WP005 - The full Chef burger, with a side of THE DIGITAL
Software Defined Talk Members Only Podcast White Paper Exegesis, episode #05
Coté & Brandon Whichard, August 2017.

This week’s paper is sent in by Matt Ray: “Continuous Automation for the Continuous Enterprise.” There is no lead-gen, so you can go directly to the PDF.

To hear our analysis and commentary, become a member of Software Defined Talk.


This is an excellent white paper, useful for both buyers, the Chef sales process, and the peanut gallery (us and anyone who’s actually listening to this podcast). 

For buyers, I think there’s a lot to like in here. Metrics are always good, and the capabilities map in figure 5 is a simplified alternative to all the DevOps Reports stuff. This paper gives you way to think about what “cloud strategy” is and enough re-usable metrics to put into your own PowerPoints. It’s also packed up well enough to email around internally to people you need to convince. To some extent, you could even email it to your Morlocks and say “implement this.”

In the sales process, your inside-sales people could easily send it to someone stand-alone to make the case to setup a face-to-face meeting. The paper follows the market’s adaptation of Presentation Secrets of Steve Jobs/hero’s journey by throwing up a desire, an enemy (slow moving IT), a journey into the unknown (transformation!), and a triumphant way and results that the young Skywalker can transform themselves to awesomeness. What’s key to this rhetorical technique is to use industry studies to show that the way you do IT currently doesn’t match business need, then show that others have fixed this problem, then describe the vendor-free methods and technologies these companies do, and then at the very end give the product pitch. If you can throw in a maturity model or some metrics to baseline and judge the hero’s journey, you get bonus points. All of that is done here, very crisply.

Now, does it show that Chef are the only people you can give money to get The Gift of the Goddess? If you stare at it long enough, and have enough competitor context: not at all. But, in this space, that case is pretty impossible to make unless you table-flip the entire model and suggest that there’s a “turns out!” angle on it all. In the late 2000’s Puppet and Chef had some good USPs, but they were pushed down the stack of EBIT(A) when containers came in, container orchestration folks and Kubernetes, and cloud platform trouble makers like Pivotal and other PaaS-infrastructure software market marauder. Their re-alignment around a Chef platform (see below) is a good reaction to this.

(We try to do this at Pivotal with out “value-line” argument, which is high-level value-prop’ing mysticism - which, of course, I think works very well - as do public cloud providers in the area of “why are you managing data centers at all?”).

There are some potential methodological holes and traps, but without digging more, who knows. It’s not too useful for stealing content because it’s strongly Chef aligned; in contrast, to the DevOps Reports (which is nominally Puppet funded and aligned) is good for lifting content.

Also worth noting is our extreme bias for all this: Matt Ray works at Chef and is not only a co-host of Software Defined Talk but a long-time friend. We want him and, thus, his company to do well with all of this; he’s hella vested and an ancient strike-price, bruh!

Sidebar: The Best white paper I’ve ever read (at the time)

First of all, it’s been awhile since I read it, but Chef’s 2015 white paper on Standard Bank’s “DevOps Journey” is one of the best white papers I’ve read. It has everything: helpful advice from a customer, a concise story, and it hustles Chef without trying too hard. I literally picked this paper up as a print-out at a Chef booth, and carried in my bag for awhile ‘till I got bored one day and read it. It was great! CHOP CHOP!

Back to their new white paper.


  • From what I can tell, Chef, Inc. (as it were) makes its money in two ways: (1.) helping people speed up their configuration management/provisioning tasks, regardless of how “cloudy” they are, and, (2.) genuinely transforming how they do IT to be more DevOps and “cloud-native,” as we like to say in Pivotal land. There’s all sorts of new products and things Matt Ray can tell you about.
  • Chef was one of the original DevOps monetizers and, along with Puppet, became synonymous with DevOps for a long, long time. For many people, the Venn diagram of “Puppet and Chef” and “DevOps” was a perfect circle.
  • So, this paper is the continuing journey to tell organizations that, if you want to do the digital, you should get Chef in. A metrics paper is all about 

  • First, it’s taken me about 40 years to understand what “quartile” means, i.e., “top 25%.” “Percentile” is another that bugs me out, e.g., “99th percentile” means the opposite of what you think at first.
  • Second: good Lord, how could you possibly model out that an investment in “the app” made this difference?
  • Without getting into the methodology too much - I mean, who has the time - you can track down through a few McKinsey pieces to find the McKinsey Digital Quotient study: 2,500+ companies used, judged by 18 practices measured. One of the source lines for the above chart says they used CapitalIQ as well, an excellent date store for financial data. Presumably, with 2,500+ companies they put together some search criteria and did some “small data” analysis to categorize companies and then slice it with financial returns over 5 years. That’s a good “wet-finger in the wind” model. ¯\_(ツ)_/¯
  • Whatever, I mean, I’d use this chart. It does some backing up of an intuitive point. The DevOps Report does similar work, but with more “science-ing” behind it that I’ve never been able to understand, despite Nicole lavishly and kindly offering to explain it to me.
  • What’s also valuable about this chart, in the context of this Chef white paper, is that it’s sourced from somewhere else. So, you can use this chart without have to connect to Chef. E.g., if you’re selling a rival product, you can go back to McKinsey and not be putting competitor logos in front of prospects.
  • It’s worth spending a lot of time on this topic because the whole PDF hinges on it: doing the digital increases your treasure. That’s an old Bain value-prop.
  • Their prescription for the tech strategy (pg. 5): do cloud-native (use of cloud and bursting), microservices, use containers.
  • Defeating shadow IT - page 7 extracts from the Chef survey to, very loosely, show that using cross-function teams will satisfy The Business’ need to do more, more quickly, thus, subverting shadow IT and making mainstream IT more valuable. This is an important part int he argument being built.
  • That said, “more likely” stats from surveys can be tricky: 3% is 200% more than 1.5%, but it’s still just 3%. On the other hand, 80% is 400% more than 20%, which would be more meaningful. Often, you use “more likely” when the spread between losers and winners is small. So, when using “more likely” statements, you need to put in the actual numbers to be completely honest.
  • Even in this example, those numbers aren’t too impressive: 17% better is pretty tight. How much more does it cost in loaded risk and budget to get that 17%? 
  • But, the point is made nonetheless.
  • Page 9 introduces the “show what do you need to do to get all this awesomeness” (figure 5) in a capabilities chart. This is good stuff and well done.
  • Metrics are cat-nip for organizations looking to shift from “traditional” to “new-fangled” IT:

  • Having actual numbers in their metric suggestions is awesome, and bold in a good way. Often, when paper suggest metrics they don’t give you actionable ranges like this, suggesting that it must be tailored to your business needs and such. I think that’s largely BS (most organizations have the same problems and need the same things from IT), so it’s great that Chef just lays out some figures. Of course, you’ll notice that these are sourced from the DevOps Report.
  • “On demand” is a little hallow: is it 5 milliseconds, 30 seconds, or 3 hours?