$63.2k in new MRR in 3 months, with a 1.47-month payback

This is an AI SaaS company. Customers subscribe monthly or annually, starting from a low-cost trial.

The goal was to scale towards $1,700/day, over $50k a month, while keeping payback under 2.5 months.

The catch was that the account was bidding to trials. Trials were cheap (under $50), but there was no way to see which campaigns or keywords turned into paying customers. Scaling on trial volume alone risked buying sign-ups that never upgraded.

Before adding budget, I rebuilt measurement around revenue, then scaled only what paid back.

• Stripe revenue inside Google Ads. Every paid upgrade is uploaded back to Google Ads against its click ID, valued at its first month of MRR. A custom Stripe Payback Period column (cost ÷ new MRR) shows payback by campaign, ad group and keyword.
• Bidding to upgrades, not trials. I tested bidding towards paid upgrades instead of trials, and cut keywords that produced trials but no upgrades.
• Landing page testing. Two offer pages were split tested. The winner converted 8% better at a 19% lower cost per purchase and became the primary page.
• Competitor pages. Dedicated comparison pages against the main competitors, paired with a Competitor campaign running AI Max. Searchers comparing against one competitor spent 2 to 3 times more per order than the rest of the account.
• Feature and use-case pages. Keyword sets matched to a dedicated page for each core feature and use case, so searchers landed on the page that answered their query.
• Scaling with guardrails. Budgets rose weekly or fortnightly, campaign by campaign, with a 2.5-month payback ceiling. Brand and non-brand were reported separately every week so brand efficiency never hid non-brand performance.

Over August to October, $93.1k in spend returned 701 paid upgrades and $63.2k in new MRR at a 1.47-month payback. Spend was up 44% on the previous three months.

September was the biggest month in the account's history: $35.4k spend, 268 paid upgrades and $24.3k in new MRR at a 1.46-month payback.

Spend rose 38% on August, and upgrades (+40%) and new MRR (+39%) kept pace, so payback held flat at 1.46 months. The upgrade rate also improved from 11.8% to 13.2%.

October held the level. Spend eased 10% to $32.0k, and 241 upgrades brought in $21.4k in new MRR at a 1.50-month payback. That makes three straight months above $25k in spend with payback between 1.46 and 1.50 months.

Over July to September, spend nearly doubled (+95%) on the previous three months. Monthly spend climbed from $19.3k in June to $35.4k in September, the 8th straight month of rising spend and sign-ups.

Non-brand carried the growth. 83% of September spend went to non-brand, returning 100+ upgrades and $9,486 in new MRR at a 3.09-month payback. October was similar: 78% of spend went to non-brand, returning 88 upgrades and $7,259 in new MRR at a 3.44-month payback.

On the back of these results, the client moved more budget into Google Ads.