Chronometrica vs Clockworks User Search for Content-Backed TikTok Creator Discovery

Compare the record you need to review

Chronometrica searches public TikTok videos. Its video records include author identity, caption, video URL, available public post metrics and query context. One author can appear in several records. This makes the video the starting point for finding people whose content deserves study. Chronometrica product details.

Clockworks supports niche-keyword discovery. Its records contain profile facts, including names, signatures, follower and video counts, bio links and user IDs. Clockworks product details.

The table follows the checked Chronometrica contract, Chronometrica input schema, Clockworks contract and Clockworks input schema.

A useful selection rule follows from these record types. Write down what the reviewer must see before accepting a candidate. If that includes a concrete example of the person's content, start with video records. If it consists of profile facts, start with profile records.

Define what “worth studying” means for this assignment

The workflow question behind this comparison came from a product-builder research post. It asked how people research creators, hooks, formats and trends. That question helps define the task. It is not an independent market survey or evidence of measured demand for either product. Research discussion.

“Worth studying” needs a clear meaning before collection starts. A creator may be useful because a video shows a clear opening, explains a difficult subject or uses a format you want to understand. Those are different reasons to select someone.

For this workflow, define the result as a short list of people, each with a relevant source video and a written reason for selection. The reason should describe something visible in the content.

For example, a selection note could say that the opening introduces the viewer's problem before the demonstration starts. That is a content judgment that another person can check. A label such as “great creator” gives the reviewer much less information.

Also define the boundary of the assignment. Studying an example does not establish audience quality or commercial suitability. Those questions require a separate review. Keeping the purpose clear helps you collect the evidence that this decision needs.

Set a small search sample before you collect

Start with a narrow topic that matches the assignment. Write a small set of search phrases that a relevant video might use. Avoid combining several unrelated research questions in the first collection. A narrow first sample is easier to inspect and correct.

The checked Chronometrica schema requires `queries`. It also provides controls that determine how much material the Actor attempts to collect. Chronometrica input schema.

1. Set `expandQueries` to control whether related phrases can be added.
2. Set `maxResults` as the overall unique-video cap.
3. Set `resultsPerQuery` as the per-term maximum before the overall cap.
4. Set `maxPagesPerQuery` to bound attempted pages.
5. Choose `searchDistribution` with the effect of the overall cap in mind.

Balanced distribution can lower the effective per-term limit. These controls define an observed sample. They do not promise complete discovery of all relevant videos or creators. Chronometrica input schema.

This setup example uses documented fields and valid bounds. The query is an example, not a claim about search volume or returned content:

{
  "queries": [
    "ugc ad examples"
  ],
  "expandQueries": false,
  "maxResults": 6,
  "resultsPerQuery": 6,
  "maxPagesPerQuery": 1,
  "searchDistribution": "balanced"
}

Replace the query with a phrase that fits your assignment. The saved TikTok Content Idea Research Task is another starting point. A Task is a saved input preset. Review its settings before use so that your collection matches your question.

The first collection should help you assess relevance and record structure. Expanding a poorly chosen query only gives you more material to review.

Turn video records into creator candidates

A video search returns video records. It does not follow that each record represents a different creator. Chronometrica can return several videos from one author, so you must group records before reporting the number of people found. Chronometrica product details.

Use stable author identity to form these groups. Review uncertain identity matches before you merge records. If the available evidence does not resolve a match, keep that uncertainty visible. A clean-looking list is less useful than an accurate one.

Keep two views of the work. The first is the source collection, with the returned video records. The second is the creator shortlist, with one entry per selected author. Each shortlist entry should point back to a selected source video.

Choose the video that best supports the stated reason for selection. The most relevant example may not be the one with the largest visible count. If the assignment concerns hooks, the opening needs to support the selection. If it concerns demonstrations, inspect the demonstration.

Retain the other collected videos. They let a reviewer check how you reached the decision. They can also show whether your selection depends on one isolated example within the observed sample.

Work through a fictional example

The following example is fictional. It is not live TikTok data, a benchmark or a customer result.

Suppose a collection contains six relevant video records from three authors. Author A appears in three records. Author B appears in two. Author C appears in one.

