When Your Own Analytics and the Tracking Platform Disagree on an Influencer's Sales: A Reconciliation Process

The short answer

Your own order or analytics system and the tracking platform's own attribution dashboard rarely count the exact same number of sales for an influencer — that's the normal state, not a bug. The fix isn't chasing a perfect match: it's a standing reconciliation process — match orders by order ID, explain each gap with an already-documented cause, and pay commission on whichever source your contract names as the system of record.

Why the two numbers rarely agree

A tracking platform and your own order/analytics system are built to measure two different things with two different methods, so a gap between them is the normal state — not a symptom that something is broken. The most common, already-documented causes are:

• The in-app browser strips referrer and UTM signals on TikTok and Instagram, so a sale lands as an unknown source in your own analytics tool even though the tracking platform still caught the click via its own click ID. See the four concrete causes in why influencer sales don't show up in Google Analytics.
• The attribution windows aren't the same. Your own analytics tool and the tracking platform often set different default windows for how long a click still counts — a purchase that falls outside one window but inside the other only counts in one place. See how Meta, TikTok, GA4 and Safari's cookie cap each set their own window in attribution windows in influencer marketing.
• VPNs and private DNS break IP-based fallback matching. When the tracking platform loses the cookie, many setups fall back on IP address to still match a click to an order — but that doesn't work if the customer is on a VPN or a private DNS service. See do VPNs and private DNS tools break influencer tracking's IP fallback.
• Cashback and deal sites overwrite the last cookie. If the customer clicks through a cashback site after the influencer's link already set a cookie, that last cookie can win the attribution inside the tracking platform's own system — while your own order system still shows the influencer's code was used. See cashback and deal sites can overwrite an influencer tracking cookie.
• A discount code and a link count separately until they're deduplicated. A customer who both clicks the link and enters the code at checkout risks being counted twice across the two tracks, if they're never matched against each other. See the double-counting section in discount codes vs tracking links.
• Dark social isn't a tracking failure at all — it's a missing click. A screenshot or a link pasted into a private message was never a trackable click to begin with, so the tracking platform has nothing to attribute, even though the sale genuinely happened. See dark social.

None of these causes require inventing a new technical explanation — the point of this article isn't to re-explain them, but to give you a process for systematically matching each individual gap against one of them, instead of guessing.

The reconciliation process, step by step

1. Agree a system of record before the campaign goes live. Write into the contract which system governs commission calculation if the two disagree — usually the tracking platform, since it's the one generating the link or code and triggering the payout itself.
2. Export both datasets for the same period. Pull your own order/analytics export for the full campaign window, and pull the tracking platform's own list of attributed sales for the exact same date range.
3. Match by order ID, not by totals. A matching total can easily hide the fact that ten orders are missing on one side and ten different orders are missing on the other — only an order-by-order match shows the real gap.
4. Split the unmatched orders into two groups. Group A: orders in your own system that the tracking platform hasn't credited. Group B: orders the tracking platform has credited that don't exist in your own order system — the latter usually means a cancelled, refunded or test order the platform is still counting.
5. Explain each gap with one of the known causes above, order by order, rather than writing off the whole gap as "normal noise".
6. Pay on the agreed source — don't use the reconciliation to renegotiate the commission basis case by case, unless the gap is unusually large or unexplained.
7. Log the reconciliation with a date and both raw numbers, even when they match. A logged trail is what actually protects you if a gap later turns into a dispute with the influencer.

Which number should you actually pay commission on?

• IF the contract already names a system of record → pay on that number, and use the reconciliation purely to catch errors in the platform's own figure (e.g. a refunded order that's still credited).
• IF nothing was written down about which system governs → default to paying on the tracking platform's number, since that's the figure the influencer can see and dispute themselves — but investigate any gap above a set threshold (e.g. 10%) before the next payout.
• IF the gap is a one-off, unusually large spike in a single campaign → look for one concrete, explainable cause (a single cashback push, a VPN-heavy audience segment) before concluding a system is failing generally.
• IF the same percentage gap recurs campaign after campaign → that's a structural tracking loss (in-app browsers, attribution windows), not a reconciliation failure — budget for it rather than trying to close it to zero.

