WP002 - Not a [Very Good] DevOps Talk
Software Defined Talk Members Only White Paper Exegesis  @cote & @bwhichard, July 2017.

To hear our podcast talking about this, become a Software Defined Talk patron so you can access the members only podcast. Otherwise, just enjoy Software Defined Talk itself.

Not actually a DevOps Talk, Coté

This week, we take a look at Coté’s (my) 2016 “stump speech.” The talk that he gives by default and most often at conferences, MeetUps, EBCs, and other meetings where I’ve been asked to “talk about DevOps/cloud-native/etc.

We’ll use the most recent recording (given July 10th, 2017 at the Austin DevOps Meetup); here are the slides as given.

Why you care

At least in what I’ve typed below, I’m trying to show “how a PowerPoint hustler works,” so to speak. I like hearing how people actually do things, tactically, in the creative process; so here’s a bunch of that!

More broadly what this talk tries to do is de-hype all the hype around cloud, DevOps, cloud-native, and containers by explaining what you’d actually do with all this stuff to make your organization better (either to run better, make more profit, etc.). 

While I don’t spend a lot of time on the business justification, much of the time is spent going over “how do we actually get from our sorry state to the unicorn utopia.” That kind practical advice for the rest of us has been the theme of my work for a long time.

Methodology - My process for putting together a talk

I usually write up a prose version of talks before putting them into slides. I more or less did this here with my Cloud Native Journey and then Crafting Your Cloud-Native Strategy booklets. You can see me starting to do this for my WIP Cloud-native Enterprise Architect talk.

I rehearse a new talk about twice, sometimes three or four times. I hate rehearsing. It’s personally embarrassing to find out how poorly prepared I am and it seems like a waste of time. But, it’s important to do that to make sure it all flows together like I think it will, and that the time works. Once I’ve done it a handful of times on my own - sometimes in hotel rooms the morning before I’m supposed to talk! - I just give it over and over again and view that as rehearsing. Often, I’ll find “small” venues (like meetups) to test it out in front of a live audience; I’d intended to do that here, but the scheduling got pushed way back and I’d already given this talk many times.

The “methodology” is based on me hunting out anecdotes, surveys, and all the “facts” that provide the skeleton for the corpus. Think of it like that middle-school research paper exercise where you write down a bunch of facts (and their citations!) on index cards. I do this daily, just with anything that might be useful some day. For the broader points - like how I want you to think about what DevOps “really is” and what cloud-native is, I pull both from official propaganda (when it’s Pivotal stuff), but mostly from my own thinking and conversations I’ve had with people. A lot of what’s in here was driven by discussions I had with Andrew Clay Shafer the first year I was at Pivotal, talking with large organizations trying to figure out all this stuff, how Pivotal Labs operates, and, you know, my 20 year history studying this stuff.

There are times when I’m dubious by omission, esp. the obvious one of not mentioning competition or, mostly, not drawing case studies from competition’s customers. But, I try to cite all the things I’d want, e.g., in one of these exegesis exercises in the foot-notes, or at least link to them. That lack of detailed citation as a pet-peeve of mine came from my frustration started the lack of that in most of the material I’d read and use while doing strategy and M&A at Dell. People would always put something ludicrous like “Source: IDC.” Then you’d hunt down the research and it too often didn’t say exactly what they were claim or was missing some context that made their argument weaker.

Macro, content, etc.

The goal is to explain what I think the point of DevOps, and, thus, cloud-native, and thus the best “cloud strategy” is. Over the years I’ve found that most explanations of “digital” are too broad and not laden with enough real-world, G2000, “boring” enterprise examples. They rarely ever have actual tactics and mini case-studies of failures wrapped in. I don’t want to tell people what the end-state looks like, I want to tell them how to get their shit together so that they can start to get to that end-state.

Also, of course, I want them to be interested in Pivotal. I want them to at least think that the reason you do DevOps, cloud, etc. is to improve how your organization does software and, indeed, that it’s not only important to improve how you do software but that your “business” needs to. I don’t hit up this imperative as much as I do in other talks - this is the “full nerd” version, but the idea is scurrying around in there.

Micro

Commentary on the recording, in order with occasional time codes.

