How we classify Job Title names into Job Categories

every company calls their developers something different their marketers too and don't get me started on the creative ways people describe leadership roles but hidden in this chaos is an opportunity an opportunity to understand your users better, to segment them more effectively, and to speak their language here's how we turned the mess into a system step 1 start with the raw data during the signup process, you should capture the job titles of all your new users some are clear ("software engineer"), others less so ("ninja coder supreme") the first step is to get them all in one place select case when title = '' then 'unknown' else lower(title) end as user title from users step 2 create your categories the magic isn't in collecting titles it's in grouping them meaningfully we identified key categories that matter for b2b saas dev (developers, engineers) devrel (developer relations) product (product managers, analysts) exec (c suite, founders) marketing sales design support academia step 3 build your classification logic here's where it gets interesting start with the most common patterns case when lower(title) like '%developer%' then 'dev' when lower(title) like '%engineer%' then 'dev' when lower(title) like '%cto%' then 'exec' when lower(title) like '%marketing%' then 'marketing' but don't stop there think about variations "dev" could be "developer", "développeur", or just "dev" "eng" might mean "engineer" or "engineering" some titles combine multiple keywords step 4 handle the edge cases this is where most people give up don't edge cases tell you something when title is null then 'unknown' when title = '' then 'unknown' when title in ('test', 'asd', 'a', ' ') then 'unknown' people enter test data they make typos they leave fields blank plan for it step 5 keep it maintainable break your categories into logical groups we use this order technical roles (dev, devrel, security) product roles (product, design) go to market roles (marketing, sales) support roles (customer success, support) leadership roles (exec) create or replace table user categories as ( select user id, email, title, title category from user data ) because personalization at scale requires structure because speaking to a developer differently than a ceo isn't just polite – it's effective because understanding who your users are shapes what you build for them start with a small scope test your assumptions, and refine your approach as you go the goal isn’t to categorize every job title, but to understand enough about them to make more informed decisions that's how you turn chaos into clarity one title at a time