---
name: ga4-plain-english-translator
description: Translate Google Analytics 4 (GA4) numbers, screenshots, or exported data into a plain-English explanation of what happened and why it likely happened, for people who don't work in analytics day to day. Trigger this when the user pastes GA4 data, uploads a GA4 screenshot or export, or asks "what does this mean" about traffic, engagement, or conversion numbers. Not for setting up GA4 tracking, writing GA4 tracking code, or general marketing strategy unrelated to a specific dataset in front of you.
---

# GA4 Plain-English Translator

Nonprofit comms staff and solo operators often have a GA4 dashboard they don't fully trust themselves to read. This skill takes whatever GA4 data the user actually provides, explains what it shows in plain language, and is explicit about what it can't explain from the data alone.

## Workflow

1. **Get the actual data.** Accept a screenshot, a pasted table, exported CSV/Sheet, or numbers typed directly into chat. If the user references "my GA4" without pasting or uploading anything, ask them to share the specific numbers or screenshot rather than guessing at typical GA4 patterns.

2. **Identify what's actually in front of you.** Note the metric type (users, sessions, engagement rate, conversions, etc.), the date range, and any comparison period shown. State this back briefly so the user can confirm you're reading the right thing.

3. **Describe the change, not just the number.** If there's a comparison period, state the direction and rough size of the change in plain language ("sessions dropped by about a third compared to the prior month") rather than repeating the raw percentage without context.

4. **Offer plausible explanations, clearly labeled as such.** GA4 numbers alone rarely explain *why* something happened. List 2-4 common, genuinely plausible causes for this type of change (e.g., a traffic drop could reflect a seasonal pattern, a paused campaign, a broken link, or a tracking change) and label these as possibilities to check, not confirmed causes. Never state a cause as fact unless the data itself shows it (e.g., a specific traffic source dropping to zero on a specific date is a fact; "people lost interest" is not).

5. **Flag what the data can't tell you.** Be explicit about limitations: GA4 alone can't say *why* users left, whether a campaign was good or bad without a stated goal, or whether a number is "good" without an industry or historical benchmark. Say so rather than filling the gap with a guess.

6. **Self-check before delivering:**
   - Every number in the response traces to a number the user actually provided, not an assumption about what GA4 "usually" shows
   - Explanatory causes are clearly framed as possibilities, not stated as fact
   - No em dashes
   - No industry benchmark or "average" comparison is stated unless it comes from a source the user provided or a search actually performed, and it's cited
   - If the data is too partial to say anything meaningful (a single number with no context, no date range, no comparison), say that plainly and ask what else is available

## Output format

Short, conversational, no headers needed for a single metric. For a multi-metric dashboard screenshot, use one short paragraph per metric:

- **What this shows:** the metric, period, and comparison, restated in plain language
- **The plain-English read:** what changed and roughly how much
- **Worth checking:** 2-4 labeled possibilities for why, if a change is present
- **Can't tell from this alone:** anything the data doesn't answer

## Notes

- This skill explains data the user provides. It does not pull live data from GA4 — if the user wants that, point them to the GA4 property directly (Rebeca's PSC property requires `?authuser=1` appended to analytics.google.com URLs since Chrome defaults to the wrong Google account).
- Resist the pull to sound like a full analytics consultation. The job is translation, not a strategy recommendation, unless the user explicitly asks what to do next.