Influencer Attribution Windows: How Long Should Creator Campaigns Get Credit?

Download the free attribution-window planner

An attribution window can make a good creator look bad. It can also make a weak campaign look much better than it was.

Close a YouTube test before day 90 and you may miss people who found the video late. Leave a code open for six months without a rule and the campaign may get credit for sales that came from email, paid search, or repeat use.

There is no perfect number. Set the clocks before the post goes live. Be clear about what each clock can tell you.

Which three clocks are teams mixing up?

Teams often use "attribution window" for three different things.

1. A platform reporting window is the rule an ad platform uses to count an action after a click, view, or engaged view.
2. An observation window is how long your team keeps gathering results from a creator post.
3. A decision window is the period that will guide a renewal, change, or stop decision.

They do not have to match.

TikTok makes this easy to see. Its current ad-group help page lists click windows of 1, 7, 14, or 28 days. It lists view windows of off, 1, or 7 days. For some app setups, it lists engaged-view windows of 1 or 7 days. A separate page for TikTok's self-attributing network lists a smaller set of options. In other words, the choices can depend on the campaign and setup. They can also change.

Those settings control paid TikTok reporting. They do not tell you when to stop watching an organic TikTok sponsorship. TikTok also says its Ads Manager setting does not change reporting from a mobile measurement partner.

Write all three clocks into the brief. If you do not, people will use the same word while arguing about different questions.

How should you use the 7/30/60/90-day table?

These dates are checkpoints inside one watch period. They are not four rival windows. They help the team plan. They are not platform rules.

Some sales take more than 90 days. Some evergreen posts keep getting views and clicks well past that point. Day 90 is the floor, not the deadline. A creator code still needs an end rule. Keep the first source on the lead in your customer relationship management (CRM) system. Report early creator actions and later sales steps apart.

How do you pick the window?

1. What exact action counts?

"Conversion" is not a useful answer on its own.

Is the goal a click, a sign-up, an approved account, a funded account, a purchase, a sales call, or paid revenue? Each step can take a different amount of time.

Track the path when you can. If the goal is funded accounts, ask for clicks, sign-ups, opened accounts, funded accounts, and a measure of user quality. This is not extra dashboard work for its own sake. It tells both sides where people leave.

Low clicks with strong down-funnel rates point to a very different fix than strong clicks with weak approvals. The first may need a better call to action or a more natural ad placement. The second may be a landing-page, offer, audience, or product problem.

2. How long do customers take to act without creator media?

Use the brand's own conversion-lag curve. A company may say its sales cycle is 30 days while most customers act in three days and a small group takes months. The full curve is more useful than one average.

Ask:

• How many customers act on the same day?
• How many have acted by day 7 and day 30?
• When does the curve start to flatten?
• Does lag change by product, device, or first purchase?

Many normal buyers may act after day 30. If so, day 30 is too soon for the final outcome. Fast offers should also stay open through at least day 90. That lets the team see the full content tail. Use fast sales to read the early checks. Do not use them to close a YouTube test early.

3. How old is the post, and how much attention has it had?

Calendar age is not enough. A post can be 30 days old and still have an immature read if it has reached far fewer people than that creator's work normally reaches by day 30.

For YouTube, compare the sponsor video's views with the creator's normal views at the same age. A new video may be only seven days old. Do not compare it with the final views of old work. The new video may be far below its same-age norm. If so, weak sales may point to low reach. They do not yet prove that the audience will not buy.

Agentio says it studied more than 10,000 YouTube creator integrations run through its platform. Nearly 40% of the views came after day 30. So did 30% of the clicks. A second Agentio report on short credit windows says brands should spread spend credit across at least 90 days. YouTube later cited the playbook and rounded the view finding to 40%. The Agentio sample does not cover every YouTube video. Its 90-day rule is not a promise for one post. Yet the work shows why day 30 is an early read, not a final one.

Classify the content before launch. News may get most views fast. The topic can lose value soon. An evergreen guide or review may gain views for months. Search and YouTube picks can keep it moving. Day 90 is still the first final decision point for both types. Yet the content type tells you if day 90 is final or just one more check.

Long-form posts can also get a second life from a news cycle or a creator's next video. Short-form reach is often faster, but old posts can resurface there too. Use the real curve, not a stereotype about the app.

4. What choice will this report drive?

A tracking check and a renewal decision need different clocks.

Do not wait 90 days to notice a broken link. Do not use a 7-, 30-, or 60-day read as the final verdict on a YouTube creator.

5. How strong a claim does the brand need?

A tracked sale tells you that the buyer used a creator path and acted inside the rule. It does not prove that the ad caused the sale.

The longer the window, the more likely it is that the buyer also saw a search ad, an email, another creator, or the brand's own content. That does not make the long view useless. It means the language should get narrower.