How often should you reconcile?

Reconcile before every commission payout run — not just once at the end of the campaign. Wait until the end, and you've typically already paid out several rounds on a number nobody checked, and any error is far more expensive and awkward to fix after the fact than to catch before the money is sent.

Worked example: reconciling one campaign

The numbers below are hypothetical and for illustration only. This is not a real Make Influence customer case, and none of the figures are benchmarks.

An influencer's discount code was used on 240 orders in your own order system over a 14-day campaign window. The tracking platform's own dashboard shows 216 attributed sales for the same influencer over the same period.

• Gap = 240 − 216 = 24 orders, i.e. 24 ÷ 240 = 10% of your own order count.

After matching order by order, you find:

• 9 orders: the code was used directly at checkout with no preceding click on the influencer's link — the tracking platform's click-based attribution has no click to match against, even though the customer genuinely used the code.
• 8 orders: explained by VPN or private-DNS use breaking the IP-fallback matching (per the VPN/IP-fallback article above).
• 4 orders: the customer clicked through a cashback site after the influencer's link, and the last cookie won the attribution inside the tracking platform's own system, even though the code still shows as used in your own order system.
• 3 orders: remain unexplained even after checking every known cause.

That gives: 9 + 8 + 4 = 21 of the 24 orders explained (21 ÷ 24 = 87.5%), with 3 orders (12.5%) remaining genuinely unexplained. In this example, the contract's system of record is the tracking platform's figure — so you pay commission on the 216 orders, log the 24 explained/unexplained orders as documentation, and let the 3 unexplained orders stand as a logged but non-actioned remainder unless the pattern recurs in the next campaign.

Comparison: your own system vs. the tracking platform's dashboard

Make Influence's operational perspective

In Make Influence's experience, most commission disputes don't happen because one of the two numbers is "wrong" — they happen because nobody wrote down which number governs before the gap showed up. Once that happens mid-payout, the conversation with the influencer gets unnecessarily hard, because it looks like a post-hoc justification rather than an agreed rule.

We recommend building reconciliation in as a standing, recurring step before every payout — not as an exception process only run when someone questions the numbers. A couple of hours of matching per campaign is cheap insurance compared with having to explain an undocumented gap after the money has already been sent.

FAQ

Which number is "right" — ours or the tracking platform's?

Neither is automatically more correct than the other — they measure different things in different ways. The point of reconciliation isn't to make them match perfectly; it's to document which number is agreed as governing, and explain the gap.

How big a gap is normal?

There's no single universal benchmark figure — it depends on the platform, the product, and which of the causes above apply to your specific setup. What should concern you isn't the size of the gap in one campaign, but whether the same percentage gap keeps recurring unexplained, campaign after campaign.

Do you need to reconcile for a flat per-post commission rather than per sale?

No — if the commission is a flat fee per post rather than a percentage of sales, there's no commission amount that depends on how many sales were actually counted. Reconciliation is still useful for reporting, but doesn't govern the payout itself.

Does reconciliation require a special tool?

Not necessarily. For most single campaigns, a spreadsheet matching order IDs from both exports is enough. If you're running reconciliation across a large roster of influencers or affiliates at once, doing it by hand can take long enough that a purpose-built tool becomes worth considering.

Does a large gap mean fraud?

Rarely. The large majority of gaps come down to one of the known technical causes above — not deliberate manipulation. Suspected outright fraud, such as fake clicks or cookie stuffing, is a separate problem with an entirely different investigation process.

What if the platform already paid out on a wrong number?

Catch it in the reconciliation before the next payout runs, and correct it going forward rather than trying to claw the money back — it's far cheaper in trust than a retroactive repayment.