A network dashboard reports one conversion. Your click tracker reports zero. Before changing links, pixels, or attribution settings, determine whether the two screens are even measuring the same event. An affiliate network may count a pending, approved, reversed, or delayed transaction; a click tracker may record a click but not receive a conversion notification at all. Neither screen alone establishes a paid commission.
First win: reconcile one event without changing the campaign
Open a blank ledger and record the network's event time, timezone, order/reference identifier (redact personal data), product, gross amount, commission status, and any network click/sub-ID. Next record the tracker's reporting timezone, date range, campaign, click ID or sub-ID, and conversion event definition. Save screenshots or exported rows with capture time. Do not paste buyer names, emails, card data, or full order records into a third-party support ticket.
The first win is a single row that says either matched, not comparable, or unresolved. A dashboard disagreement is not proof that either vendor is broken.
| Check | What to compare | What a mismatch suggests | Safe next move |
|---|---|---|---|
| Event definition | Network sale/lead vs tracker conversion goal | Different event types | Label separately; do not force a one-to-one count |
| Time | Both timezone settings and inclusive date windows | Midnight rollover or reporting lag | Expand the window around the event, noting the adjustment |
| Status | Pending, approved, rejected, reversed, paid | Network counted an event that has not earned commission | Track each state, not just the initial sale |
| Identity | Network click/sub-ID vs tracker outbound click identifier | The two systems cannot join this transaction | Inspect the original recorded outbound link and available network fields |
| Redirect | Destination after every redirect and parameter survival | Click/sub-ID stripped or rewritten | Test with a harmless non-purchase click only where program rules allow |
| Notification | Network postback/API/pixel configured, enabled, and acknowledged | Tracker never received the event | Inspect delivery logs and vendor instructions before editing |
| Filtering | Unique vs total clicks, bot filters, deduplication | Count definitions differ | Compare raw eligible rows rather than headline totals |
| Scope | Correct account, property, offer, and campaign | Screens describe different traffic | Correct the report filters, not the live links |
Diagnose in the least destructive order
1. Freeze the evidence. Capture both reports and the original link configuration before any edits. If the network offers a transaction export, use it as the network-side event record. Do not assume the commission is approved or payable merely because a conversion row exists.
2. Align the comparison. Match reporting periods, timezones, campaign scope, and the event being counted. A tracker goal named “sale” may be a manually configured event; its name does not prove it received a sale confirmation. A network conversion may be recorded after the click and appear in a later reporting period.
3. Look for a join key. A click/sub-ID passed into the network and returned with the conversion is the strongest ordinary way to reconcile one event. A landing-page UTM parameter by itself is not a guaranteed transaction identifier. If the network did not retain the identifier, classify the event as network-observed, tracker-unattributed, not as a fabricated match.
4. Inspect the notification path, if one exists. Document the expected sender, receiver, event type, authentication method, and delivery acknowledgment. A conversion pixel placed on a checkout page you do not control is not an available remedy. Some programs provide postbacks or reporting exports; others do not. Check the specific program's rules before assuming an integration exists.
5. Test without buying. Where permitted, perform one ordinary outbound click and verify that the click is recorded and the destination loads with only allowed parameters. This tests the click path, not the conversion path. Never create a fake transaction or use a self-referral to manufacture proof. A real conversion test requires separately authorized, compliant conditions.
6. Escalate with a narrow evidence bundle. Send support the anonymized reference, event timestamp/timezone, expected goal, observed identifier behavior, and configuration screenshots. Ask whether the network emits a conversion notification for that status and whether it includes the tracking identifier. Do not send secrets or buyer PII.
Example: a truthful reconciliation
Suppose the network displays one pending order at 23:55 UTC, while your tracker shows no conversions for the corresponding local calendar date. Your ledger first tests timezone and status. If the network record has no retained click ID, even an expanded tracker date range cannot prove which click drove the order. Report: “One pending network event; zero tracker-confirmed attributed conversions; identity unresolved.” That statement is useful, and it does not inflate earnings.
When to stop troubleshooting
Stop changing settings when you have established that the program cannot return a click-level identifier or conversion event to your tracker. Keep network-side sales and tracker-side click metrics as separate evidence streams, and judge the campaign by the reliable measures you actually possess. Do not buy tracking software merely to solve a data-sharing limitation that software cannot remove.
For a small campaign, a spreadsheet with event timestamp, source, offer, outbound click count, network transaction count, approval status, and payout confirmation may be sufficient. The first decision is whether you need click-level attribution at all. A more complex tracker becomes useful when it can receive lawful, reliable event evidence and that evidence changes your decisions.
What next
After you classify one discrepancy, document a repeatable weekly reconciliation: capture raw rows, compare statuses, note reversals, record only approved commissions as approved, and record cash only when actually received. If you cannot identify an event reliably, mark it UNKNOWN. That is more actionable than assigning it to the last campaign you remember launching.
Disclosure: Demandward may recommend affiliate tools elsewhere. This diagnostic guide does not require purchasing anything and contains no affiliate CTA.
Next useful check: If the click identifier is missing before the network receives it, follow the safe redirect-chain audit on one permitted, non-purchase click. You can return to Demandward home without buying a tracker.
For more practical guidance, visit Demandward. This article does not require purchasing software.