(Hiring & [Managing People) & Products]

Thoughts on the OS2 Lectures in Stanford’s CS183C
(ideas which surprised or resonated with me)
Part I of II

I. Micromanagement vs. the Jeopardy Approach

Mariam Naficy observed that if you can’t let people fail, you can’t let them grow. She brought up the notion in the context of capital: when you’re cash strapped you can’t afford mistakes and wind up micromanaging. We’re not in that situation, thankfully, but I could certainly relate to the more emotional reasons for failing to give employees adequate wiggle room. She explained that at first she knew everything going on at the company, but now she doesn’t “which is hard for a person who is detailed-oriented.” I can tend toward perfectionism and worry about how Alation will be seen if things aren’t just right. But I also want Alation to be way bigger than I can personally track with high granularity and getting there will require delegation with high-standards but less oversight. Eric Schmidt offered an excellent strategy for doing so when you’ve got smart people (as we do): early in his talk he described how on arriving at Google he would “just ask questions” (e.g. “what’s the plan for this?”) and later he explained how he got the lawyers to work out the right structure for the Google-to-Alphabet transition by telling them to “solve for x” (specifying the goals / properties of the optimal solution, but not suggesting any methodology). Both quotes strike me as articulations of a Jeopardy Approach to management — phrasing your directions in the form of a problem. I love this idea and I’ve seen it work amazingly (including for me sometimes when I’ve had the restraint and humility to execute it correctly and commit to it completely).

II. Reconciling (Seemingly) Incongruous Perspectives on Hiring

Attracting and retaining talent is top of mind for Alation here in OS2, so I paid extra attention to the speakers’ viewpoints in that regard.

Eric Schmidt offered support for the Google founders’ belief that the industry generally overvalues experience and undervalues raw intelligence and flexibility. But he also recommended that companies “hire a CFO who’s gone bankrupt.” Taken together, I think the lesson may be that someone really smart can reverse-engineer declarative knowledge, but emotional learning and values can perhaps be gleaned only from experience.

Meanwhile, Schmidt supported Bob Taylor’s approach of pitching the dream (of ARPANET in Taylor’s case) and not trying to convince anyone who didn’t immediately “get it” to join, while Jennifer Pahlka admired Todd Park’s persistent-persuasion approach which had been compared to “water on stone.” Here too, I think the perspectives are actually compatible: you can’t talk someone into having a passion, but you can talk them into taking an action. Jen understood the opportunity right away. But she needed someone to objection handle all the concerns about the job, effect on her family, etc.

III. Factoring Firefox feature functions:

Palliating palpable pain vs. offering unanticipated experience improvements

I loved John Lilly’s assessment that early Firefox users came for the popup-blocking and stayed for the tabbed browsing. The first version of Alation’s query tool, Compose, was filled with features that we thought would be game-changers capable of radically improving the lives of data analysts who were using Teradata SQL Assistant. But many of our would-be users did not really perceive themselves to need help (much as IE users in 2004 probably didn’t realize that having one webpage per window was an issue before they’d gone to the other side). Luckily for us, our browser-based tool worked well on Macs, whereas the competing tool offered a really uncomfortable experience for non-PC users. That earned us an initial cohort. Then, for mostly separate reasons, we implemented Offline Execution, which addressed this huge pain point people had of not getting their results for long-running queries when they carried their laptops between WiFi hotspots to get to meetings (thereby dropping the desktop client’s connection to the DB). That led to a huge usage uptick.

Since then, we’ve tried to always explicitly ensure that everything we release addresses at least one big problem people know they have.

(Happy Ending: many of the features we thought would help sell the product up front have in fact become sustaining features once users give Compose a try #yaywearentalwaystotallywrong)

Appendix

A few moments that totally resonated, with brief reactions in italics.

• John’s anecdote about being taught to use “let’s take this offline” as a “safe word” at Reactivity:
  Yup, learning to say LTTO totally changed my work life!
• Jen’s statement that the user experience with the government need not feel so different than in the consumer realm:
  Here here! The same goes for enterprise software (and not just if you’ve got the split market and bottom-up approach of a Dropbox or Square)
• “The problem with communication is the illusion that it has occurred”:
  That’s brilliant! Probably the single biggest lesson we’ve been learning at Alation (and I’ve been learning personally) as we scale is how hard it is to transfer knowledge in a manner that’s not too lossy.

A long tangent down memory lane turned into a post of its own: here’s Part II.