How to track creator post performance automatically (views, likes, CPM)

The cheapest way to track what creator posts do is to never ask for the numbers. When a creator submits a link to a post on X, YouTube, TikTok or Instagram, CollabPal reads its views and likes at submit time, re-reads them whenever you press Refresh, and shows them on the submission, on the campaign's insights, and on the creator's page. Cost per 1,000 views is worked out from what you actually paid. If the campaign pays on the numbers, the same reading decides the payout.

What gets measured, and when

Any submission whose link is a post we can read. That is independent of how the bounty pays: a flat-reward bounty for a review video still gets the video's numbers. The reading happens three times:

• At submit, so the review shows the post's numbers alongside the work. A slow platform doesn't hold the creator up; a miss is picked up overnight.
• Whenever you press Refresh numbers on the dashboard, a bounty or a creator page: every linked post not checked in the last hour is re-read, up to 25 per press, oldest first. Nothing runs on a schedule, because every read costs money and most days nobody is looking.
• On demand, from Refresh on the submission, with a ten-minute cooldown so nobody can burn money by leaning on the button.

Which platforms return numbers depends on which providers the workspace's server has configured. The bounty builder says so when you pick a platform, and a social-metrics bounty on a platform without numbers can be saved but not paid until they are on.

What the campaign insights tell you

Open any bounty and the top of the page is the campaign's performance, computed from its measurable, non-rejected submissions:

• Total views and likes across the linked posts, and how many posts reported numbers.
• Median views per post. The median, not the mean, so one viral post doesn't describe a campaign of ordinary ones.
• Engagement rate: likes per view across posts that reported both.
• Cost per 1,000 views: everything approved on this campaign divided by its views. This is the number to compare campaigns, platforms and creators on.
• Engagement over time: each post's current numbers, counted in the week it was submitted.

The same figures roll up per creator, so a creator's page shows their median post, their engagement rate and what their reach cost you. That is the input to moving them between groups and to setting their rate.

How do you know the post is the creator's?

The provider reports who posted it, and that is checked against the account the creator connected. A post by someone else is refused at submit on every bounty. A post that cannot be tied to a connected account still counts in the insights, but it cannot decide a payout: a social-metrics bounty will not approve it, and a milestone bonus is held, until the creator connects that platform. Numbers that release money have to be the creator's own numbers.

Paying on the numbers

Three ways the numbers can drive pay, all of them set in the bounty's one-sentence builder:

• Social-metrics bounty: a post has to reach a threshold before it can be approved, and pays a flat amount, an amount set at approval, or a CPM rate per 1,000 views or likes with a cap per post.
• Step bonus: on a threshold bounty, so much per extra 1,000 views above the threshold, up to a cap.
• Milestone bonus: on any bounty, a fixed amount if the post reaches a number. A $150 walkthrough with a $100 bonus at 50,000 views is a flat deal for most creators and a performance deal for the one who breaks out.

At approval the reviewer sees the verdict in plain words: "$100 bonus: the post reached 62,400 views" or "No bonus yet: 31,000 of 50,000 views", and the Approve button shows the total.

What to do with the numbers

• Compare cost per 1,000 views across campaigns before deciding what to run next month. A $50 review bounty at $0.40 per thousand beats a $300 walkthrough at $2.10 unless the walkthrough converts better.
• Promote creators whose median post is above the campaign's median into the Trusted group; give them the bigger bounties.
• Set a creator's rate from their numbers. The rate pre-fills custom-reward approvals, so you stop renegotiating.
• Export the submissions table to CSV when finance asks; every row carries its views, likes and what it was paid.

Do it with an agent

claude mcp add --transport http collabpal https://www.collabpal.com/api/mcp

For each open campaign, give me total views, median views per post and cost per 1,000 views, then rank creators by median views. Refresh the numbers on anything checked more than a day ago first.

The agent uses list_campaigns and get_campaign (which carries the insights), list_submissions and get_submission for per-post numbers, refresh_post_metrics to re-read a post, and get_creator for a creator's roll-up. A read-only connection is enough for all of it.

Frequently asked questions

Do creators have to connect their accounts for tracking?

No, tracking works from the link alone. Connecting is what lets the numbers decide a payout, because it proves the post is theirs.

How fresh are the numbers?

Read at submit and then whenever you press Refresh on a page or a submission. Each figure shows when it was last checked.

Can I track posts from campaigns that pay a flat amount?

Yes. Any submission that links a readable post is measured, whatever the reward. The insights and the cost per 1,000 views work for flat campaigns too.

What is a good cost per 1,000 views for creator content?

Depends on platform and intent, but for startup product content $1 to $5 per 1,000 views on short video is common, and under $1 means a creator found an audience you should pay more attention to.

Does this replace platform analytics?

No, it reads public post numbers so you can compare creators and campaigns and pay on them. Deep analytics stay in each platform; this is the layer that ties them to what you paid.