Tracking fails silently. A campaign with broken tracking looks exactly like a campaign with a bad offer: spend goes out, nothing comes back, and the conclusion is usually wrong.

These are the errors that produce that outcome, roughly in order of how often they happen.

1. The postback was never tested

The postback is the message the offer sends your tracker when a conversion happens. If it is misconfigured, your tracker records clicks and no results.

Catch it: fire a test conversion before any real spend. Most networks provide a test URL. Four days of buying against a postback that never fired is the most common expensive mistake in this business.

2. Wrong tokens for the traffic source

Every source uses its own parameter names for placement, creative, campaign and device. Copying a tracker template written for a different source produces empty columns in every report.

Catch it: send a few clicks and open the report. If placement or creative is blank, the tokens are wrong, not the traffic.

3. Losing parameters on redirect

A landing page that strips query parameters, or a redirect chain that drops them, breaks attribution between the click and the conversion.

Catch it: click through your own funnel and check that the click ID survives to the offer page. Do this every time you change the landing page, not only at setup.

4. Counting the same conversion twice

Two tracking systems both reporting, or a pixel that fires on page reload, inflates results. This one is worse than under counting because it makes a losing campaign look profitable and you scale it.

Catch it: compare your tracker’s numbers against the network’s payout report weekly. They should be close. Persistent disagreement in your favor is a warning, not good news.

5. Ignoring approval rate

Your tracker records conversions. The network approves a share of them. Optimizing on recorded conversions while a segment is being rejected means you are scaling exactly the traffic that gets refused.

Catch it: pull the approval rate by placement and country, not just overall.

6. No consistent naming

Campaign names improvised over months make comparison impossible six months later, when the comparison is finally worth doing.

Catch it: decide a naming pattern before the second campaign and never deviate.

Where setup errors originate

Most of them come from configuring the tracker and the traffic source separately and assuming they agree. Mondiad’s walkthrough of a full campaign setup covers where the parameters go on the source side, which is where the mismatch usually starts.

The check that catches most of it

Before every campaign: one test click through the full path, and one test conversion. Two minutes. It catches items one, two and three, which together account for most tracking failures.

After every campaign week: tracker numbers against network payout numbers. That catches four and five.

The reason this matters so much before scaling is set out in cheap traffic and profitable traffic are not the same thing.

And on reading the data once it is arriving correctly, see SEO traffic vs social traffic.

FAQ

How do you know if affiliate tracking is broken? Clicks recorded with no conversions across every creative and segment is the signature. A genuinely bad offer usually produces at least a few.

What is a postback? The server to server message the offer sends your tracker when a conversion occurs. Without it your tracker sees clicks but not results.

Why do tracker and network numbers differ? Small differences are normal, from timing and deduplication. Persistent large gaps mean a configuration problem or double counting.

Should you optimize on recorded or approved conversions? Approved. Recorded conversions that get rejected later will lead you to scale the traffic the advertiser is refusing to pay for.