Analyse a selection
The Analyste tendances (Trends Analyst) studies the cards you give it and looks for an original adaptation for Fortnite. You will know how to start an analysis, read it and put it in its proper perspective.
Before you start
Section titled “Before you start”- A connected AI account. The analysis uses the agent and the model you have configured, and uses up your quota. Collecting the sources does not.
- The refresh of the sources must have finished.
Start an analysis
Section titled “Start an analysis”- In Tendances Maps (Trends Maps), tick the checkbox at the top left of the cards, 8 at most. Sélectionner les 8 premières (Select the first 8) ticks the first cards in the view.
- In the bar that appears at the bottom, click Analyser la sélection (Analyse the selection).
- A conversation opens, and the agent works in it.
For a single card, open it and click Analyser cette inspiration (Analyse this inspiration).
The analyst only reads: it does not change your files or the editor, does not start any other agent and does not build any prototype.
Read the result
Section titled “Read the result”The result is saved in each card, under Analyse enregistrée (Saved analysis), with the date and the project it was made for. A comparative summary stays in the conversation.
| Verdict | Meaning |
|---|---|
| À explorer (To explore) | An original lead, backed by evidence, worth prototyping. |
| À surveiller (To watch) | A mixed or short-lived signal. |
| À écarter (To rule out) | Specific evidence goes against this lead. |
| Preuves insuffisantes (Insufficient evidence) | The concept cannot be judged seriously. |
Six criteria are scored from 0 to 4, or Inconnu (Unknown) if there is no evidence on which to score. There is no overall score.
| Criterion | Question |
|---|---|
| Audience observée (Observed audience) | How many players, in what sample? |
| Dynamique récente (Recent momentum) | In which direction, and over what period? |
| Engagement | Do players stay and come back, with figures to back it up? |
| Différenciation (Differentiation) | What original variation can you offer in UEFN? |
| Faisabilité UEFN (UEFN feasibility) | What is the smallest possible prototype? |
| Lisibilité du concept (Concept clarity) | Is the promise clear and honest? |
Each card also gives À réinventer (To reinvent), Risque principal (Main risk) and Prochain test (Next test). Réexaminer avec l’analyste (Re-examine with the analyst) replaces the card’s previous analysis.
What an analysis does not prove
Section titled “What an analysis does not prove”Island Pilot says so itself on each card: “une hypothèse à valider” (a hypothesis to validate).
- It is not a revenue forecast or a promise of placement in Discover.
- The analyst has not played the game. It reads the card’s data. With only a title and a badge, it scores Inconnu for anything that depends on gameplay.
- It has not watched the videos: the title and the thumbnail do not prove the content.
- A high audience proves neither retention nor revenue.
- A badge proves neither the audience nor the appeal of a game.
- An analysis is dated. It is not redone on its own when the figures change: check its date before relying on it.
The only way to find out is the proposed test, on your island, with your players.