PLAYBOOK / OCTOBER 15, 2026 / BY DRAMAGLITCH
Don’t measure a game like a video.
A fan can spend three minutes inside a playable experience and still never watch your show. Plan the measurement around the behavior you actually want.
Start with a question the pilot can answer
“Increase engagement” is too broad to evaluate. A useful first question could be: does an interactive mystery generate more qualified visits to the next episode than a static campaign page for the same audience? Another could be whether existing fans return to a new weekly challenge. These are different objectives and need different experiments.
Decide which audience you are serving, where the campaign will run, how long you will observe it and what outcome changes the next decision. Write those choices down before launch.
Use a simple event vocabulary
| Signal | Definition to agree before launch |
|---|---|
| Playable ready | The experience has loaded and can accept an interaction. |
| Play start | The player makes the first meaningful move. |
| Completion | The player reaches a defined ending or completes the challenge. |
| Return click | The player activates the specified episode or series link. |
| Verified episode start | The destination confirms playback through an agreed, permitted integration. |
| Repeat play | A player returns within a defined period, where consent and measurement allow it. |
A return click is not an episode view. A page load is not a play start. If the destination does not provide playback confirmation, report that limit rather than renaming a click as a viewer.
Make the denominator visible
Completion rate could mean completed sessions divided by starts, or completed sessions divided by all landing-page visits. Both can be useful, but they answer different questions. A game may have excellent completion among starters and still lose most visitors while loading. Report the full funnel and specify session and deduplication rules.
For repeat play, define the time window and whether the unit is a person, a browser or a session. A browser identifier does not establish a unique person, and cross-device activity complicates that further. Be precise about what your implementation can actually observe.
Separate activity from incremental impact
Where feasible, randomly allocate comparable traffic to playable and non-playable variants. Keep the audience, campaign period, destination and creative promise consistent. Choose the primary metric, minimum detectable effect, sample-size plan and stopping rule before launch. A short test with few outcomes may be inconclusive, even when one percentage looks larger.
If random assignment is unavailable, describe the comparison honestly. Comparing this week’s game traffic to last month’s unrelated campaign can suggest a hypothesis, but seasonality, creative and audience changes prevent a clean causal claim.
Respect attribution and privacy limits
UTM parameters can describe a campaign visit, but do not prove a later purchase or viewing session came from the game. Cross-app attribution, consent requirements and platform policies can limit what is available. Use the least data necessary and agree retention, access and deletion rules before connecting production analytics.
Do not send email addresses, message contents or sensitive fan data as event properties. Do not fingerprint players to repair a measurement gap. For an early pilot, aggregate session signals and a clearly defined destination click may be enough.
Give the result a decision
- Continue: the primary outcome improves with enough evidence and acceptable production and distribution costs.
- Revise: players complete the game, but the series destination receives little qualified traffic. Revisit the story connection and final invitation.
- Fix delivery: visitors fail to reach a playable state on the devices or in-app browsers that matter.
- Stop or retest: results do not justify expansion, or the evidence remains too weak to make a confident decision.
This framework is DramaGlitch’s proposed measurement approach, not a published performance benchmark. Implementation and analysis belong in the agreed project scope.
Plan a measurable experiment