A discrepancy is the gap between two systems that should be counting the same thing. Your tracker says 1,000 clicks, the network says 930, the traffic source says 1,080. Some difference is normal; a large or sudden one usually means something in the setup broke.
A discrepancy is a difference between the counts two measurement systems report for the same activity: impressions, clicks, conversions or revenue. Every link in an affiliate chain counts on its own, the traffic source, the affiliate’s tracker, the network and the advertiser, and each applies its own definitions, filters, time zone and timing. Perfect agreement almost never happens, so the useful question is whether a gap is stable and can be explained.
The normal causes are predictable. Clicks fall between systems because some visitors leave during redirects, bots are filtered with different rules and one platform counts unique clicks while another counts all of them. Conversions differ because of status (pending versus approved), attribution windows, cross-channel deduplication, time zones that move a conversion to another day and postbacks that arrive late. In display advertising the long-standing rule of thumb, rooted in IAB guidance, treats impression gaps of up to about 10% between a publisher and a third-party ad server as acceptable.
Alarming discrepancies have their own signatures. A sudden fall in tracker conversions while the network still shows them points to a broken postback or a lost click ID. A large click gap on a single placement suggests bots or accidental clicks that never load the page. Conversions that the tracker records and the network later rejects point to quality filters, fraud checks or duplicates, and a gap that widens week after week often means the advertiser is reversing conversions once the hold period ends.
Auditing works best starting from one click. Send a test click through the whole chain, note the IDs each system records and confirm the conversion arrives everywhere with the same value and time. Then compare daily totals by source in one time zone and one status, and keep a short log of setup changes so a new gap can be tied to whatever caused it. Persistent gaps that cannot be explained are worth raising with the network, backed by data at sub ID level rather than totals.
A media buyer compares a week of a 30 USD CPA offer across three systems, all set to the same time zone.
| Metric | Traffic source | Tracker | Network |
|---|---|---|---|
| Clicks | 52,000 | 48,900 | 47,600 |
| Conversions | – | 401 | 412 (371 approved, 41 pending) |
| Revenue | – | 12,030 USD | 11,130 USD approved + 1,230 USD pending |
The 8.5% click gap between source and network is in the normal range for paid traffic with several redirects. The 11 conversions the network shows but the tracker never received all arrived in the postback log with an empty click ID, every one from a prelander variant whose button had been hard-coded without the token: a setup fault, fixed in minutes, not shaving. The 41 pending conversions explain the revenue difference and will settle after the hold period. Numbers are illustrative.
How to diagnose affiliate tracking problems: which tracker, network and platform discrepancies are normal and which are alarming, plus broken postb...
Core · 10 min readHow to build an affiliate dashboard that drives decisions: the five to eight metrics that matter, segmenting by source, offer and GEO, leading vers...
Beginner · 6 min readHow affiliate tracking really works: clicks, click IDs, postbacks, server-to-server flows and why clean attribution is the foundation of a profitab...
Beginner · 12 min readIf you run campaigns on gut feel, you are flying blind with your own money.
A difference between the counts two systems report for the same clicks, conversions or revenue, caused by different definitions, filters, timing and tracking methods.
For clicks passing several redirects, gaps from a few percent to the low teens are common; for impressions the industry rule of thumb is under about 10%. Conversions should match closely once statuses are aligned.
Follow one test click through every system, compare by source and day in a single time zone and status, and check postback logs for missing or empty click IDs.
Rarely as a first explanation. Broken postbacks, lost click IDs, pending statuses and deduplication explain most gaps; persistent unexplained ones should be raised with sub ID data.
References are listed as plain text on purpose; look them up by title and publisher. Updated: 2026-10-09.
Run them inside a network with real tracking, roughly 48-hour payouts and a dedicated manager.
Join the network