← All work  ·  Case 04

Samwells

A 19-day Shopee CPAS remarketing campaign on a Rp1.8M test budget, about USD 111. Small money. The point is that the forecast held.

Role: Performance Marketing Associate · RevoU Period: Apr–May 2024 Channels: Meta Ads Manager · Shopee CPAS Budget: Rp1,800,000 (≈USD 111)
CREATIVE TESTINGSOCIAL COMMERCEShopee CPASproduct → saletest / learn / rotate
3.4%
Purchase value variance to forecast
143
Purchases, against 121 forecast
8.8
ROAS, against 8.5 forecast

What I did

Built the media plan before spending anything. Three custom audiences on view, add-to-cart and purchase windows, with exclusion windows so they couldn't overlap and bid against each other. Budget split evenly across all three so each one was a readable test rather than a guess.

Every input was forecast from the account's own historical CPM, CTR, CVR and AOV: 121 conversions, Rp15.24M purchase value, 8.5 ROAS. Then I ran it and checked my own numbers.

Forecast against actual

Purchase value
Forecast Rp15,237,342 · actual Rp15,751,431, within 3.4%
Conversions
Forecast 121 · actual 143, 18% over
Cost per acquisition
Forecast Rp14,852 · actual Rp13,139, 11.5% under
Average order value
Forecast Rp125,725 · actual Rp110,150, 12.4% under

Volume beat the forecast by 18% while basket size came in 12% under it. The two nearly cancelled, which is why revenue landed within 3.4% of plan. Knowing which of the two moved is what separates a lucky result from a repeatable one.

Meta Ads Manager campaign structure for Shopee CPAS
Campaign, ad set and ad structure built against Shopee Collaborative Performance Ads.
Static bundle offer creative for Samwells
Video creative cut for feed and story placements
Ad preview as delivered
Second ad preview as delivered

Copy led with the bundle price rather than the product. That's what the testing pointed at.