Quantified bullets read more credibly than unquantified ones. A line that says "reduced average build time from 11 minutes to 3 minutes" lands harder than "improved build performance." Recruiters know it. Hiring managers know it. The result is that many candidates start sprinkling percentages everywhere - most of which they cannot defend in an interview.

The right move is not to bolt fake metrics onto every line. It is to dig through your actual work, find the real numbers that already exist, and use them well.

Where real numbers actually live

You probably have access to more genuine metrics than you realize. Look here:

  • Dashboards and reporting tools. Analytics platforms, error monitoring, CRM reports, financial dashboards. If you used them at all, there are numbers attached to your work.
  • Performance reviews and 1:1 docs. Quarterly check-ins often record the projects you shipped and the outcomes attached to them.
  • Project recaps and retro docs. Post-launch reviews often summarize the "before" and "after" you need.
  • Internal slide decks. If your team presented a project to leadership, that deck almost certainly contains the numbers you are looking for.
  • Your own notes. Old commit messages, ticket numbers, internal posts. Memory fails; written records do not.

Honest categories of numbers to look for

  • Scale. Users, customers, accounts, sessions, transactions, dollars, rows, requests per second.
  • Change over time. Before and after a project - conversion, retention, latency, error rate, satisfaction score.
  • Throughput. Tickets closed, deals closed, features shipped, articles published - when they are meaningful to the role.
  • Team and scope. Size of the team you led, number of stakeholders you coordinated, span of the surface you owned.
  • Time. How long a project took, or how much time you cut from a process.

What to do when you genuinely cannot find numbers

Not every bullet can or should be quantified. When you do not have a clean metric, do not invent one. Describe the outcome in concrete, defensible language instead.

"Rewrote the onboarding flow that had been on the team's backlog for two years; the new version is the one every new account now goes through."

That bullet has no percentage in it, and it is still strong. It names a specific artifact and a specific outcome (it shipped and is now in use).

The interview test

Before you put a number on a bullet, ask yourself: if an interviewer asked "how did you measure that?", could you give a credible, ten-second answer? If yes, the number belongs on the page. If no, either find a real version of that number or use language instead.

Recruiters and hiring managers are not impressed by big percentages. They are impressed by the candidates who can confidently explain where the number came from. The bullet "boosted conversion by 87%" is worse than "boosted checkout conversion from 2.3% to 3.4% over Q2" - because the second one tells the interviewer exactly where the conversation will go next, and shows the candidate is ready for it.

One pass per resume

Do not try to quantify everything in one sitting. Take one pass through your resume looking only for places where a real, defensible number can replace a vague claim. By the end of the pass, you will almost always have surfaced numbers you had forgotten you had - and you will have a sharper, more credible document than the one you started with.