Achievement Bullets That Actually Prove Impact (With Before/After)

The bullet is where resumes are won and lost. The format choice (chronological vs. hybrid vs. functional) decides the frame; the bullets decide whether the recruiter believes you can do the job. Most bullets are bad in a predictable way: they describe the job that existed, not the outcome the person produced. A bullet that says "Responsible for managing a team of engineers" describes a job description; it doesn't describe a person.

This post walks through the only bullet formula that consistently survives a 6-second scan, plus 12 before/after rewrites across five common functions. By the end you'll have a template you can apply to every line on your resume.

The formula: Verb · Impact · Method

A bullet that does its job has three parts:

1. Verb (past tense, specific). "Built", "Owned", "Reduced", "Migrated", "Closed". Not "Was responsible for", not "Helped with", not "Worked on". The verb is the first word the recruiter reads — make it earn its place.
2. Impact (a number, a name, or a magnitude). "by 34%", "for 18 enterprise customers", "across 4 regions", "from $1.2M to $4.8M ARR". This is where you prove the work mattered. No number = no proof.
3. Method (one short clause). "by deprecating two legacy services and consolidating onto Redis", "by running a 9-week onsite-discovery program with the top-30 accounts", "by rewriting the bullet copy for 12 paid-search campaigns".

Most bullets stop after the verb. Strong bullets do all three.

A template:

[Past-tense verb] [object/scope] [quantified impact] by [one-sentence method].

That's it. It's not a flashy framework — it's just the minimum a recruiter needs to believe the bullet.

Engineering bullets

Before

"Worked on the checkout service for the e-commerce platform."

After

"Rewrote the checkout service in Go, cutting p99 latency from 1.4s to 280ms and lifting completed-checkouts by 8.2% over 6 weeks, by replacing 3 synchronous downstream calls with an event-driven flow."

The before version describes a job; the after version describes a person who shipped something. The numbers (1.4s → 280ms, 8.2%) are falsifiable — the recruiter can probe them in the screen. The method ("3 synchronous downstream calls → event-driven flow") proves you actually did the work, not just sat on the standup.

Before

"Mentored junior engineers."

After

"Mentored 4 mid-level engineers through promotion to senior over 18 months, by running biweekly 1:1s with explicit promotion criteria, code-review pairing, and OKR alignment with their managers."

Mentorship is the hardest thing to make tangible on a resume, but specificity wins: "4 mid-level engineers" "through promotion" "over 18 months" turns a vague claim into a track record.

Before

"Improved the CI/CD pipeline."

After

"Cut CI build time from 18m to 4m and deploy time from 14m to 90s across 22 services, by parallelizing the test matrix, caching dependency installs, and replacing serial deploys with a fan-out Helmfile rollout."

Notice that "improved" is vague. "Cut from 18m to 4m" is not. Any verb that can be replaced by a more specific verb should be.

Sales bullets

Before

"Met or exceeded quota every quarter."

After

"Hit 138% of annual quota ($2.4M ARR booked vs. $1.75M target) by reorienting from inbound MQLs to outbound on a hand-picked list of 80 mid-market accounts in fintech."

"Met or exceeded quota" is what every salesperson says. Numbers + the strategic shift behind them is what one salesperson says.

Before

"Closed enterprise deals."

After

"Closed 7 net-new enterprise logos averaging $480k ACV in the financial-services vertical, including the first ResuAI ANZ enterprise customer (Westpac), by leading 9-month POCs with security-first technical onboardings."

Naming a customer (when it's public-knowledge that they're a customer) is the single highest-impact thing a sales bullet can do. It's instantly verifiable on the company website or LinkedIn.

Before

"Built out the SDR motion."

After

"Hired and ramped 4 SDRs in 6 months, growing weekly outbound activity from 200 to 1,400 touches, with average new-SDR ramp to quota dropping from 14 weeks to 8."

Specificity about who you hired, what changed, and by how much converts a manager bullet from "ran a team" to "ran a team that did X."

Marketing bullets

Before

"Managed paid social campaigns."

After

"Owned $1.8M annual paid-social spend across Meta, TikTok, and LinkedIn, dropping blended CAC from $86 to $52 over 12 months by killing 3 underperforming creative concepts and reallocating to a single-thread BOFU retargeting pod."

