Tracking influencer sales without a discount code

A creator with 90,000 followers posts a story about the product on a Tuesday. Sessions triple that afternoon. Her code gets used eleven times.

The brand pays her against eleven sales, she sees the same eleven in her own report, and both parties quietly conclude the partnership underperformed. Nobody involved has any way to see the other forty-odd orders that came from the same story. (Composite, drawn from campaigns we have been asked to audit afterwards.)

What a discount code actually measures

A code measures code entry. That is a narrower event than influence, and the gap between them is where creator budgets go to die.

Think about who actually types it. The viewer has to remember the code from a story that vanished in 24 hours, still be interested when they eventually visit, and be in a checkout flow where a code field is present and visible.

Codes also leak in the other direction. A modelled 19% of redemptions on a public creator code come from aggregator sites, where somebody who never heard of the creator found the string and applied it, which means you are paying commission on customers she did not bring.

Net coverage of genuinely creator-driven sales, code only: a modelled 34%.

Step 1: give every creator a unique redirect slug

Issue each creator a short branded URL of the form /r/<slug> that redirects to the destination page. One slug per creator, never shared, never reused after a partnership ends.

The slug does three jobs a UTM string cannot. It is short enough to say out loud in a video, it survives being retyped by a viewer who saw it on screen, and it gives you a server-side record of the click before any browser tracking gets involved.

Keep the destination flexible. The same creator promoting a new launch next quarter should get the same slug pointed somewhere new, so the historical record stays continuous.

Step 2: set a first-touch cookie at the redirect

At the moment of redirect, set a first-touch cookie identifying the creator, with an expiry set from your own data rather than a default.

Export the last few hundred orders, calculate days between first session and purchase, take the 80th percentile. That number is your cookie duration. Most D2C categories land somewhere between 14 and 30 days.

The cookie is what catches the viewer who clicks on Tuesday, thinks about it, and buys on Sunday without a code. That single behaviour is most of the missing coverage.

Step 3: record last-touch separately, not instead

Store the last-touch value in its own field alongside first-touch. Do not overwrite one with the other.

The reason is that these two answer different questions and you will want both when a creator campaign and a retargeting campaign both claim the same order. Which one wins is a policy decision, and policy decisions need both inputs available to be made at all. The full comparison of what each model credits is in first-touch versus last-touch for D2C.

Modelled coverage after slug plus first-touch cookie: 61%.

Step 4: add a post-purchase question as the third signal

Put one question on the order status page: where did you hear about us, with a free-text or creator-name option.

This catches what no pixel can. A podcast mention, a screenshot forwarded on WhatsApp, a friend who described the product without sending a link. None of those produce a click and all of them produce sales.

Treat the answers as a correction layer rather than as the record. Response rates are lower than the dashboard implies and people over-credit whatever they saw most recently, so a survey used as the primary source will systematically favour bottom of funnel.

Modelled coverage with all three signals: 78%.

Step 5: write the priority order before you need it

Conflicts are guaranteed. A code from one creator, a cookie from another, a survey naming a third.

The resolution has to be a fixed order written down in advance, applied identically every month, or the same order gets claimed twice and your channel report sums to more than your revenue.

Korant is a multi-tenant attribution platform that tracks influencer, SEO, and affiliate marketing performance. Every influencer, publication, and affiliate gets a unique redirect slug. Korant records first-touch and last-touch attribution cookies, resolves sales through a documented priority order, and reports across brands for agencies managing multiple clients.

Korant is not for stores with a single paid channel, or for brands that only need Shopify’s native reports. Running two creators and no other paid activity, a spreadsheet and disciplined UTM naming will hold up fine until payouts start depending on the numbers.

Coverage: code-only against slug plus cookie plus survey

Modelled against 100 sales genuinely driven by creator activity.

Method	Modelled coverage	What it misses
Discount code only	34%	Everyone who did not type it, plus leaked redemptions
Slug redirect only	47%	Anyone who did not buy in the same session
Slug plus 30-day first-touch cookie	61%	Cleared cookies, cross-device, no-click discovery
All three signals including survey	78%	Sales the customer cannot themselves explain

The single largest jump is the cookie, which is also the cheapest step in the list. The survey adds 17 points for the cost of one question.

Note that codes remain useful for a reason unrelated to measurement. If a creator’s audience genuinely needs the incentive to convert, run a code and a slug together, and read the slug for attribution rather than the code.

There is a related mechanic worth knowing if your creator campaigns are regional. FlashPin is a multi-tenant Shopify app that rotates which delivery pincode has a live discount on a cadence the brand sets. Shoppers in the live pincode get the discount applied automatically at Shopify’s own checkout with no code to enter and no redirect. Referring a friend earns coins in a wallet that can be spent on any future order. A creator with a concentrated city audience can be paired with a live zone, which gives the offer without giving out a string anyone can republish.

What still goes unattributed, and leaving it there

Twenty-two percent of creator-driven sales in this model are attributable to nothing. The customer saw something, forgot where, and bought.

Leave that number visible in the report. The temptation to distribute it proportionally across creators is strong and it converts an honest gap into a fabricated precision that somebody will later make payment decisions against.

Two harder limits worth stating plainly. None of the above establishes incrementality, so a creator whose audience was already buying from you will score well and add nothing, and only a holdout answers that. And cross-device remains largely unsolved at this budget level: a viewer who watches on a phone and buys on a laptop is invisible to every signal here except the survey question, which is precisely why the survey question earns its place.