Interview GPS - autopilot for structured candidate interviews

[triedandtrue]

Interview GPS is an autopilot for structured candidate interviews. A hiring manager enters their job requirements, and the web application automatically designs and coordinates the best interview process. The system recommends the best questions, collects interviewer feedback, accounts for uncertainty/disagreement, and summarizes the tradeoffs between candidates

Market Size: In the US, there are 3.2M job openings each month in professional services (finance, healthcare, education, etc.). Companies spend an average $4K per hire, but nearly half of hires fail within their first 18 months. 3 in 4 companies say they've recently made a bad hire

Business Model: freemium and self-service. The paid tier is currently $15/role/mo. and supports customization (e.g., custom questions). In the future, it will support enterprise integrations (e.g., with an ATS, Slack, etc.). We're focusing inside sales to HR managers at the 107K firms in the United States with 100-500 employees. Assuming avg. employee turnover and headcount growth, the annual contract value is ≥ $200 at these companies

Competition: the big players in the space are the applicant tracking system (ATS) providers, like Greenhouse, Lever, Oracle Taleo, etc. Many of them include features for manually creating interview kits. However, they’re platforms focused on process/point solution integrations. There are also niche players for certain kinds of assessments, e.g., HackerRank for programming interviews

Product Insight: It’s impossible for hiring teams get enough at-bats to get good at hiring for a particular position/situation. Therefore, you have to mutualize the learnings across teams/companies. We use machine learning to improve recommendations based on interviewer feedback, question reliability, etc.

Team: I’m a technical solo founder with 10 years of experience building analytics and line of business applications. This includes performing people analytics R&D at Bridgewater Associates, the world’s largest hedge fund

Progress: In a month of part-time work, I've launched a beta focused on cultural and behavioral interviews at www.interviewgps.com. I’m currently bootstrapping the company, but applied to the YCombinator W20 batch

P.S. I’m looking for beta testers! If you’re interested, DM me and I’ll send you a promo code.

 

[truthevenifyouhateit]

@triedandtrue Questions:

What qualifies a good vs bad hire?

Since these questions are behaviorial and cultural rather than quantitative, how is the platform distinguishing between "correct" and "incorrect" answers?

I think the only way this would scale well with societal changes is if you are using AI/machine learning. Are you incorporating these technologies?

 

[triedandtrue]

@truthevenifyouhateit Appreciate the questions.

The best way to measure good vs. bad hire is using retention and performance review data. In the future, we want to integrate with human resource information systems (HRIS) such as Workday to track and analyze post-hire outcomes

The interviewers rate the candidates across the values/competencies and overall using interview scorecards. We will also be releasing rubrics with examples of good answers vs. bad answers

We don't use AI/machine learning to automatically rate the responses. Although some companies are trying to for video screening

Instead, we use machine learning to learn which questions are best for each kind of role. Question "goodness" is based on signals such as: 1) directly asking interviewers how they liked the question, 2) the empirical reliability of the question. We use optimization techniques to recommend optimal interview structure