The before version sounds like an intern; the after version sounds like a person who knows where their dollars went.

Before

"Increased website traffic."

After

"Grew organic blog traffic from 12k to 89k monthly sessions over 9 months by shipping 34 long-form posts targeting bottom-funnel commercial intent ('best resume builder for engineers', etc.), with 7 posts landing on page 1 of Google."

Specificity about the kind of traffic and the kind of content matters. "Increased traffic" could mean anything; "bottom-funnel commercial intent" tells the reader you understand the difference between top-of-funnel vanity traffic and traffic that converts.

Before

"Worked on email marketing."

After

"Rebuilt the lifecycle email program from a 9-email broadcast cadence to a 22-step behavioral sequence in Iterable, lifting trial-to-paid conversion from 11.4% to 16.8% over the first quarter post-launch."

"Worked on email" is unfalsifiable. "11.4% to 16.8% trial-to-paid in Q1 post-launch" is a specific claim that the recruiter can probe.

Operations bullets

Before

"Improved fulfillment operations."

After

"Cut the average ship-to-customer time from 4.1 days to 1.8 days across 3 fulfillment centers, by renegotiating two 3PL contracts and migrating last-mile from USPS Priority to a regional carrier blend (Veho + LSO + DHL eCommerce)."

Operations bullets are often the hardest to write because the work isn't "feature-shaped" — it's process changes. The way to make process work specific is to name the exact change (carrier names, contract renegotiation, etc.) and the exact metric (days, percentage points, dollars).

Before

"Reduced costs."

After

"Reduced annual compute spend by $340k (38% YoY) across the data-platform footprint, by rightsizing 23 Snowflake warehouses, deprecating 4 dormant dbt jobs, and moving 2 batch pipelines from hourly to 4-hour cadence."

"Reduced costs" is the single weakest bullet in operations. Always specify what costs, by how much, and how.

Design bullets

Before

"Redesigned the onboarding flow."

After

"Redesigned the new-user onboarding flow (8 screens → 4 screens, no skip-able steps), lifting D1 retention from 38% to 51% and D7 from 19% to 27% in the A/B test that ran for 3 weeks across all new mobile signups."

Design bullets need to push hard against vague verbs like "improved", "elevated", "delighted." The work being designed isn't the bullet; the user-behavior change that resulted from the design is the bullet.

Before

"Built the design system."

After

"Built v1 of the component library in Figma (62 components) and shipped the React implementation (47 components covering ~88% of UI surface), cutting new-feature design-to-shipped time from 6 weeks to 2.5 weeks across 9 product squads."

The "62 components in Figma + 47 in React + 88% of UI" detail makes the bullet feel real. The "6 weeks → 2.5 weeks" anchors it in an outcome the business cares about.

The "could this be on someone else's resume?" test

Before every bullet ships, ask one question: could I drop this bullet, unchanged, on another candidate's resume in the same function? If yes, the bullet is too generic. Real bullets carry specifics that can only have come from your work — a customer name, a tool, a dollar figure, a date.

A bullet like "Improved team velocity" passes onto any engineering manager's resume. A bullet like "Lifted average sprint velocity from 38 to 56 story points by pairing with the platform team to remove 4 blocking dependencies before the quarter started" only fits the person who actually did that work.

This test catches 60-70% of bullet-level issues by itself.

A few small mechanics

• Past-tense for old roles, present-tense for current. Don't mix.
• No personal pronouns. "Led", not "I led".
• End bullets without a period (the tradition is mixed, but no-period reads cleaner on a single-page resume) — just be consistent.
• One line per bullet on screen. Two-line bullets get skimmed. If the bullet wraps to a second line, cut.
• 3-5 bullets per role, not 8. Bullets after the 5th get ignored. Pick the strongest.

Now do this for your resume

1. Open your resume side-by-side with this post.
2. Bullet by bullet, ask the "could this be on someone else's resume?" question.
3. For each bullet that fails, apply the Verb · Impact · Method formula. If you can't think of a number or a name, the work probably didn't matter as much as you remember — pick a different bullet to write instead.
4. Run the rewritten resume through our analyzer with the target JD. The per-bullet rewrites will catch anything you missed.

Most resumes go up by 10-15 score points just from this exercise. The format is the frame; the bullets are the picture.