The launch dashboard went blank before the review
At 4:40 p.m., a project lead opened the team’s Notion dashboard before a customer review. The “Ready this week” count showed nothing. Five minutes earlier, the team had added a new row to the delivery database and changed an owner field. “The formula is broken,” someone said. Another teammate reached for the formula editor, ready to replace the expression while the review deck was already open.
I stopped the edit. A blank result can mean several different things: the formula may intentionally return an empty string, a referenced property may be empty, the formula may point to a renamed field, or a database view may hide the rows that would explain the number. A blank is not automatically a calculation error, and changing the formula before checking its inputs can erase the clue.
Notion formulas can use database properties, built-ins, and functions, and their result can depend on the values in each row. The official formula guide also documents empty() as a function for checking empty values; notably, Notion treats 0, an empty string, and an empty list as empty for that function. That detail matters when a dashboard uses zero to mean “no outstanding tasks” but a formula uses empty() as if zero always meant missing data.
Trace the result from one row to the dashboard
Start with a row that should appear in the count. Open the source database itself, not just the dashboard card. Check whether the row exists, whether its status and owner are filled in, and whether it meets the intended date range. Then open the formula property for that row and read the result. The formula editor can show a live preview for a database row, which helps separate a row-level result from a dashboard or view issue.
Next, inspect every property named in the formula. A field may have been renamed, replaced, or left blank on newly added rows. If a formula returns text only when several conditions are true, one false condition can route the row to an empty-string branch. Read the expression from the outside inward: identify the first condition, the values it checks, and the final value returned by each branch. Compare one row that works with one that does not, changing no data while you compare.
Then check the formula’s meaning of “empty.” Notion’s official syntax reference says empty(0) is true, just as empty("") and an empty list are true. If zero is a valid result in the workflow, do not use empty() alone to distinguish zero from missing input. Decide what the dashboard should show for a genuinely missing value, a measured zero, and a value that has not been entered. Make those states explicit in the formula and test each with a representative row. Avoid pasting a speculative rewrite into the live database during a deadline.
| Symptom | First check | Safer next move |
|---|---|---|
| One row is blank | Referenced properties in that row | Compare inputs with a working row |
| New rows alone are blank | Defaults and required fields | Add a test row and fill each input deliberately |
| Zero looks like “missing” | empty() logic | Decide whether zero is a valid result before changing the condition |
| Source rows look right but card is blank | View filters and grouping | Open the underlying database and compare its filters |
| Formula preview works but dashboard does not | Dashboard source and selected view | Confirm the linked database and view match the intended data |
After the row-level check, inspect the dashboard card or linked database view. Confirm that it points to the expected database and view, then review filters, grouping, date windows, and hidden properties. A formula can calculate correctly on every row while a filtered view shows no rows. Conversely, the view may contain the right rows but a card or rollup may summarize a different property than the one the team expects.
Fix the cause without losing the working version
Create a small test row or duplicate the database in a safe workspace if the change affects many records. Write down the expected result for three cases: a missing input, an explicit zero, and an ordinary value. Then edit the formula in the test copy and verify those outcomes. The goal is not simply to make a blank disappear; it is to preserve the distinction the team relies on. A dashboard that shows “0” for missing owners may look reassuring while hiding incomplete data.
If a property name changed, update the formula to reference the intended property and confirm the output type. If the expression deliberately returns an empty string for a nonmatching status, check whether that condition still reflects the team’s rules. If the underlying row is incomplete, decide whether to fill the value, make it required in the workflow, or label the result as pending. These are different fixes; the formula should not silently invent business data to make a card look complete.
For a dashboard count, verify that the view includes the rows you intend to count and that the aggregation is set on the right property. Test the change against a handful of known rows, then review the whole view before using it in a customer meeting. If the formula feeds an automation or another database property, check those downstream uses too. A formula update can change more than the visible card, so keep a copy of the original expression and note what you changed.
When the result remains unexpected, use Notion’s formula editor feedback and official troubleshooting guidance rather than guessing from the word “blank.” The formula guide explains how to open the editor and inspect available properties, functions, and the row preview. The help article on common formula errors covers cases where unexpected empty values can affect formulas and automations. Compare the exact property names and row values against those references; do not assume a blank means Notion lost the record.
The project lead in the review room did not need a clever formula rewrite. The useful clue was that the new rows had a different owner state, and the dashboard view filtered for assigned work. The team checked the source records, made the owner state visible, and kept the existing formula while they tested the view rule. The meeting started with a clear explanation instead of a last-minute database change. A quiet blank had pointed to an unmade decision in the workflow, not necessarily a broken calculation.
Record what the dashboard is meant to say
Before changing anything, write down what the card should mean in plain language. “Ready this week” might count only rows with an owner, an approved status, and a due date inside the current week. If the team cannot agree on those conditions, a formula cannot resolve the ambiguity. Use one example row for each expected state and have the project owner confirm the result before changing the shared view.
After a successful correction, keep a short note with the property names, the view filters, and the three test cases. This helps the next teammate understand whether a blank means missing input, intentionally excluded work, or an actual calculation problem. It also prevents someone from removing a helpful condition just to make the card display a number. The dashboard is useful when its meaning is stable, not merely when its cells are filled. Recheck the card after saving so the team sees the verified view. before the review.
Frequently asked questions
Does a blank formula result mean Notion has an error?
Not necessarily. A formula may return an empty string for a branch, use an empty source property, or be hidden by the current view. Check a specific source row and the formula preview first.
Why does empty() treat zero as empty?
Notion’s formula syntax reference says 0, "", and [] count as empty values for empty(). If zero is meaningful in your process, design and test the condition with that behavior in mind.
Can I fix the formula directly in the production database?
If it affects many rows or downstream automations, test the intended cases in a safe copy first. Keep the original expression so the change can be reviewed or reversed.
See Notion’s official guides to formulas, formula syntax and functions, and common formula errors.
Quick takeaway Check the source row, empty-value logic, and dashboard view before editing. Test missing, zero, and ordinary values in a safe copy.