Explorer Mode Basics
Explorer Mode Basics
Open any dataset from Analytics Studio and you land in a lens — this is Explorer mode. The screen shows one big number and nothing else: Count of Rows, the entire dataset. That number is the answer to the most boring possible question ("how many rows are there?"), and every useful thing you do from here is refining it.
Learn six moves and you can answer almost any single-dataset question: measure, group, trellis, history, limit, and filter.
1. Change the measure
The measure is the number you're aggregating. Click the measure at the top of the lens — it reads something like Count of Rows — and you get two dropdowns: the aggregation function and the field.
Change Count to Sum and pick Amount, and the lens now shows total pipeline instead of a row count. Same dataset, completely different question. Average, Min, Max, Median, and Unique count are all there too.
You're not limited to one. Click the + next to the measure to add a second measure — say, Count of Rows alongside Sum of Amount. Now you can see volume and value side by side, which is usually far more informative than either alone. A hundred small deals and one enormous one produce very different stories from the same total.
2. Group by a dimension (bars)
A single number is rarely the answer. Click Bar (in some releases labelled Group) and pick a dimension — Industry, Stage, Account Owner — and the lens splits that measure into one bar per value.
You can group by a second dimension too. Group by Stage, then add Industry, and you get a stacked or clustered breakdown depending on your chart type. Ordering matters: the first grouping is the primary axis, the second nests inside it.
3. Trellis: one chart becomes many
Trellis breaks your single chart into a grid of small charts — one per value of a chosen dimension. Instead of one bar chart of amount by stage, you get a bar chart of amount by stage for each region, laid out side by side.
This is the right tool when you want to compare patterns, not totals. A stacked bar tells you which region is biggest; a trellis tells you whether each region's stage distribution looks the same shape. Those are different insights, and trellis is far better at the second.
You can trellis by row, by column, or both — a two-dimensional grid — though beyond a handful of values per axis it stops being readable.
4. Undo, redo, and history
Explorer mode is meant to be poked at, and it protects you accordingly.
Undo
Steps back one action. Every click — measure change, grouping, filter — is a step, so you can retrace freely.
Redo
Steps forward again if you undid one too many. Nothing you explore is destructive.
History
A panel of every state this exploration has been in. Click any entry to jump straight back to it.
The history panel is the underrated one. When you've clicked your way down a rabbit hole and the chart from four steps ago was the good one, you don't undo four times and hope — you open history and jump directly to it. It also doubles as a record of your reasoning, which is handy when someone asks how you got to a number.
5. Query limit and sorting
By default a lens returns every group. Open the limit control (often shown as a row/record count setting) and set it to 10, and you get ten groups back.
Which ten depends entirely on your sort. This is the pairing that trips people up:
- Sort the measure descending + limit 10 → your top 10.
- Sort the measure ascending + limit 10 → your bottom 10.
- No sort + limit 10 → ten arbitrary groups, which is almost never what you wanted.
So "top 10 accounts by revenue" is not one setting, it's two: sort descending, then limit. Get in the habit of setting them together.
6. Filters
Filters narrow the rows the query runs over, and how they behave depends on the field type.
Measure filters
A numeric range — amount greater than 50,000, quantity between 1 and 10. Drag the slider or type exact bounds.
Date filters
Absolute ranges (a fixed start and end) or relative ones — current quarter, last 12 months — which stay correct as time passes.
Dimension filters
Pick values from a searchable list. Include or exclude, one value or many.
Relative date filters are the ones to reach for. A lens filtered to "1 Jan – 31 Mar" is wrong on 1 April. A lens filtered to "current quarter" is right forever, which matters enormously once the lens becomes a dashboard chart someone checks every Monday.
Filters stack, and they're all visible in the filter bar at the top of the lens, so you can always see exactly what's been excluded. If a number looks surprisingly small, that bar is the first place to look.
Putting it together
A realistic exploration takes about thirty seconds: sum of Amount, grouped by Account Name, filtered to closed-won in the current fiscal year, sorted descending, limited to 10. That's your top ten customers this year — and every one of those clicks wrote a line of SAQL you can go and read.
Which is exactly where we're headed. Next: the three modes a lens can display its results in — chart, table, and raw SAQL.
Discussion
No comments yet — be the first to start the discussion.