The Primacy of Process

by Tim

Navigating the seas of the startup world is never a simple task. There’s a whole big world of possibilities out there, and it’s hard to know if you’re following the right route. The dismal success rate for startups shows how tough the going can get.

In the storms of hard decisions, uncertain futures, or plain bad luck, many companies lose their way. The instinct to bail the overflowing water and let the ship drift for the moment is strong, and sometimes reasonable. Most times, however, abandoning planning and process spells doom for a startup. As the COO, creating and enforcing processes is part of my job. I use processes to tame the disparate parts of the business, channeling the torrent of activity my co-founders unleash daily1.

In order to mark and measure our path to future greatness, we recently adopted a planning system called Objectives and Key Results (OKRs). OKRs are a measurable way of setting goals and evaluating progress company-wide, from the freshest hire to the highest executive. Every employee has at least one objective, which in turn has 3-5 key resources that measure the objective’s completion. OKRs should be ambitious and a little uncomfortable. On a 0-1 scale, the right score is about 0.7; a lower score indicates poor completion, and a higher score indicates poor goal-setting. OKRs started at Intel, then spread across the corporate world. Their most famous advocate, John Doerr, brought the system to Google in the company’s early days; they still use the system today2.

Here are our Q4 2014 OKRs:


Objective: Public launch

Key Results:

  • 100 landlords using Castle
  • 50% smooth onboards3
  • 90% of landlords would recommend Castle to their peers


Objective: 100 landlords using Castle

Key Results:

  • 750 leads generated
  • 75 leads closed
  • 25 inbound leads generated


Objective: Public launch

Key Results:

  • 95% production site uptime
  • 15 new user stories4 based on feedback from beta users deployed
  • 10 automated tests written
  • System for auto-deploying code to production built
  • System for creating daily backups of the production database built


Objective: Prepare Castle for public launch

Key Results:

  • 2 interns hired
  • 100% of QA user stories completed
  • 250 leads added
  • 25 leads closed

The system has already proven useful, forcing us to set specific targets for sales and development—a tech startup’s two most important activities. As an engineer, I have an unending appreciation for all things measurable, and I look forward to measuring (and blogging about) our progress at the end of the year.

  1. Once during an interview, the interviewer pointed to his bike and asked which part represented me best. I chose the frame, since it holds all the moving parts together in a discrete and firm manner.
  2. I recommend watching Rick Klau of Google Ventures explain the system and its history at Google here.
  3. A smooth onboard is one in which the user has no questions from the start of signup to the first rent collection cycle.
  4. User stories are how we describe features. For example: “As a tenant, I can activate automatic rent payment by text.”