Use tracked results to run the work. Use a test when the question is whether the campaign caused more sales. Our YouTube lift study guide covers Brand Lift, Search Lift, and Conversion Lift. The creator media mix guide covers a second path for a large, ongoing program.

When is a read mature?

This is where teams get themselves into trouble. They agree to test a creator, then judge a slow outcome before the content or funnel has had time to work.

We look for five things before we treat a read as mature:

1. The post has been live for at least 90 days if it is on YouTube.
2. The post has had a fair chance to reach its normal audience.
3. The brand's usual conversion lag has had time to pass.
4. Tracking worked from the first click through the final event.
5. There is enough data to tell a weak signal from normal noise.

If one of those is missing, label the read early. Do not dress it up as a verdict.

What counts as a sign of life?

The final goal still matters most. But when it is not ready or it lands a little short, look at the path around it.

Useful signs may include:

• The post is still gaining views.
• Click rate is healthy, even if later steps are slow.
• A small group of users is moving far down the funnel.
• Customer quality is high.
• Comments show real interest or useful objections.
• The video underperformed the creator's normal views, but results per view were sound.
• One angle worked better than the others.

None of these turns a miss into a win. They tell you whether there is a reason to try again and what to change.

This is also why we prefer more than one touchpoint when the budget allows. One post gives you one view count, one angle, and one chance to land on a strong piece of content. Three posts with different angles give both sides far more to learn from. The window must be applied to each post by its own age, not only to the campaign's first launch date.

How should the window change by channel?

The product and action still come first. Use the channel to adjust the plan.

What about YouTube?

Keep a seven-day health check for links, clicks, and early audience response. Treat days 30 and 60 as directional checkpoints. Do not issue the final creator verdict before day 90, even for a fast-moving offer. At day 90, extend the read when the video's same-age view curve, click curve, or normal buying lag is still active.

Compare each video with the creator's same-age norm. Use the content type set before launch too. News and evergreen guides may mature at a different pace. Weak views plus weak sales can point to low reach. Normal views plus weak sales point to a different issue.

What about TikTok and Instagram?

Use day 7 as a health check because short-form response can move fast. Treat days 30 and 60 as directional, and keep watching through at least day 90 for most brand programs. Extend the window if the offer takes research, approval, or a large payment. Separate organic results from paid results when the creator post is also used as an ad.

Do not copy a paid platform's window into the organic report. The two windows answer different questions.

What about podcasts?

Podcast listeners may hear an ad away from a screen. They may search the brand, use a code later, or return to an older episode. Vanity URLs, offer codes, branded search, and post-purchase surveys can help fill the gaps.

First, ask what the delivery number means. IAB Tech Lab says podcast measurement still rests largely on server logs because most apps do not send client-side playback proof. Its current final guideline is version 2.2. Version 2.3 was released for public comment on July 21, 2026, with comments open through August 19. The draft adds more guidance on open RSS, video podcasts, server logs, and measurement windows. Do not call it final yet.

Ask the publisher which guideline and method it uses. Also ask about nearby ads. Podcast listeners have less public space than YouTube viewers to tell you that several ads were stacked too close together.

Is the campaign ready to judge?

Complete this before launch. Revisit it at each readout.

A thoughtful, imperfect window is better than a copied "best practice" that no one can explain.

What data should you keep?

Save the post date and each result date. Save the action, value, creator, asset, and platform. Keep the safest level of detail that your privacy rules allow.

Then the team can rebuild 7-, 30-, 60-, 90-, or later-day views. If the dashboard saves only one 30-day total, the missing tail cannot be recovered.

The creator campaign data standard covers the shared fields. The creator UTM guide covers link names and setup.

Which mistakes cause bad calls?

• Using one window for links, codes, surveys, CRM influence, and lift results.
• Changing the window after seeing which view looks best.
• Comparing creators when their posts are different ages.
• Comparing a new video with the lifetime views of old videos.
• Treating a paid platform default as the rule for an organic deal.
• Closing long-form work before it has had a fair chance to travel.
• Treating a day 7, 30, or 60 checkpoint as a final YouTube verdict.
• Treating day 90 as a forced cutoff when an evergreen view or conversion curve is still active.
• Keeping a code open forever and counting every reuse as a new customer.
• Calling attributed sales incremental sales.
• Hiding the funnel until the final event arrives.

The last mistake is costly. If a creator sends qualified people and the brand loses them at an application step, the answer may be a product or page fix. Without the middle steps, both sides are guessing.

Choose the window before the numbers come in

An attribution window is a rule for when you will look, what you will count, and what you will do next.

Set it before launch. Judge each post at the same age. Use days 7, 30, and 60 to operate the campaign, not to end the YouTube test. Make the first final decision at day 90 or later, and keep observing when the content or conversion curve is still active. If a campaign only works when it gets credit for every sale forever, the window is hiding the problem.

For help planning a creator test with clean readouts and fair renewal rules, talk with Creators Agency.