The Custom Report Builder

When none of the built-in reports answers your question, build your own: pick the data, narrow it, group it, chart it, and save it so it runs on demand or on a schedule.

A custom report is a saved definition, not a one-off export. Once saved it behaves like any built-in report — you run it, export it, schedule it, and share it — and it re-queries live data every time, so last month's report reflects this month's corrections.

Who can build one

The builder is an administrator tool. Employees never see custom reports — not the builder, not the "Create custom report" action, and not saved reports shared with the account.

Some datasets carry payroll-sensitive figures and are limited further, to Payroll and HR Administrators. If a dataset doesn't appear in your list, that's why — your role can't reach that data anywhere in the platform, and a report is no exception. See Roles & Permissions.

Anatomy of a report

Every custom report is built from the same seven parts:

The builder lays these out as panes side by side. Available Fields is searchable and grouped by business area — Person, Employment, Compensation, Time & Attendance — with an Add all per group, so you look where a field belongs rather than hunting an alphabetical list.

Each chosen column can be renamed; the row then shows "Renamed from…" so the original is never lost. Columns also carry badges telling you how they behave: GROUPED BY or TOTALLED PER GROUP. Reorder by dragging, or focus a row and press Alt + ↑ / ↓.

Data Sources shows the dataset with a plain-English Shape of the result — "One row per Department, then Schedule Status · 5 columns · Sorted by Department, a to z" — which is the quickest way to confirm you built what you meant.

Preview runs the report against real data whenever you want to check your work; Update (or Save on a new one) commits it. A preview is a sample rather than the finished figure — see Reading the preview for what it tells you about itself.

Building one, step by step

1. From Reports, click Create custom report.
2. Give it a Name.
3. Choose the Dataset (see the list below). The column picker repopulates from your choice.
4. Pick your Columns. Start with the few you'd actually read; you can add more once the shape looks right.
5. Add Filters to narrow the rows.
6. Optionally set a Group-by and a Sort.
7. Optionally add Charts.
8. Check the Preview, then save with a visibility.

Datasets

There are 32 to choose from. A dataset is the grain of your report — one row per employee, per punch, per payslip line, and so on. Choosing the right grain is most of the work.

The work-schedule dataset is a useful one to know: it brings Work Schedule, Schedule Status, Scheduled Hours and Scheduled Cost, which is how you compare what you rostered against what you paid.

Joined measures

Joined measures bolts extra figures — hours, leave, time-clock — onto a report without changing its dataset. It's only offered where it makes sense: datasets with one row per employee. On a per-day or per-transaction dataset the section tells you plainly that there are none to include.

Filters

Filters decide which rows qualify. Each row is a column, an operator, and a value; stack as many as you need.

The set as a whole reads as All of (every condition must hold) or Any of (one is enough) the following — the equivalent of AND and OR. + Add condition adds a row; + Add group nests a whole sub-set with its own All-of/Any-of, which is how you express "in Nassau, and either hourly or on contract".

A condition can also be set to Ask on run — instead of hard-coding the value, the report prompts for it each time it runs. One saved report then answers "which department?" for every department, rather than needing a copy per team.

Whoever opens that report sees the questions in a panel headed Asked each run at the top of the results: one input per asked condition, and a Run button. Answering changes only the values the author opened up — the structure, the other conditions and the dataset's own constraints stay as saved, and running with a different answer never edits the report. Two people can hold different answers on screen at the same time without affecting each other or the saved definition.

Where a report covers a period, you can use a preset — Yesterday, Month to Date, Quarter to Date, Year to Date, Last 12 Months — or Custom Range. Presets are relative, so a scheduled report keeps meaning the same thing each time it runs. Yesterday runs to the end of yesterday like every other preset, which makes it the one to pair with a daily schedule: a report that runs each morning covers the whole of the day before rather than a partial one.

Conditions the dataset always applies

Some datasets carry their own built-in constraints. These appear above your filters with a padlock and the words Always applied, written out in the same plain language as your own conditions — and you can't edit or remove them. They're shown rather than hidden so a total that looks short has a visible explanation instead of being a mystery.

Filter values read as words, not codes: a department filter shows the department's name rather than its internal ID, so a saved report still makes sense to whoever opens it next.

Employment level, across the datasets

Employment level is now available to filter and group by on many more datasets — performance reviews, recognition, salary scales, tax contributions, time entries, scheduled hours, and the punch datasets among them. It is the dimension most often needed to split a report the way an organisation actually pays people (full time, part time, contract, casual), and previously you had to approximate it with department or work title.

Report warnings

Before a report runs, the builder tells you what it's actually going to do. Two kinds of message appear:

• Blocking (red) — the report can't run as configured. Fix it before saving.
• Advisory (amber) — it will run, but something probably isn't what you meant.

These exist because two mistakes used to be invisible until the results arrived, if at all: grouping quietly changes which columns come back, and a filter left without a value quietly matches nothing. Read the warnings before you trust a total.

Reading the preview

The preview tells you what it is actually showing, which matters because it is a sample and a total is not.

Charts

A report can carry charts as well as a table. Click Add chart and configure each one; use Move chart up / Move chart down to order them.

Showing a breakdown: stacked or small multiples

Once a chart has a breakdown you choose how to show it:

• Stack series — the series sit on top of one another in one chart. Best for reading the total while still seeing the mix.
• Small multiples — one panel per series, shared scale, as the control itself puts it. Best for comparing shapes across series, because the shared scale keeps the panels honest.

They're alternatives, not a combination — pick one. Both need a breakdown column to exist first, and a KPI tile or Pie can't be broken into series at all.

Saving & sharing

Save the report with a visibility:

Sharing a report never shares access to data: each person still sees only the employees their own role and scope allow.

The builder spells out the consequence before you commit — changing a report to account-wide says "Everyone in your account will be able to see this report from now on", and narrowing one back to yourself warns "Everyone else will lose access to this report. Anything scheduled from it keeps running." That last clause matters: restricting visibility does not stop a schedule that's already sending it. After saving you land on the saved report rather than back in a list.

A report you'll want to build variations of can also be saved as a template to start from next time.

Running & exporting

A saved report runs on demand from the reports list, and exports as Excel (.xlsx), CSV (.csv), or PDF (.pdf). Excel for onward analysis, CSV for loading into another system, PDF when it's going to a person rather than a spreadsheet.

Scheduling

Any saved report can run on a recurrence and email its output:

• How often — Daily, Weekly (choose the day), or Monthly.
• Who gets it — TriblockHR users, and any email address you type in. That last part matters: an accountant or auditor with no login can receive the monthly pack directly.
• What they get — the export format you chose, generated fresh at run time.

Pair a schedule with a relative date preset. Month to Date on a monthly schedule keeps reporting the right window forever; a Custom Range will still be reporting last March next March.

The Runs log records every execution with its status — Queued, Running, Succeeded, or Failed — plus row counts, duration, who ran it, and the error when one fails.

When it won't save

The builder validates a chart before letting it through. The messages say exactly what's wrong:

That last one is the common trap: charts reference columns by name, so pruning the column list can silently orphan a chart you built earlier.