A checklist for source-based research
A competitor audit records a defined sample of accessible public posts and separates observations from explanations. Its output should be a sourced finding or a testable hypothesis, not a claim that a particular hook, tag or campaign caused a result.
- Define the question. Explain why each account belongs in the comparison.
- Record coverage. Keep the number of posts and start/end dates separately for every account. Equal post counts do not guarantee equal periods.
- Save source links. Attach a source URL to each cited observation.
- Check fields. Keep missing values missing and verify formats rather than assuming every TikTok item has the same shape or duration.
- Describe the distribution. Report the median, range and eligible sample count where useful, and inspect outliers separately.
- Define ratios. State whether you divide likes or comments by views, followers or another denominator. Exclude unavailable or zero denominators.
- Inspect dates and captions. A cluster may suggest a theme; it does not prove a funded campaign.
- Count distinct posts per hashtag. A repeated tag does not establish its effect on discovery or engagement.
- Compare in context. Neighboring posts are contextual comparisons, not controlled experiments.
- Write one test. Specify an original change, outcome and measurement period for your own account.
Illustrative audit note
Suppose a top post has 20 times the recorded views of the sample median and appears among several posts using the same tag. That describes an outlier and a tag cluster. Funding, distribution and the reason for the difference remain unknown. A possible next step is to test an original creative variation and measure it on your own account. These numbers are hypothetical.
Choosing the sample
A modest initial window can help inspect the data, but no universal post count guarantees a reliable conclusion. Report coverage and uncertainty, expand the sample when the question requires it, and avoid treating a small window as full account history.
Keep the limits visible
Record the source URL, publication date and timezone, observation time, returned post count and date coverage. A public account may return incomplete data or no usable result. Missing fields remain unavailable. Views are not unique viewers, verified reach, watch time or conversions. A public post does not establish why a result occurred, and a mention does not prove a paid partnership.
Availability
V5 launches with Instagram research. This is an educational guide to public TikTok research, not an offer of new TikTok collection, a free paid report or an MCP integration. Existing customers retain their agreed terms. Contact support to confirm existing access; use current pricing for new V5 offers.
Common questions
What is a good engagement rate?
This guide supplies no universal benchmark. Use a clearly defined measure and a relevant comparison set, with dates and coverage visible.
Should I copy a competitor’s top video?
No. Use it as a source to inspect, then develop an original test. A public result does not prove the mechanism behind it.