Checklist 03 · by Deniz Aydın · updated 2026-08-14
Postbacks, tested
Most traffic problems are sub-ID taxonomy problems wearing a costume. The S2S chain has four hops: network to campaign URL, URL through prelander, prelander to offer, offer back to tracker, and every hop can drop the click ID without an error message anywhere. Nine checks below, in wiring order. The rule that governs them all: no spend until a real test conversion lands in the right column with the right sub-IDs attached, because a broken pipe and a bad campaign produce identical charts and only one of them deserves a budget cut.
Pick the tracker and pay for it
Voluum's Profit plan runs $149 a month for a million events; RedTrack starts at $99 with Solo at $149. Binom sells a license if you would rather own the server. Any of them works; a spreadsheet does not. The tracker is the second-source that makes every other number on this site checkable.
Map the sub-ID taxonomy before the first campaign exists
Fix the slots: sub1 source ID, sub2 creative, sub3 GEO, sub4 format (classic or in-page), sub5 prelander. Write it down and never improvise a slot. Week-six reports on improvised taxonomies require archaeology, and archaeology bills by the hour.
Carry the network click ID end to end
The network's click-ID macro rides in the campaign URL, through the prelander, into the offer link, and back in the postback. One dropped parameter anywhere and conversions arrive unattributed, which looks exactly like a bad campaign instead of a broken pipe.
Use S2S, not pixels
Push traffic converts on offer pages you do not control, behind redirects, on devices where third-party pixels die quietly. Server-to-server postbacks survive all of that. If an offer only supports a pixel, price that risk into the payout you accept.
Fire a real test conversion from every network
Not a preview click: a real click from RichAds, Push.House, or PropellerAds test tools, followed through to a test conversion the offer side confirms. It must land in the right tracker column with the right sub-IDs. One test per network, repeated after any URL change.
Verify the payout value in the postback
Our sweepstakes account lost three weeks of margin to a payout mismatch the postbacks caught: the offer reported $1.80 while the network settled lower on a subset of GEOs. Pass the payout parameter and reconcile it weekly against the affiliate network's own report.
Set duplicate and fraud rules in the tracker
Same click ID converting twice, conversions with zero-second click-to-convert times, and GEO mismatches between click and conversion all get flagged automatically. Cheap traffic includes bots; the tracker is where you refuse to pay them twice.
Send network postbacks too, when bidding needs them
PropellerAds CPA-goal bidding and RichAds automated rules optimize on the conversions you send back to the network. Forward a filtered postback stream: real conversions only, no test events, or the network optimizes toward your QA clicks.
Reconcile weekly against a second source
Tracker versus affiliate-network dashboard, every Friday, same UTC day boundaries. Single-digit percentage gaps are timezone noise; anything past 10% is a broken pipe or a payout dispute waiting to be had while it is still small.
When the wiring passes
A passing chain unlocks everything else on this site: the whitelist process assumes per-source data it can trust, and the fatigue schedules in the creative brief assume per-creative CTR that is actually per-creative. If you would rather hand the whole wiring job over, it is a fixed €700 setup and it ends when the test conversions pass.