The brief
Ginyaki is a restaurant group in Islamabad that was launching two ordering apps, Ginyaki and GinXpress, alongside its website ordering. The Meta side of that launch is in the Ginyaki case study. On Google the job was installs for both apps and online orders from people already searching for the brand or for food nearby.
What I did
App campaigns for each app. Ginyaki and GinXpress each got their own app install campaign. The Ginyaki app campaign ran on Target CPA and came in at Rs22 an install.
Search for people ready to order. A search campaign on Target CPA picked up people searching to order. It had an 18.85% click-through rate and a 35.37% conversion rate, and GA4 recorded 128 purchases from it worth Rs274,117.
Performance Max, moved onto order value. The first Performance Max campaign was set to maximise conversions and returned 1.25 in conversion value. The commercial Performance Max campaign bid on Target ROAS instead, optimising on GA4 purchases with their real order values, and returned 4.27.
Results
Across the account, 2,692 app installs (1,284 for Ginyaki at Rs22 and 1,408 for GinXpress at Rs35.20) and, on the commercial Performance Max campaign alone, 292 GA4-tracked orders worth Rs597,030 and a 4.27 ROAS on its total conversion value. A second search campaign in November came in at Rs6,916 an order, well above the rest, and was switched off so the budget stayed on the campaigns above.
The account has several purchase actions that overlap, so where a campaign's conversions came from more than one of them the figures above use GA4 purchases, and no order is counted twice.
In short
Installs bought on cost per install, orders bought on order value. Once Performance Max was told what an order was worth, it returned more than three times what the first version did.