…but they lost some momentum as the keynote wore on.
fintanr Actually on the Brian Stevens note, though he’s been there a while, it is interesting to see just how many new faces Diane Greene has brought in.
sogrady: Indeed, a lot of new people. Wasn’t too surprised to hear that the Cloud team is hiring faster than any other Alphabet BU.
So quick thoughts: what did Google do right and what did it do wrong with Next?
fintanr: Right – Dr Fei-Fei Li, great story telling.
sogrady: Big +1 on that. She is excellent.
fintanr: They brought out the machine learning aspect big time, and doubled down on data aspects which will be absolutely key.
On the enterprise front, the SAP partnership is good.
Negatives – overall story arc, logistics, timings. Lots of grumbling from people about registration on day 1, folks we all know were outside in the queue for up to an hour of the day one keynote. The keynotes need to run on time (but in fairness when do we ever see a keynote run on time).
Logistics & timings are easy to fix, story arc will take more work.
rachel: On the enterprise note, you could tell that they worked to curate an enterprise-friendly message.
For all the news about their recent deal with Snap, that wasn’t the focus. They brought out Disney, Colgate, Verizon, Home Depot, HSBC, Schlumberger, etc as their customer case studies.
monkchips: The AI and Machine learning section to me felt like a bit of a missed opportunity. Google should have led with Dr Fei-Fei Li out of the gate for day one. Everyone at the keynote was wearing women’s day ribbons, but in ML they didn’t need to wear ribbons. It was kick ass women kicking ass. Dr Li, the hiring of Jia Li, then Sara Robinson, doing the demo… I was in my happy place.
On day zero at the analyst event Dr Li really swept me away with her story telling. It was like right ok, great journey, that’s really interesting. Where do I sign up?
sogrady: Which was, it’s worth noting, the first time she’d ever presented to analysts.
monkchips: Missed opportunity because Google is so strong in the space, TensorFlow is going nuts, they have real clients.
So there were two interrelated stories – one the next wave of cloud is deep learning and ML and two, it’s being led by women at Google.
sogrady: What about you, @rachel? What could be improved?
rachel: They definitely need to work on their narrative. It’s a hard journey to move the culture from tech-dominated conversations, and you can tell they’re still early in that process.
And to be fair, there are definitely still contingents of people who expect that as the conversation, see:
Google NEXT '18 keynotes should demo the most DOPEASS TECH ON PLANET. That's what they have & we want to see #GoogleNext17
— Chris Gaun (@Chris_Gaun) March 8, 2017
monkchips: Haha. Right.
rachel: So trying to find the balance between showcasing their tech and communicating their message coherently is where they have the most obvious chance for improvement.
sogrady: That’s fair.
monkchips: @rachel so right. And I believe that’s one area where they could have used references more effectively. HSBC for example is a fantastic reference for Google. But I am not sure the setting matched the stone.
sogrady: For my part, I thought Google did a good job of articulating its data and ML stories, but still struggles with telling anything but a very technology-centric story – the enterprise case studies notwithstanding.
As aside on the data story, I was curious whether you all felt – as some did – that Google should have held the Spanner news for Next?
fintanr: Honestly I think it would have got lost at Next.
monkchips: Tend to agree with @fintan. Our industry spends too much holding things for the big reveal. Then it’s like oh awesome you have 100 things to announce in 3 days.
sogrady: Surprisingly, we’re all in agreement. Without getting into a conversation outside the scope of this chat – what are conferences actually for? – I thought that viewpoint was outdated.
Announcements are great and you need some, but dropping too much at once is counterproductive, IMO.
Ok, next question. What was the most important announcement or takeaway from next, in your opinion?
monkchips: Google Home is really good at Spotify! 🙂
fintanr: For me it’s mainly takeaways over announcements – in the ML area, TensorFlow still has crazy momentum, Google understands the importance of useful datasets and combining them together for developers.
On saying that, the Kaggle acquisition is a big deal IMHO.
It has the chance to pull the next wave of data scientists directly onto Google’s platform from day 1.
sogrady: So your most important announcement or takeaway is like three things then? 🙂
fintanr: LOL, yes.
rachel: In terms of announcements, I think Kubo is pretty cool.
sogrady: Yeah, that was a real surprise when we got the preview on that. Wait, Pivotal is supporting…Kubernetes?
rachel: It feels like it may have more immediate impact on the Pivotal ecosystem, but it also supports the Google open cloud concept. Mobile workloads are a key aspect of how Google will be able to compete against other cloud providers, and Kubernetes is a path to that. Seeing Pivotal, which has already embraced an alternate platform in PCF, also supporting Kubernetes is definitely an interesting progression in the ecosystem. It will be interesting to watch Kubo’s adoption both from Pivotal and Google’s perspective.
sogrady: Concur. @monkchips: biggest takeaway?
monkchips: My big takeaway is totally selfish. Google needs to change, knows it needs to change, and it is not really a coincidence they signed up as RedMonk clients pretty much at the same time as running the event. That point at which Web Scale companies feel they want to engage on an advisory basis is interesting, and it was great to have a couple of days to get a feel for where we can help. Like story-telling? Hello – we can definitely help. Doubling down on developer experience. Yes please. Google isn’t just more open to us though, that’s my point, it’s really beginning to smooth off the arrogance, which will definitely help the firm to compete with Amazon and Microsoft.
