A creator posts a reel on a Tuesday evening. By Thursday you have eleven new orders, three of them with her discount code. The other eight could be hers, could be your email from Wednesday, or could be people who were going to buy anyway. Influencer tracking is the work of deciding which, order by order.
What is influencer tracking?
Tracking influencers means joining what a creator posted to the orders it earned, then saying how sure you are about each join. It covers three jobs: knowing which posts went out, collecting every signal that ties a buyer to a creator, and grading those signals so a report separates proof from guesswork.
Most setups stop at the first half. A spreadsheet of posts and a column of code redemptions shows activity and some sales. It cannot say what happened to the orders with no code, and those are often the majority.
The four signals that tie a sale to a creator
Every method of influencer tracking uses some mix of four signals. They differ in what they prove and in where they break.
| Signal | What it proves | Where it breaks |
|---|---|---|
| A redeemed discount code | The buyer had the creator's code at checkout | Codes leak to coupon sites; a shared code proves less than a single-use one |
| A tracked link | The buyer clicked the creator's link before buying | Lost when the buyer switches device or types the address |
| A captured post and its time | The creator posted shortly before the order | Timing alone cannot rule out coincidence |
| The buyer's own answer | The buyer says who sent them | People misremember, misspell handles or skip the question |
A single-use code is the strongest of the four, because it names one buyer and one source. A shared code is weaker: once it reaches a voucher site, redemptions stop meaning that the creator sent the buyer. Generate codes per creator, ideally from their handle so a code is readable at a glance, and keep a list of which code belongs to whom.
A tracked link carries the creator's name in its parameters, usually UTM tags such as utm_source and
utm_campaign, or a click id from a short link. It is strong when the click and the order happen on the
same device within the same session, and silent otherwise.
A captured post gives you the time and the content: when the reel went out, what it said, which product it showed. On its own it is circumstantial. Combined with a link or an answer it becomes useful context.
The buyer's own answer is the signal most stores never collect. A short question after checkout, such as "Who told you about us?", reaches the buyers who saw a post and then searched for the brand by name. Rewarding the answer with a small discount on the next order raises how many people reply.
How to track influencer sales without discount codes
You can track influencer sales without discount codes by combining tracked links, captured posts and a post-purchase question, then grading each order by the strongest signal it has. No single replacement is as strong as a single-use code, so the method depends on joining several weaker signals and being honest about the result.
- Give every creator a tracked link. Build each link with the creator's handle in its parameters, so the order carries their name when the buyer clicks through. Use a short link for bios and stories, where long URLs are cut off.
- Capture their public posts with timestamps. Read posts from every account the creator holds, on every network they use, so you know what went out and when. A post published two hours before an order is worth checking; one published five weeks before usually is not.
- Ask the buyer. Put one question after checkout and accept free text. Match the answer to your creator list, allowing for typos and near-spellings: "my friend Lila's video" should land on the creator with that name.
- Set a window per signal. Decide how long after a post or a click a sale can still count. Give hard signals a longer window than circumstantial ones.
- Grade every order. An order with a matching link click and a matching answer is close to proven. An order with only a post in the window is possible, not proven. Report the two separately.
The step stores skip most often is step three. Links fail across devices and posts fail on timing, but a buyer who types a creator's name into a question box has told you directly. It is first-party data that no pixel or analytics tag sees.
What this method cannot do is read private data. Story viewers, direct messages and the creator's own analytics screens are not public. A setup built on public posts and your own store data will not see them. Ad platforms are a separate source again: a creator's whitelisted ads need their own tracking through the ad platform.
Which influencer drove the sale when two creators fit?
When two creators fit the same order, split the credit by role rather than picking a winner. One creator introduced the buyer, often weeks earlier. Another may have assisted with a later post. One converted them, through the code or link used at checkout.
Take an order placed on a Friday. The buyer clicked creator A's link on Monday, saw creator B's post on Thursday, and redeemed creator B's single-use code at checkout. The answer to "who sent you?" names creator A. A fair reading is that A introduced the buyer and B converted them. Giving all the credit to B, because the code is the last touch, under-reports A for every order like it.
Some orders cannot be separated. Two posts went out on the same day, the buyer used no code, clicked no link and skipped the question. The honest answer is that the sale is possible for both creators and proven for neither. Reports that force a single name onto every order look tidier and are wrong more often.
