# Insights

> Build ClickClacks Insights: chart any event over time with counts, unique people, sessions or property math, split it by a breakdown, combine series with formulas, and save it for your team.

- Canonical URL: https://clickclacks.io/docs/guides/insights
- Section: Product guides
- Last updated: 2026-09-26

Insights answers “how many, how often, and split by what?” Pick an event, choose what to count, and see it over time. Add series, a formula or a breakdown, then save it for your team.

## When to use it {#when}

- “How many people signed up each day this month?”
- “What is our average order value, week by week?”
- “Which plan uses the export feature most?”
- “What share of visitors start a trial?” (two series and a formula)

For steps in order, use [Funnels](https://clickclacks.io/docs/guides/funnels.md). For whether people come back, use [Retention](https://clickclacks.io/docs/guides/retention.md).

## Build an insight {#build}

1. Go to **Insights** and click **New insight**. It starts as **Any event** counted as **Total events**, drawn as a daily line.
2. Under **Series**, pick an event. The picker lists **Your events** and **Autocaptured** ones, and you can type a name that hasn’t arrived yet.
3. Choose a measure, such as **Unique people**.
4. To narrow this series only, click **Where** and add up to five property filters.
5. Click **Add series** to compare another event on the same chart. You can have up to five series, lettered A to E.
6. Set the date range, source and segment in the [filter bar](https://clickclacks.io/docs/guides/filters.md). Those apply to every series.
7. Name it in the header and click **Save**.

A new series copies the one before it, counted as Unique people. So to compare “signed up” and “started trial”, add a series and change only the event.

## Measures {#measures}

| Measure | What it counts |
| --- | --- |
| Total events | How many times the event happened. |
| Unique people | How many different people did it. |
| Per-person average | Events divided by people: how often each person does it. |
| Sessions | How many visits included it. |
| Sum / Average / Median / Minimum / Maximum / 90th percentile of a property | Math over a numeric event property, such as an order’s amount or a load time. |

The property measures ask for a **Numeric property**, and the picker lists only numeric properties seen on recent events. Send numbers as numbers (`49`, not `"49"`) so they show up here.

## Breakdowns {#breakdowns}

Click **Add breakdown** to split every series by a property. Choose **Event property** (such as Country or UTM source) or **Person property**, which uses each person’s newest traits from identify, such as `plan`. Then choose **Top 5**, **Top 10** or **Top 25**. Everything else is grouped as **Other**.

Referrer and UTM breakdowns use the _visit’s_ values. See [Referrer and UTM](https://clickclacks.io/docs/guides/filters.md#referrer).

## Formulas {#formulas}

Click **Add formula** to combine series with letters, numbers, `+ − × ÷` and brackets. With two series it starts as `B / A * 100`, which is B as a percentage of A. You can give it a name, such as _Signup rate_. The formula takes effect when you press Enter or leave the field, and only once it reads correctly.

## Charts and intervals {#charts}

- **Line**, **Stacked bar** and **Table** show the metric over time.
- **Bar**, **Big number** and **Pie** total the whole range. Pie needs exactly one series, no formula, and a breakdown.
- **Interval**: Hourly, Daily, Weekly or Monthly. A range of a single day is always hourly, and Hourly works for ranges up to 31 days.
- **Rolling window**: 7-day or 28-day average or sum, with the Daily interval. Sums only apply to counts and sums.
- **Cumulative** adds each point to the ones before it. It works for Total events, Unique people, Sessions and Sum. Averages don’t add up, so it is unavailable for them.

Click a legend entry to hide or show a line. With **Compare** on, you get previous-period lines, a “was” figure, the change, and a _Previous_ column in the table.

## Drilling into a point {#drill}

Click any point, or focus it with the keyboard and press Enter, to see what is behind it. Series counted by people open a list of people from the most recent 50 matching events. Other measures open the events themselves. You can’t drill into a formula line, into _Other_, or into a person-property breakdown.

## Saving and sharing {#save}

- Insights are saved for the whole project. The list has **All** and **Mine** (the ones you made), with search and sorting by **Recently updated**, **Name A–Z** or **Created**.
- **Unsaved changes** appears in the header when you have edited a saved insight. **Save** overwrites it, and **Save as new** makes a copy.
- The whole insight lives in the page address, so you can paste the link to a teammate who is a member of the project.
- **Export CSV** downloads the table.
- **Add to board** works on a saved insight with no unsaved changes. The tile stays linked, so later edits show up on the board too.
- Deleting an insight removes it for everyone. Any board tile linked to it will say so. You can undo from the toast.

## Tips and gotchas {#tips}

- **Charts are capped at 5,000 points.** If you hit it, shorten the range, widen the interval or lower the breakdown limit.
- **Hidden and merged events** in [Lexicon](https://clickclacks.io/docs/guides/lexicon.md) are folded out of the event picker. A merged name counts as the event it was merged into.
- **A link that doesn’t quite fit** is adjusted rather than rejected, and the page says what changed.

## Recipes {#recipes}

### What share of visitors start a trial, day by day?

1. Click **New insight**. Set series A to `$pageview`, counted as **Unique people**.
2. Click **Add series** and set series B to your trial event, such as `Trial started`, also **Unique people**.
3. Click **Add formula**. Keep `B / A * 100` and name it _Trial rate_.
4. Save it as _Trial rate_.

The formula line is the percentage of that day’s visitors who started a trial.

### Which campaigns bring the most sign-ups?

1. Set series A to `Signed up`, counted as **Unique people**.
2. Click **Add breakdown**, choose **Event property** and **UTM campaign**, then **Top 10**.
3. Switch the chart to **Bar** to rank campaigns over the whole range.

Because UTM is visit-level, sign-ups count toward the campaign that brought the visit, even if the sign-up page itself had no tags.

### What is our average order value, week by week?

1. Set series A to your order event, such as `Order completed`.
2. Choose **Average of a property** and pick `amount` as the numeric property.
3. Set **Interval** to **Weekly** and the dates to **Last 90 days**.

### Why did sign-ups drop on Tuesday?

1. Chart `Signed up` as **Unique people** over **Last 30 days**.
2. Add a breakdown by **UTM source**, then try **Browser** and **Country**, and look for the one value that fell. To check one source at a time, use the Source picker in the filter bar.
3. Click Tuesday’s point on that value to see the people behind it.

If one browser fell to zero, it is likely a bug. Check [Friction](https://clickclacks.io/docs/guides/friction.md) for that day. If one campaign stopped, it is likely traffic. To hear about drops like this next time, set up an [alert](https://clickclacks.io/docs/guides/alerts.md).

## Related {#related}

- [Filters, dates and comparisons](https://clickclacks.io/docs/guides/filters.md)
- [Funnels](https://clickclacks.io/docs/guides/funnels.md): Steps in order, and where people drop off.
- [Boards](https://clickclacks.io/docs/guides/boards.md): Put a saved insight on a board.
- [Events and identify](https://clickclacks.io/docs/events.md): Send the events and properties you want to chart.