  • (09:00) “If you came for… seniors and pets” - this is an allusion to the pre-recorded call in show from the Mr. Show. The hosts of Reconcilable Differences joking make reference to this as I do here. I’m copying them, as I do much of Merlin Mann’s crew’s little asides.
  • I go over what the talk is about. I spend too much time on the cover slide. Part of my presentation style is to tell people over and over again what I’m going to say, what I’m saying, and what I said. It’s good to sprinkle that with what happens. So, here I’m explaining what we’re talking about here: improving how you do software. I also explain who the audience should be that this applies to: don’t want to talk DevOps to storage admins.
  • I like to throw in terms like “digital transformation” and stuff a lot to (as we’re doing here) defend the use of jargon by explaining it. My belief in de-jargoning words by understanding how they works comes, this comes obliquely from my E325M english class at UT, where Prof. Trimble told this anecdote-cum-aphorism: one semester when we were reading Mencken, a student said he used too many book words. I said: get a dictionary. Along with reading all of Hunter S. Thompson’s works and letters available at the time, this class really formed my approach to writing, and my style.
  • I tell a story about renewing my passport. I’m very “free associating” in my professional life (my family doesn’t appreciate it, so I do it less at home). I try to always use recent, relevent examples of “boring,” enterprise applications that cloud-native can apply to. Things that all of us have experienced. We already know about hot dogs.
  • (01:50) “Or as we would say around Pivotal land, ‘cloud-native’” - there’s a vendor battle around what to call all this stuff. DevOps would be easy, but it’s hard to own as a name and associated sales-cycle. “Cloud-native” either means “container orchestration” (hence the foundation around it), or the way I use it here, to mean “DevOps, plus using a cloud platform (like Pivotal Cloud Foundry), plus the business knowing what to do with this small batch process.” That is, I see DevOps as a sub-set of “cloud-native.” You’ll noticed that I cycle through these two terms as if they’re synonyms, often saying them together. Occasionally I talk about DevOps as itself. I mean, you gotta call it something.
  • “Like Boyd was describing…” - the host of the Austin DevOps Meetup, Boyd Hemphill, has spoken before me. He’d given a talk about a disasters job he had trying to manage rapid change. I try to always see the talk(s) before me and (for better or worse) make call backs to them if the same audience has seen them. I don’t like getting up early, but I do this with keynotes sometimes. I feel like it’s respective and keeps up the continuity for the conference. It also ingratiates you to your peers and fellow speakers. As ever, I’m trying to throw out as many examples and citations as possible. I loathe bold claims that aren’t backed up with proof, if only “this thing happened to someone once.”
  • “Several people here said they were hiring… they lowered their referral bonus this year, so I’m not too motivated” - this is a good joke to get out at first. I didn’t really set out to do this when I started talking, but I try to quickly get the audience to laugh and start portraying myself as a goof-ball. Getting people to laugh relaxes them (and me!). This makes them listen better and (if you’re being wicked) be easier to convince. The goof-ball part is the same: I like to quickly disarm people’s defensive (and, thus, offensive) behavior when this weird person has come into to talk with them. There’s a lot to be said for this kind of Bush Jr., awshucks couching. Also, this joke about hiring starts to get the mercantile “bias” problem off the table. Of course I’m here to make you think like I do and, if you’re a qualified lead, figure out how to give Pivotal money. Let’s get that off the table so we can start talking. (You’ll notice that I mess up on time and totally glide past the soft-pitch at the end - this happens frequently!)
  • “I hadn’t bathed for a few days” - again, I’m a goof-ball. Put down your shields…and please, don’t do anything confrontational, I’m a pussy cat.
  • Bio slide - establishing my credibility, my background, and doing a bit of CTA’ing. All I really want people to do is subscribe to my podcasts. I want people to know that I’ve spent 20 years on all this stuff.
  • “I can align things” - I stopped coding in 2006. I really don’t know how to code anymore, but I put a lot of thought into presentations, even ones I give over and over. People generally sneer at “decks,” but I like to think I show them what can be done with them if you slice and dice it all together into a story - and repeat yourself a lot. John Willis has a good description of my approach here in his 2017 DevOpsDays Austin talk.
  • The Register… See all the wonderful ways that people who don’t know how to spell…” - it’s always good to troll the trolls. This also hopefully drives CTA to more content.
  • “There’s always one guy who’s here…” - I like to heckle the audience a bit, esp. the people I see frequently.
  • “I just posted this…on my Twitter account” - one of my new tactics to get people to follow me in Twitter is to tell them to go to my Twitter account to get the slides, hopefully following me. 
  • Castlevania and Dune - just some random, nervous talking from me, but pulled back to the narrative: so you’re gonna change how you computer, what should you be doing? Here, the problem is laid forth in a joke for old people who would remember this scene in the meeting (I even mix up “Picard” on purpose here). This slide has proven to be some of my best PowerPoint joke-work.
  • “I brought my son” - I don’t usually do this, but I was hoping he’d be able to watch whatever it is I do. There was another son brought there and they were talking too loud during the first talk, so I’d put him the hallway. Many road-travelin’ speakers like me dream of taking their kids: John Willis and Andrew Clay Shafer have pulled it off, even getting their kids to speak at conferences.
  • How software is done wrong, grocery stores and submarines - everything about how I do things comes together here in these two slides: The Big Lebowski, 80s movie references (all nerds like Sean Connery), and things people misunderstand about how software is done. These “one big picture” slides are a departure from my usual text and chart heavy slides. This year I wanted to try out what all the kids are doing. I’d noticed that I just sort of use the slides as wallpaper for my talking (something I picked up from Bruce Sterling’s SXSW talks and hearing John Roderick talk about the few times he’s down presentations) so I wanted to test the theory that I don’t really need much in the slide, unless the data in the slide is the point. My thoughts on this approach so far: ¯\_(ツ)_/¯
  • The 80s submarine analogy - this is an example where I really like talking. This story should be cut shorter, but I give this long version pretty much every time.
  • (11:38) “I bring these two things up…” - often times, I’ve lost the narrative thread of what I’m talking about. I feel like the audience is like, “what were we talking about?” as well. I’ll commonly forget what the next slide is, and sort of have to shuffle through to get back on track.