What this checkpoint covers
Gristmill Bakery is an invented artisan bakery with three shops and a growing online bread-school business. It posts long baking tutorials and Shorts, and it runs paid video ads for its online classes. The questions draw on the retention curves guide, the ad creative hook test design and the creative production playbook.
Task
Answer each question in a sentence or two with the reason. Then compare with the explanations, count a right action with a wrong reason as a miss, and reread the linked guide behind each miss.
Questions
- A sourdough tutorial has a healthy CTRShare of impressions that result in a click. but loses a large share of viewers in the opening seconds. The social lead wants a new thumbnail. What do you advise?
- Across one video's curve there are a dozen small wiggles and one sharp dip that does not recover. Which deserve attention, and how do you check?
- A short croissant video has a higher average percentage viewed than a long bread-scoring tutorial, so the team wants to make only short videos. What is wrong with the comparison?
- The paid team wants to test a new ad that has a different opening, music, offer and ending against the current one, both in the same ad set, and to read it after two days. What would you change?
- The hook test needs more purchases per arm than the budget buys in a month. What are the options?
- Nobody can say which version of last month's best ad ran on which placement, and the studio never sees results. What is missing?
- The best-performing ad's cost has crept up while its FrequencyAverage times a person saw an ad. has climbed. How should the production cadence respond?
Answers explained
- Fix the opening, not the packaging. A healthy click-through rate means the title and thumbnail attract the right click; a steep early fall means the opening breaks or delays the promise. Open on the result promised, cut the preamble, and check the traffic source filter before blaming either. See the intro: the first drop.
- The sharp dip that does not recover. Small wiggles are normal attrition. A sharp, unrecovered dip marks a moment that gave viewers a reason to leave. Watch that second with the curve open, and confirm the shape repeats across videos or traffic sources before rebuilding a template around it. See dips, spikes and plateaus.
- Absolute retention cannot compare different lengths. Short videos keep a larger share by construction, and loops push the share up. Use relative retention to judge each against videos of similar length, then give Shorts and long-form different jobs rather than choosing one. See relative retention and comparing videos.
- Change one thing and separate the audiences. With four differences, a result cannot be credited to the hook. Keep everything after the opening identical, use the platform's split test so delivery cannot favour one version, budget equally, and read only once the planned duration and purchases per arm are reached. See design.
- Test on a more frequent event, or wait. The quantity that matters is conversions per arm. Either run longer, raise the test budget, or test on a higher-funnel event such as a class enquiry and accept that the result speaks to that event. Do not read early because the budget is tight. See sample size reasoning.
- A naming convention, a library and a results review. Name files by campaign, concept, format, placement, version and date from the first file, keep approved assets and their results in one library, and hold a regular review where results reach the people who made the work. Record each A/B TestControlled comparison of two variants. against the concept it tested. See asset naming, versions and the creative library.
- Plan refreshes before fatigue, from the backlog. Rising cost with rising frequency means the audience has seen it enough. Keep new concepts in production on a fixed cadence, built from recorded learnings, so a fresh version of the winning idea is ready when the current one tires. See cadence, inputs and outputs.
Where to go next
If the diagnosis questions caught you out, redo the video retention exercise for a woodworking school. Then move to the final module with the complaint wave exercise for a train ticket app. To revisit platform playbooks and adapting one idea across formats, return to the intermediate social media level.
