📎 Webclip
Hiring programmers with a take-home test
Basecamp argues that there is no perfect hiring process for programmers, but that take-home assignments are fairer than whiteboards, brain teasers, and long screening pipelines. They use them only after narrowing the applicant pool, and they keep the assignments small enough to reflect real work without demanding unpaid multi-day effort.
The examples in the page show two late-stage assessments, one for Security, Infrastructure & Performance and one for Research & Fidelity. In both cases, the task was based on a real problem, candidates had about a week, and the assignment served as one step in a longer evaluation that still included interviews and team review.
Reading notes#
- Basecamp rejects whiteboard grilling and brain teasers as hiring methods for programmers.
- The company asks for real code only from late-stage candidates, not from everyone who applies.
- Open source experience helped in the early days, but it would exclude many good programmers.
- Previous-job code is often unavailable or hard to attribute to one person.
- For one opening, Basecamp received more than 1,300 applications and could not review that many code samples.
- Candidates are narrowed first through cover letters and, to a lesser extent, resumes.
- The Research & Fidelity opening gave 40 people the take-home test, which Basecamp later judged too many.
- The Security, Infrastructure & Performance opening gave 13 people the take-home test.
- Basecamp says future take-home groups should be fewer than 20.
- The company rejects unpaid assignments that demand 20 to 40 hours of work.
- For programmers, Basecamp uses assignments estimated at 3 to 5 hours.
- That commitment is still substantial, so it is reserved for candidates already in serious contention.
- Basecamp does not use recruiter screening mills, automated resume tools, or interview travel.
- The company says it spends weeks preparing the job opening and the hiring process.
- The rules are stated up front: applicants need a tailored cover letter and must be ready for the assignment if they reach the top group.
- The SIP assignment was based on a real issue in rack-ratelimit and asked candidates to implement a feature.
- The task was not an abstract algorithm puzzle and instead resembled work on the job.
- Jorge Manrubia’s submission included tests and documentation and was considered comprehensive, though possibly more polished than necessary.
- Jorge spent closer to 7 to 8 hours, and Basecamp says this extra effort was not required.
- Completing the assignment was necessary but not sufficient to get hired.
- From 13 candidates, 10 moved to a first phone interview, 6 to team interviews, and 2 to an extra interview.
- The R&F assignment was based on enhancing datalist autocomplete for HEY.
- Nabeelah Ali’s solution kept everything in a single page and was less elaborate than Jorge’s but still strong.
- Nabeelah also spent more than the estimated 3 to 5 hours, partly because she refactored and polished the work.
- She says she would rather do that than a live coding challenge.
- Basecamp argues that the alternative to take-home work is often travel, whiteboards, or abstract assessments.
- Other options mentioned include pair programming and code review, but the page says each has limits.
- The post says hiring and applying are hard, and that judging programmers without seeing real code risks bias, credentialism, whiteboard puzzles, and brainteasers.
