From 29 to 129 paying customers a month, profitable every month

Rezi is an AI resume builder used by more than 4 million job seekers. Users sign up free, upload a resume, then upgrade to Pro on a monthly or lifetime plan.

The goal was to grow paid upgrades from people who weren't already searching for Rezi, at a cost per upgrade below breakeven, as organic search traffic declined.

The catch was that non-brand search barely contributed. In March it drove only about 1 in 10 paid upgrades. Keywords were on broad match, bidding optimised to sign-ups rather than paid upgrades, and mobile took a similar share of spend to desktop while rarely converting.

Before adding budget, we fixed what Google was optimising towards, then scaled only while cost per upgrade stayed below breakeven.

• Bidding to paid upgrades, not sign-ups. We verified sign-up, resume upload and purchase tracking, then made Pro upgrades the only conversion Google bids towards.
• Cutting the waste. Non-brand keywords moved from broad to exact and phrase match, and "free" searches were excluded account-wide. Ads moved to desktop only, where users could sign up and upload a resume in one session, and the click-to-upgrade rate nearly doubled from 4.46% to 8.76%.
• A clear breakeven line. We agreed a breakeven cost per upgrade with Rezi up front and only added budget once non-brand converted consistently below it.
• Competitor pages. We wrote and wireframed comparison pages against Teal, Jobscan and Kickresume, each with a switching discount, for a dedicated Competitor campaign. In July it was the most efficient non-brand campaign.
• Ad copy built on paid conversions. Copy is refreshed quarterly, keeping elements that convert to paid and cutting those that only draw clicks. The top description now runs in every ad, converting at 8.8%.
• Scaling with guardrails. Google forecast room to triple spend overnight. Instead, the core Generic campaign's budget rose 57% over six weeks, with upgrades checked before every increase. Brand and non-brand were reported separately every month so brand efficiency never hid non-brand performance.

September was non-brand's biggest month since we took over: 129 paid upgrades at 25% below breakeven, up from 29 in March.

Over April to September, non-brand delivered 687 paid upgrades and stayed below breakeven every month, averaging 20% under. It held there through two new competitors entering the auction and a full website migration.

Monthly report · non-brand, Apr to Sep 2026

In the first full month, upgrades nearly quadrupled at a similar cost per upgrade. The core Generic campaign then posted its best efficiency to date in September, 28% below breakeven, with upgrades up 14% on August.

Non-brand carried the growth. Its share of Rezi's paid upgrades rose from about 1 in 10 in March to more than half in September.