Content Performance Agent
Reports engagement and conversion by piece.
- Industry
- Banking
- Department
- Content
- Team
- Content Planning
- Kind
- Investigator
- Approver
- Head of Content
- Code
- CON-12
It reports engagement and conversion by piece (guide reads, calculator use, account openings and loan applications started from content) and explains why pieces over- or under-perform, separating traffic source, search ranking, rate competitiveness, format, timing and tracking causes. It ranks the likely causes with evidence and updates the content performance report. **The problem it solves:** Why the pieces named performed as they did: reads, calculator use and applications against their benchmark or earlier period, when the change started, and what changed in publishing, distribution, search, rates or tracking around then.
How it works
6 steps, and a person signs off before step 6.
- 01Looks it up
Finds the record the request is about in the internet and mobile banking platform: views, engagement and application and account-opening events by piece and channel for the period, with rate-change dates..
- 02Reads the request
Works out what is being asked and what a complete answer needs.
- 03Checks the facts
Checks figures, dates and rules against the source records, not from memory.
- 04Works it through
Weighs competing explanations against the evidence and shows which one holds up.
- 05Asks a person
Hands the work to Head of Content with its reasoning before anything goes out.
- 06Acts
Once approved, it acts in the content reporting dashboard: it updates the content performance report with the period's figures, verdicts and causes. and records what was done.
What it works with
Your knowledge, your systems, your people.
Knows
The content measurement guide: the funnel from read to application, benchmarks by format and pillar, conversion events (account opened, application started, calculator completed), attribution rules and known tracking changes in the internet and mobile banking platform.
Reads from
the internet and mobile banking platform: views, engagement and application and account-opening events by piece and channel for the period, with rate-change dates.
Acts in
the content reporting dashboard: it updates the content performance report with the period's figures, verdicts and causes.
Hands back
An object with pieces (a list, each with piece_id, title, views, engaged_seconds, conversions, conversion_rate, vs_benchmark_percent and verdict (out_performing, on_target, under_performing or too_early)), causes (a list, each with piece_id, cause, evidence, confidence and how_to_confirm) and recommendations (a list with piece_id, fix and expected_effect).
Asks a person when
The report credits account openings or loan volume to content, recommends retiring a piece, or finds a tracking fault in the banking platform that affects other teams.
Settings
Your limits, set by your managers.
| Setting | Default | What it controls |
|---|---|---|
| Under-performing below (% of benchmark) | 70 | At or below this share of its benchmark a piece is under-performing. |
| Out-performing above (% of benchmark) | 130 | At or above this share of its benchmark a piece is out-performing. |
| Fewest sessions for a verdict | 500 | Below this a piece is too early to tell. |
| Conversions counted | account_opened, application_started, calculator_completed | Comma-separated events. |
| Attribution | last touch | How conversions are credited to pieces. |
| A drop this large across pieces is a tracking suspect (%) | 30 | A same-day drop this large across many pieces is checked as tracking first. |
Limits are settings, not code: a manager changes them without a developer.
Rules it always follows
- Give no verdict on a piece with fewer than 500 sessions.
- At or below 70 percent of benchmark is under-performing; at or above 130 percent is out-performing.
- Count only account_opened, application_started, calculator_completed, credited by last touch.
- A drop of 30 percent or more starting on one day across many pieces is a tracking or release change until shown otherwise.
- When a product's rate changed in the period, test rate competitiveness as a cause before blaming the content.
- Never state lending volumes or revenue the records do not give.
See the Content Performance Agent agent on your own work.
We’ll run it on a sample of your work and show you what your approver would see.