We do our best work with companies that know they need to change, so we’re kind of a bellwether.
But working with open source communities, coming across as more humble in the industrialisation of “site reliability engineering”…is huge.
fintanr: +1 @monkchips the whole CRE thing is going to be fascinating to watch over the next twelve months, and is a huge opportunity.
sogrady: Got it. For my money, it’s the level of commitment on display. Google’s biggest issue, amongst the people I talk to and some in the media is credibility. Next should be helpful in that regard: 10K people at the event, solid new tech, messages that resonate like the Customer Reliability Engineering (CRE), branching out from startups to enterprise markets, ~$30B in CAPEX investments over the last three years, et cetera.
rachel: $29.4B…. 🙂
fintanr: What’s $600M among friends @rachel 😉
sogrady: It’ll take a while for Google to move the needle because of skepticism based in part on Google’s willingness to deprecate businesses outside of search, but Next was a step in the right direction IMO.
monkchips: @fintanr and @sogrady tell us more about CRE. It could be a thing.
sogrady: The idea, according to Google, is basically to extend the company’s Site Reliability Engineering (SRE) practices into the customer’s instances, such that – as one example – when a customer’s app goes down Google folks get paged along with the customer ops staff.
fintanr: The part that I found most interesting is how these teams are being created, they are taking the best people from Site Reliability Engineering and moving them into CRE – and from all of our conversations this is far from lip-service, this is choosing someone excellent from SRE and then letting them choose the next set of folks to come join them in CRE.
That is a serious commitment to make.
monkchips: @fintanr so much +1. I heard that. They’re literally stealing all the best SRE talent and putting it to work for customers.
sogrady: To the extent that Google can blur boundaries between support of infrastructure and apps, that becomes interesting as a potential differentiator. The key will be if they can scale it because people don’t scale as well as software and they’re already short on people.
fintanr: The trade off is that it will still be Google infrastructure specific. Some of the issues discussed in the SRE book were very much they kind of thing that only web scale folks encounter. I think Google can scale the CRE offering on their own infrastructure. Passing the lessons to a wider audience may be a bit harder.
monkchips: Aha – their own infrastructure. Good point @fintanr. Tell me more about Web Scale companies, cloud native vs on prem. Seems that could even be a weakness for AWS.
fintanr: It’s partly an operating model and approach, the web scale and cloud native folks expect to fail, design and manage for it. More traditional on prem struggles with that mindset.
monkchips: So kind of like the mistake we made with WebSphere and WebLogic in the late 90s. Designing for problems very few customers have.
sogrady: I think that’s less true today, actually, than it was back in the day.
Consider the challenges banks, for example, have had with mainframes.
They struggled a bit when online banking allowed for major spikes in transactions or lookups as customers checked from browsers when they did their banking when they got home at night.
When they went mobile, and people started checking a dozen times a day because it was in their pocket, things got even more interesting.
So yeah, not everyone needs to be 24/7, but changes in availability driven by mobile among other trends do introduce new demands for a lot of businesses.
monkchips @sogrady what? We’ve gone through internet access, search Expedia style access, mobile access, and you’re saying challenges of mainframes? Have you seen what Walmart is doing with Node and mainframes? I was thinking less literally about transactions. Whether for example the complexity overheads of microservices are going to be for everyone.
But this is good. We might even have to talk about Spanner!
sogrady: I think we’re getting into scope creep here 🙂
Which brings us to our last question: what did Next mean for Google’s competitors?
monkchips: Throw down!
It’s been funny seeing Oracle running around saying AWS looks “old-fashioned” to customers. I think the challenge for AWS is actually that Google rather than Oracle may come across technically a little crisper, more modern, you could even say “cool”. And if you combine cool with being nice to open source source, that’s going to be a thing. Because there is no doubt commercial open source companies are increasingly threatened by AWS. They want to hedge their bets etc. Now if Google can position itself well with open source, that actually serves to put pressure on Microsoft Azure, too, which has been bending over backwards to embrace open source, walk commercial open source startups into customers and so on. For IBM, which has been avoiding the “public cloud wars” by moving up the stack with its Bluemix and Watson story, Google also has some really good questions to ask, because it’s positioned similarly.
fintanr: Google put down a marker of we are here, and we are serious. We know we need to earn customer trust and we are intent on doing that.
rachel: It will also be interesting to watch the response to committed use discounts as opposed to reserved instances.
sogrady: I thought this was interesting on the competitive front:
fwiw 1 of GCP's biggest competitors helpfully emailed reporters with counterpoints in advance of #googlecloudnext so there is concern.
— Barb Darrow (@gigabarb) March 8, 2017
Whatever we all might think about what it means for Google’s competitors, there is some evidence that they were taking both Google and the event seriously.
And with that, we can bring this edition of the RedMonk Slack Chats to a close. Thanks for taking the time folks.
Disclosure: Google is a RedMonk customer, and paid for T&E to Next. Amazon, IBM, Microsoft, Pivotal and Oracle, also mentioned, are RedMonk clients.