The correct report is six observed videos and three distinct creator candidates. Reporting six creators would double the actual creator count in this example.

The researcher keeps all six source records. For the shortlist, the researcher selects one relevant video from each author. The shortlist therefore has three entries. Each entry includes its selected video link, caption and reason for study.

The arithmetic is simple, but the distinction matters. A reviewer who sees only the number six may think the research found twice as many people. Separating the two counts prevents that error without discarding useful content.

Now apply the assignment. If it asks for a concrete hook example for each person, the selected videos support the requested review. If it asks for profiles with bios and follower counts, the required starting record is different.

The three shortlist entries describe this fictional sample. They do not establish that only three suitable creators exist. A creator absent from the returned sample may still publish relevant material.

Inspect the content before you accept a candidate

A search phrase gives you a route to possible examples. It does not settle whether each example fits the assignment. Open the selected content and inspect the part that supports your reason.

For a hook study, identify what the opening does. Does it state a problem, show an outcome or introduce a question? Record the observed approach in plain language. Avoid assigning a quality label without saying what you saw.

For a format study, describe the structure that matters. You might need to inspect how the video moves from introduction to demonstration. Your note should make that structure clear enough for another researcher to check against the source.

Use the caption as part of the context. Do not assume that the caption alone describes everything in the video. The selected content and your written reason should agree.

Check the profile before confirming the shortlist entry. Confirm that the person you intend to study matches the selected source. Resolve any uncertainty that could change the entry.

Keep claims proportional to the evidence. One relevant video supports further study of that example. A broader statement about the creator's usual approach needs more content review. This keeps the shortlist useful without asking a small sample to answer a larger question.

Keep counts, missing values and search context clear

A study list needs enough context for someone else to understand its limits. Keep the query and collection settings with the source records. Record when you observed the material. This helps a reviewer understand which question produced the shortlist.

Report video count and distinct creator count separately. If you select fewer creators than you observed, show that distinction too. The returned collection and the final selection answer different questions.

Keep available public post metrics beside their source videos when they help the review. If a value is missing, preserve it as missing. Do not replace it with zero. A missing count does not establish that the post received no engagement.

Treat a visible count as context for the observed post. Explain any selection that depends on it. Otherwise, a reviewer may assume that the largest count determined the list when your actual purpose was to study a specific topic or format.

For each shortlisted person, retain the author identity, selected video URL, caption, selection reason and source context. This is a practical review record. It lets someone check both the person and the content that caused you to select them.

Use profile records when the assignment changes

The main switch condition is a change in the required deliverable. If the reviewer asks for a profile list, reconsider the starting record before doing more video collection.

A useful test is to write the required columns first. Are most columns about a person's profile, or about a specific piece of content? That question exposes a mismatch early.

Do not spend time collecting examples that the current assignment does not need. Equally, do not treat a completed profile list as a completed content study. If the next reviewer asks why a person deserves study, you still need evidence for that judgment.

This is a workflow choice. Decide which evidence must come first, then choose the Actor that supplies that type of record. Add further review only when the assignment requires it.

Use a short, repeatable review method

The full method can stay small. Use it for the first collection before increasing the search scope.

1. State the deliverable. Say whether you need a content-backed study list or a profile list. Define the evidence required for each accepted entry.
2. Write the topic queries. Keep the first set narrow enough to review. Fix query expansion, result caps and page limits before collection.
3. Collect the chosen record type. Check that the returned records contain the context your assignment needs.
4. Group video records by author. Review uncertain identities before merging them. Keep the source records available.
5. Inspect and select an example. Keep a relevant video and caption beside each proposed creator. Write the topic match as a human judgment.
6. Report the counts separately. State how many videos you observed and how many distinct creator candidates they represent. Keep missing values unchanged.
7. Confirm the shortlist. Check the actual content and profile. Record unresolved questions that could affect the decision.

After this review, decide what the next collection must resolve. A weak topic match may call for a better query. An uncertain selection may call for more content review. Increasing the cap is useful only when more records can answer the remaining question.

If your next task is to organize the selected examples, the TikTok UGC creative library guide covers that separate step. This comparison ends with choosing and reviewing the right discovery records.