Checklist 04 · by Alina Bondar · updated 2026-08-14
Whitelists earn their place
A whitelist is not a list of sources you like. It is a list of source IDs that survived written thresholds on your own offer, your own GEOs, your own clicks, reviewed on a fixed day every week. Eight checks below. They assume working per-source attribution, so if the tracker taxonomy is improvised, start with the postback checklist instead; a whitelist built on broken attribution is a horoscope with a budget.
Write the kill threshold before spend
Our default: a source ID dies after 400 clicks without a conversion, or after spending 2× the target cost per conversion, whichever lands first. The exact numbers matter less than their timestamp: thresholds written after the data arrives always forgive the data.
Start broad, on purpose
Open targeting for the first 72 hours is data collection, not a mistake, provided the daily cap holds and the kill rules exist. A whitelist copied from a forum thread inherits someone else's offer, GEO, and season; yours has to be earned on your own clicks.
Judge sources on cost per conversion, not CTR
High-CTR sources with no conversions are how bot zones introduce themselves. The tracker's per-source report (sub1 in the postback taxonomy) is the only vote that counts, which is why the postback chain gets tested before any of this starts.
Give borderline sources one structured second chance
A source that failed on 400 clicks during a weekend, on one creative, deserves one retrial on a weekday with the winning creative, at half budget. One. Sources that fail twice are done; sentiment is not a KPI, as the sweeps account learned about Bangladesh.
Cap winners at 20% of daily budget
The best source in week two has a habit of being the most fraud-attractive source by week five, because publishers notice demand. A concentration cap keeps one zone's decay from taking the account down with it.
Rebuild bids per source, not per campaign
Once a whitelist exists, flat campaign bids leak money: pay each surviving source what its conversion economics support. On our sweepstakes rebuild that meant fourteen GEOs collapsing to whitelists holding CPC under $0.03 while the blended ROAS climbed from 0.7 to 2.4.
Keep a graveyard file
Every killed source, with the date and the number that killed it. Networks recycle zone IDs and publishers rename; the graveyard stops you from re-buying a lesson you already paid for. Ours is a plain spreadsheet with 1,100 rows and counting.
Review weekly, on a fixed day
Whitelists rot. Publisher mixes shift, Chrome's 2026 engagement enforcement keeps re-shaping subscriber quality, and a list that was clean in March is a rumor by July. Friday, thirty minutes, same three questions: who died, who is decaying, who earned more budget.
The shape of a healthy list
Six to ten weeks in, expect 5 to 15% of tested sources to hold a place, concentration under the 20% cap, and a graveyard several times larger than the whitelist. The sweepstakes receipt shows the full arc on a real account, and the GEO checklist explains why the same process lands differently in Manila than in Minneapolis. Done by hand it is a few hours a week; done by us it is the media-buying retainer.