Find which videos bring new viewers

What to do

Compare the new-viewer and traffic-source data available in Studio across a consistent period. Note each video's topic and format.

Compare comparable entry points

Choose a consistent period and group videos by the kind of task they cover. Compare the available new-viewer and traffic-source data for entry-level lessons and follow-up videos. Look at which promise brought people in and which videos already assumed experience. Use the comparison to choose one follow-up or clearer starting point, then measure it in a similar period. Keep publication dates and format in the record.

Example or check

Group introductory tutorials apart from follow-up videos. New viewers may enter through different topics than returning viewers.

Related checks for analytics, discovery and growth

Turn viewer language into topics

Read comments and questions looking for a specific unresolved task. Group different phrasings around the same problem, then choose a question you can demonstrate with an actual example. A useful topic is small enough to answer completely and familiar enough that viewers can recognize it in the title. If three people ask why the same setting fails, show a working input and the common mistake. Avoid giving each phrasing a nearly identical video. Make the introduction and description reflect the problem you chose so the video is easy to identify without prior channel context.

Validate one decision with a small follow-up

Pick a change you can observe independently, such as using a shorter introduction in the next tutorial or making the outcome visible in the image. Write down what improvement would look like and which report may reveal it. Compare videos with a similar audience and subject when possible. If a result conflicts with your expectation, check whether the underlying counts or source mix changed. Keep the follow-up small enough to repeat. A series of documented comparisons is more useful than a single conclusion based on whichever report happened to look most favorable.

Use click rate and retention together

A click rate helps describe a counted impression becoming a view, while retention helps describe what happens during playback. A strong image and title can invite a click, but a long delay before the promised demonstration may lose the viewer. If one graph moves, check the other before concluding that a change worked. Read the first obvious retention change and replay that moment. Compare the viewer's likely expectation with what actually happens. If the lesson opens with setup that regular viewers already know, test a version that shows the result and then explains the setup.

Make time estimates useful

Watch-time planning starts with a consistent unit. Multiply views by average duration in seconds, then divide by 3,600 to express the result in hours. Enter inputs for the same set of views and date range. If average duration is rounded, treat the result as approximate. Compare it with Studio's own data rather than expecting an exact match. A watch-time total can be useful for comparing two videos, but it does not tell you whether the time qualifies for a monetization requirement. Use the Earn section for that separate question.

Compare like with like

Match the period, scope, and traffic source when comparing figures. A video reached through a homepage recommendation may attract people with a different intent from someone searching for a specific repair. An external link can contribute views without generating impressions in the same report. Inspect the denominator and the rules for the metric before calculating a rate. If you compare two formats, note video length and audience as well as the headline percentage. Keep a small observation table with the metric, date range, source, and sample size so the comparison can be reproduced later.

Make a clear destination for new viewers

When preparing a channel page or playlist, decide what a first-time visitor should see first. Put foundational videos before their follow-ups and use readable names that describe the next action. Check that the first lesson stands alone for someone who arrived through search. In a playlist, keep the order meaningful and distinguish a single-video link from a link meant to preserve the sequence. For channel art, check that identifying text remains visible on the device most viewers use. Look at the result from a viewer's perspective after publishing any layout change.

Choose a question before opening a chart

Start with a decision, such as whether to revise the thumbnail, shorten an introduction, or make a follow-up. Choose the data that can inform that decision. An increase in impressions says something different from an increase in average viewing duration. A report with a matched date range and traffic source is easier to interpret than a collection of headline numbers. Record the question and the most plausible alternative explanation before changing anything. This lets you distinguish what the data showed from what you inferred and choose a better next test if the result is unclear.

Treat small samples carefully

Small samples can move sharply because of one view or click. Include the raw counts next to percentages and avoid deciding from a single unusually good or bad day. A change in the traffic mix can also alter a combined rate even when each source behaves similarly. Review source breakdowns when available. For a newer video, collect enough comparable observations to make a practical decision. If you cannot get a good match, frame the result as a clue for your next upload instead of declaring that the thumbnail or title caused the change.

Review an opening at the exact drop

Find the time of a visible change in the retention graph and replay that portion as someone unfamiliar with the channel. See whether the action stops, a repeated explanation begins, or the promised result remains unclear. The graph can also reflect people skipping to a specific chapter or watching a segment again. Check the video itself before treating a dip as a mistake. Write down one editing change you can test in a future video, then record which other factors changed. This creates a useful trail even when the results are noisy.

Design a manageable revision

Keep the thumbnail, title, and opening changes separate when possible. If you change all three at once, you may improve the upload but have less evidence about what helped. Define the change before looking at the next data window. For a thumbnail, preserve both images and note what was different, such as subject scale or legibility. Check whether the traffic sources changed and whether the observation windows are comparable. If the platform offers a test for your account, read what it measures and compare that measure with the decision you want to make.