A useful tracking setup therefore says how sure it is, not only who it credits. Four grades cover most cases:
- Proven: a single-use code, a click id or a matching link click names the creator.
- Possible: the signals point at the creator but do not settle it.
- Coincidence: a post was in the window, but nothing else ties it to the order.
- Unexplained: no signal reaches the order at all.
Counting the unexplained orders matters as much as counting the proven ones. A month with many unexplained orders tells you to add a post-purchase question or fix your links, not to buy more posts.
Setting up influencer tracking, step by step
A working setup has six parts. Each one feeds the grading above.
- One record per creator. Hold every account the creator uses, on every network, under one name. A creator with an Instagram account and a TikTok account is one partner, not two.
- Unique codes. Generate at least one code per creator, from their handle where you can. Use single-use codes for gifted or high-value offers, where proof matters most.
- Tracked links. One link per creator, with consistent UTM naming across the roster, so reports group correctly.
- Post capture. Collect posts as they go out, with time, caption, mentions and hashtags. Decide whether a creator's posts count straight away or wait for your approval; a new creator's posts are worth checking before they count.
- A post-purchase question. One question, free text, with a small reward for answering.
- Windows and review. Set how long each signal stays valid, and look at the possible and unexplained orders each week rather than only the proven ones.
Discounts deserve their own care. A code that never reached the store, or was typed with a different case, fails silently. Keep the list of codes in one place and check that each one exists in the store it was meant for.
Which influencer metrics matter for ROI?
The metrics that matter are the ones tied to orders: revenue each creator converted, orders they introduced or assisted, and what you paid them. Reach, likes and follower counts describe the audience. They do not say whether anyone bought.
- Converted revenue: revenue from orders the creator converted with hard evidence.
- Introduced and assisted orders: orders the creator started or supported, with their grade.
- Cost: fees, gifted product at cost, and commission, recorded in one currency.
- Return on cost: converted revenue divided by cost. Report it as a floor, with possible sales listed beside it.
- Revenue per post: what each post returned, so you can compare a reel with a video rather than one creator's follower count with another's.
A creator with a large following and few converted orders is a reach purchase. That can be a fair thing to buy. It is still a different purchase from a creator whose smaller audience redeems codes every week, and the report should make the difference obvious.
What an influencer report should show
An influencer report should show each creator's orders grouped by grade, not a single sales total. One row per creator, with converted revenue, introduced and assisted orders, cost and return, answers the budget question for next month. One row per post answers the creative question: which format and which message sold.
Keep the evidence one click away. When a row says a creator converted twelve orders, a reader should be able to open each one. The code, the link click or the answer behind it should be there. Reports without that trail get argued about; reports with it get used.
For the store side of the join, the discounts page covers how codes are placed and traced, and forms covers the post-purchase question in detail.
Questions people ask
Can you track influencer sales without a discount code?
- Yes. A tracked link, a post captured with its time, and the buyer's answer to a short question after checkout each tie an order to a creator without any code. None is as strong as a redeemed single-use code, so a good report says which signal it used and how sure it is, rather than counting every no-code sale as proven.
How do I know which influencer drove a sale when two creators posted?
- Split the job into roles. One creator introduced the buyer, another may have assisted, and one converted them, usually through the code or link used at checkout. Credit the converting creator only when there is hard evidence, and record the others as assisting. When nothing separates them, the honest answer is that the sale is possible for both.
What attribution window should I use for influencer posts?
- Match the window to how long your buyers take to decide. A low-cost product bought on impulse needs days; a considered purchase can need weeks. Keep circumstantial signals such as a post's timing on a shorter window than hard signals such as a single-use code or a click id, because a code names its source outright however late it is redeemed.
Do creators have to give me access to their accounts?
- No. Public posts, their captions and their public counts can be read without the creator logging in or installing anything. What you cannot see that way is private data: story viewers, direct messages and the creator's own analytics screens. Plan your tracking around public posts, your own codes and links, and what buyers tell you.
How do you work out influencer marketing ROI?
- Take the revenue from orders each creator converted with hard evidence, subtract what you paid them in fees, product and commission, and divide by that cost. Record costs in one currency so the sum is honest. Report possible sales beside the figure, not inside it, so the return is a floor you can defend.