The published situation
What the source says the organization needed to change.
MoneyMe was not trying to optimize for a cheap form submission. Its commercial outcome was a funded personal loan, which happens after an application moves through checks and approval. The campaign therefore needed two things at once: enough creative range for an automated cross-channel system to work with, and a measurement path that returned funded-loan outcomes to Google Ads.
How the published programme worked
Send the commercial outcome back to the ad platform
The case says MoneyMe used offline conversion imports so campaign optimization could learn from completed, funded loans rather than stopping at an online application. That is the most transferable part of the story. Automation can only pursue the outcome represented by its conversion data.
Give automation meaningful creative choices
MoneyMe developed multiple assets, including campaign-specific video, to explain the product and establish trust. Performance Max distributed those assets across Google inventory. The machine handled placement and bidding; the advertiser still supplied the proposition, creative material, budget and conversion definition.
Establish a baseline before tightening efficiency
The campaign began with Maximize Conversions to establish a cost-per-acquisition baseline, then moved to target CPA. That sequence matters. An efficiency target introduced before the system has enough reliable outcome data can restrict learning before there is much to learn from.
Treat the published result as a lead, not a universal benchmark
Google now provides Performance Max experiments for measuring incremental lift. The public MoneyMe case does not describe a randomized control, sample size, budget, confidence interval or conversion lag. A modern replication should define those questions before launch instead of treating a before-and-after result as complete causal proof.
Our reading of the evidence
What transfers—and what the public evidence cannot prove.
Google reports that the six-week campaign increased conversions by 22%, generated more than $800,000 in revenue from newly funded loans, and reduced overall account CPA by 20%. Those figures support a practical operating lesson: paid-media automation becomes more commercially useful when downstream revenue is returned to the bidding system and the creative supply is broad enough to test. They do not prove that Performance Max alone caused the lift, or that another lender would see the same result.