How I built a quiz matching engineers with companies based on shared values

TL;DR - I turned an internal workflow (matching engineers and startups by values) into a lead gen tool, using Google Sheets, Tally.so, Zapier and Notion.

The average person spends 90,000 hours at work over their lifetime. Where people choose to spend those hours has a huge impact on the quality of their life.

We all have a unique set of values, which themselves aren't good or bad. What matters is the overlap between your values as an engineer and a startup's values - this will inform how happy you are at work.

At Commit, we spend all day, every day matching engineers and startups. Some factors are easy to match on - seniority, shared tech stacks, familiarity with particular projects and feature tasks. The hard part is understanding values alignment - a desire for autonomy vs for mentorship, an orientation towards solving hard technical problems vs messy customer problems, attention to detail in code quality, a bias for action, the belief that decisions are best made via vocal debate or through written communication with built-in documentation for posterity.

This is made all the more complex when you ask startups and engineers to tell you what they value in a work culture. You get messy, unstructured data in return - hundreds of responses, each using different terms, with no universal shared meaning.

We’ve built an internal map of this unstructured data. When a startup tells us that their culture mirrors a sports team, we know based on the specific questions we ask if they mean that they prioritize collaboration (no ego, it doesn’t matter who scores the goal), competitiveness (we play to win), or adaptability (everyone should be prepared to tap into another player’s role).

We use the internal map every day in matching engineers and startups, but saw an opportunity to open source it to help engineers everywhere. I took the questionnaire that all of our new engineers complete at Commit, and turned it into a lightweight quiz that’s open for anyone to take.

This internal map lives in Google Sheets. We have a universe of values, the questions/signals that they're held by engineers and startups, and based on this map we connect values to startups. Eventually we'll scrape job postings, but for now it's a manual process.

I started out building this quiz in Typeform but quickly ran into a blocker. Typeform works really well when you have a binary output (ie. you're a cat person, or a dog person), but doesn't work well when you have 10 possible values, and people can hold any combination of them. I had initially planned to use Typeform's calculation tool, give a score that increased by a magnitude of 10 for each value (ie. bias for action = 1, high code quality = 10, enjoy solving hard customer problems = 100), where every combination of values would then give a unique numerical output. I could have used Airtable to house the legend, but realized I would have over 4,000 different possible results, and I would have to type out descriptions for each of those 4,000 results.

I coincidentally discovered Tally.so around this time, and it solved that problem. Using conditional logic and their calculate tool, I was able to assign every value a unique "calculated field" (which I just named the title of the appropriate value), set it to 0 by default, and then when people answered a question that indicated they held that value, I used conditional logic to assign that calculated field a score of 1. In the results page, I wrote up descriptions of all the values, and then used conditional logic to hide the block of text that corresponded to a particular value when that value's calculated field score was 0.

Tally.so doesn't auto-send email results upon quiz completion, so I connected Tally.so with Zapier. Every time someone completes the quiz, Zapier will send out a pre-drafted email. In the body of the email text, I was able to inject the calculated field scores for each value, so people could see which values the quiz had indicated they held, and then I linked to a Notion doc that summarized all of the values and linked out to all of the companies that shared those values.

The overall process was a bit hacky, but the use of binary 0/1 values actually worked well with my target audience (software engineers)!

I can't recommend Tally.so highly enough - there was a bit of an initial learning curve, but it was extremely useful and saved me from having to pull dev time to build this quiz for me.

Feel free to check out the quiz if you're interested - https://tally.so/r/nPRZP3 - and reach out if you have any questions about how to use any of the tools (Google Sheets, Zapier, Tally.so, Notion) or suggestions on how I could have made this less manual!