Postback tracking that survives an audit

Client-side pixels die on pop traffic: redirect chains strip them, aggressive browsers block them, and bots fake them without effort. Server-to-server postbacks are the only conversion record worth optimizing against, which is why no campaign on this desk spends a dollar before the chain below is wired and test-fired. The tracker can be Voluum ($149 to $499 a month by plan), RedTrack (from $99), or Binom if you already run it; the procedure is identical.

The chain, end to end

One conversion travels five hops: the pop network serves a click with its click ID, the tracker catches it and issues its own click ID, the prelander carries it to the offer link, the affiliate network or advertiser records the conversion against that ID, and the postback URL fires back to the tracker, which optionally forwards a postback to the traffic source for its optimizer. Every failure we see in audits is one of these hops silently dropped, usually the last one, because nobody notices missing data that was never there.

Sub-ID taxonomy: decide it once

The tracker's sub-ID slots are the only place a conversion remembers where it came from, so we standardize them on every account: sub1 = network + zone ID, sub2 = prelander variant, sub3 = offer/payout version, sub4 = GEO-OS pair, sub5 = campaign revision date. Five slots, same meaning everywhere. The kill rules read sub1, the ladder verdicts read sub2, and neither works if the taxonomy drifts per campaign.

Three test fires before launch

1. Manual conversion: click your own campaign link from a real device, convert on the offer's test mode, and watch the conversion land in the tracker with every sub-ID populated.
2. Postback replay: fire the affiliate network's postback URL by hand with a known click ID and verify the tracker attributes it to the right campaign, zone, and payout amount, to the cent.
3. Source forward: confirm the traffic network received its copy, because a network optimizer fed no conversions will happily optimize your money toward whatever converts nothing.

The three fires take under an hour. Skipping them is how accounts run for six weeks on conversions attributed to the wrong campaign, which we have seen, twice, on accounts arriving for the $1,200 audit.

Reconcile weekly or drift silently

Every Friday the tracker's conversion count gets compared against the affiliate network's dashboard for the same window. Anything past a 2% gap is investigated the same day: usual suspects are timezone mismatches between platforms, a payout change nobody forwarded, or a postback endpoint that started rejecting long user agents. The gap itself is data. An account whose tracker and network agree within 2% for eight straight weeks has tracking you can bet bids on, and that is the entire point.