Quick Summary
- This tracker records the question, title options, dates, traffic context, observed metrics and decision for each title experiment.
- For YouTube's native test, record its reported outcome and test type; do not assume the interface exposes every metric for every variant.
- For manual before-and-after edits, keep the observation separate from a controlled A/B test because the audience and traffic can change.
- Copy the CSV header below into a spreadsheet, then use the field guide to avoid confusing impressions, CTR and watch time.
- Leave unavailable values blank or mark them “not reported.” Never convert missing analytics into zero.
You change a title on Tuesday and CTR looks better by Friday. Did the title help, or did the video start reaching a different audience? A useful test log keeps the dates, traffic source and other changes beside the numbers, so your next decision is based on what you actually observed.
Copyable tracker with a filled example
Paste this CSV into a blank spreadsheet. The second row is illustrative, not a real Scriptio or YouTube result. Replace it with your own video and Studio figures. Use one row for each experiment or manual observation.
You can also Scriptio is on a break until April 1, 2027 and open it in a spreadsheet. It contains the same header and example row shown below; no sign-in is needed to get the file.
test_id,video_title,video_url,format,published_date,hypothesis,test_method,test_type,control_title,variant_a_title,variant_b_title,variant_c_title,thumbnail_changed,experiment_start,experiment_end,before_range,after_range,traffic_source,impressions_before,ctr_before,views_before,watch_time_before,average_view_duration_before,impressions_after,ctr_after,views_after,watch_time_after,average_view_duration_after,native_test_result,other_changes,decision,lesson,review_date
EXAMPLE-01,Phone camera review,https://youtube.com/watch?v=EXAMPLE,long-form,2026-09-20,Specific camera angle may clarify the promise,manual before/after,not applicable,Phone X Review,Phone X Camera: Day and Night Samples,,,no,2026-10-05,2026-10-12,2026-09-28 to 2026-10-04,2026-10-05 to 2026-10-11,YouTube Search,1000,3.2%,80,240 min,3:00,1100,3.8%,90,270 min,3:00,not applicable,No thumbnail edit; search terms may have changed,Keep watching; no causal winner,CTR was higher in the later window but this was not a controlled test,2026-10-19
For a native A/B test, fill in the method, test type, options, start and end dates, and the result displayed in Studio. If YouTube does not show a value for an individual variant, leave that field blank or write “not reported.” Do not invent a per-title CTR or watch-time breakdown.
For a manual edit, enter the old and new titles, the edit date, the exact before and after date ranges in your notes, traffic source and metrics that Analytics actually makes available. Label the method “manual before/after.” The example's two windows have similar impressions, but it still does not isolate the title as the cause.
Field guide: what to record and why
| Field | What to enter | Why it matters |
|---|---|---|
| Hypothesis | One expected audience response, such as clearer product naming | Makes the experiment answer a specific question |
| Test method and type | Native title-only, native combined package or manual before/after | Separates controlled platform tests from observations |
| Before and after ranges | Two explicit calendar ranges of the same length for a manual edit | Makes the observation reproducible and exposes timing differences |
| Traffic source | Search, Browse, Suggested or another displayed source | CTR changes with audience and surface |
| Impressions | The count and date range shown in Analytics | Gives the CTR context |
| CTR | The displayed impressions CTR for the same date range | Indicates how often counted impressions led to a view |
| Watch time and average view duration | Only the values Studio reports for that content/range | Adds post-click context |
| Other changes | Thumbnail, description, promotion, edits or notable timing | Helps identify confounding factors |
| Decision and lesson | Keep, revise, test again, or no clear conclusion | Prevents a number from being mistaken for an action |
YouTube's native title and thumbnail tests use watch time to choose an outcome. YouTube also warns that CTR varies by audience and traffic source. Use the low-CTR diagnosis guide before deciding that a change is warranted, and the title-testing walkthrough for the native test workflow.
How to interpret the result
A “winner” is useful only for the tested video, options and audience. Save the result and consider whether the options were meaningfully different. If the test reports no clear winner, record that rather than forcing one. Use the finding to create a new hypothesis for a related video, not as a rule for every title on the channel.
For a manual comparison, do not write “the new title caused CTR to rise” unless a suitable experiment supports that causal conclusion. Use wording such as “CTR was higher during the later date range; impressions and traffic sources also changed.” That language preserves what the data shows and what it cannot establish.
What metrics should a YouTube title tracker include?
Record the displayed impressions, CTR, views and watch-time measures for consistent dates, along with traffic source, title options, test type and other changes. The exact metrics available vary by report and experiment.
Does YouTube choose a title-test winner by CTR?
No. YouTube's native A/B title and thumbnail test selects based on watch time. CTR remains useful context, but it is not the native winner criterion.
Copy and use the tracker
Copy the CSV above into a spreadsheet, delete the example row and leave unreported fields empty. Before every test, write one hypothesis and record the title options exactly as shown in Studio.
Sources: YouTube A/B testing titles and thumbnails, YouTube impressions and CTR FAQs, YouTube Analytics impressions and watch time