Discrepancy: The Gap Between What Your Traffic Source Says and What You Actually Get

Every media buyer has had this moment. The traffic source dashboard says 10,000 clicks. Your tracker says 8,500. Nobody stole anything, nobody lied, and yet the numbers don't match. That gap is called discrepancy, and if you've been buying traffic for more than a week, you already know it's not going away.

What Discrepancy Is and Why It's a Problem

Discrepancy is the difference between the same metric reported by two systems that are supposed to be counting the same thing: clicks, impressions, or conversions. One system says A, the other says B, and the percentage between them is the discrepancy rate.

It sounds like a technical footnote until you put a price on it. If a traffic source charges you for 100,000 clicks at $0.050 CPC but your tracker only recorded 85,000, you paid for 15,000 clicks you never got a chance to convert. That's $750 gone before you even look at conversion rate or ROI. Scale that across a month of spend and a discrepancy problem stops being a rounding error and starts being the reason a campaign that should be profitable isn't.

The frustrating part is where discrepancy comes from. It's rarely one clean cause. Bot and crawler traffic gets counted by the source but filtered by your tracker, or the other way around. Redirect chains lose data between the click and the landing page, especially when a link passes through a CPA network, then a tracker, then an offer, then finally the advertiser. Postbacks time out, get sent with the wrong parameters, or never fire at all. Ad blockers stop client-side pixels from ever loading. And sometimes, less often than people assume but often enough to matter, a network simply doesn't credit conversions it should.

It means you're looking at a metric made of several different failure points stacked on top of each other, and the only way to deal with it is to be able to tell them apart.

How to Deal With It, and What Counts as Normal

The first thing worth knowing is that zero discrepancy isn't achievable. Every serious source on ad tracking agrees on this. The IAB treats anything under 10% as within normal range for standard ad serving. For affiliate chains specifically, where a click passes through more redirects before it lands, up to 20% is commonly described as acceptable, simply because every extra hop is another place to lose a packet. Mobile attribution sits higher, somewhere in the 10 to 15% range, because of delayed and aggregated postbacks.

So if the goal of hitting 0% isn't realistic, what do we have to do then? It's to know your baseline, catch when it moves outside of it, and understand why.

The industry-standard fix for the biggest chunk of discrepancy is server-to-server tracking. Instead of relying on a browser to fire a pixel, the conversion gets reported directly from server to server, which means it doesn't care whether the user has an ad blocker, cleared cookies, or closed the tab before a script finished loading. S2S alone typically cuts variance down to a few percent when it's implemented correctly, meaning the click ID is consistent end to end, the postback fires from the backend rather than client-side JavaScript, and the same identifier survives every redirect in the chain.

Beyond that, the standard playbook is: keep redirect chains as short as possible, validate every new integration before you trust its numbers, log every postback attempt so failures are visible, and reconcile your own numbers against the source's official report on a regular schedule to keep your spending and ROI data accurate.

There's also a discipline piece that gets skipped sometimes. Before you scale a new traffic source, check the discrepancy rate at a zone level. Some zones can show different percentages, which overall affects the whole source's efficiency when scaled. In reality, it could be a few zones and they will decide for the whole potentially good source.

How Tracking Platforms Specifically Help

A good tracking platform separates discrepancy into layers to avoid judging by one blended percentage. It flags bot and crawler traffic on the click itself, so you can compare organic-only numbers to the source's report and see if bots explain the gap. It can give you robust analytics for a deep investigation to see which slice of traffic has the biggest discrepancy. It keeps a full log of every postback sent and received, so a missing conversion is a traceable event instead of a mystery.

Just as importantly, a tracker gives you an independent number to compare the network against. Without your own click log, you're taking the source's word for both sides of the equation. With one, you can pull an official report from the traffic source, run it against your own data side by side, and see exactly where the two diverge. That's the same mechanism that protects you against a network quietly under-crediting conversions, because a pattern of persistent, one-directional gaps between your tracker and their dashboard is the clearest signal that something on their side, not yours, needs a conversation.

What's Available in Rinku for Buyers Working This Problem

Rinku was built around the idea that discrepancy shouldn't be something you notice a month too late. Here's what's available today.

Every click gets enriched at tracking time with browser, device, geo, and a crawler flag, so filtering out bot traffic from your analysis is a one-click move. And you can also see bot percentages within breakdown entities, for example, zones. If discrepancy disappears once you exclude flagged bot clicks, you've found your cause. And a cool part is that you can set a set of rules for redirecting this traffic to a fall-back page or somewhere else where bots don't matter

Rinku's reconciliation tool lets you import an official report from your traffic source and match it against your own click log automatically. The result is a cost re-calculated per the smallest unit within the system, the zone, so you see the actual ROI for the zone within the traffic source. A normal, every-morning upload will keep your spend in order, considering some clicks were lost due to discrepancy. You can also see in the dashboard which zones never deliver any clicks to you but are counted on the source's side.

On the postback side, every request sent to a traffic source is logged, with timing and status, so a drop in credited conversions is something you can actually diagnose: was the postback never sent, did it time out, or did the network reject it. That turns "conversions aren't matching" from a support ticket into a specific, fixable problem.

The analytics layer underneath all of this lets you cut the same discrepancy question by any dimension that matters: zone, creative, campaign, geo, device. That matters because a campaign-wide discrepancy number of 8% can easily hide one zone running at 40% and ten others running clean. Without segmentation, you'd never know which one to fix, or whether it is worth fixing at all.

And because discrepancy costs money differently depending on how you're being charged, whether that's CPC, CPM, or CPA, Rinku ties this back to profit and ROI automatically to help you avoid doing math by hand. You can set a rule that flags or auto-blocks a specific zone the moment its performance crosses a threshold you define, scoped to just that traffic source or campaign, so a bad zone gets caught and cut before it drags down numbers you're actually watching.

To learn more about Rinku's capabilities, visit our documentation page.

Conclusion

None of the above makes discrepancy go away. There are tools, methods, and sets of best practices available through the web that can turn a single frustrating percentage into a set of specific, checkable causes, most of which you can actually do something about.

Start tracking your own discrepancy and know your baseline today. Try Rinku free for 14 days at rinku.cc.