The Board With One Filter

Five schemas in this record carry one facet: of two, three, four, five or six values.

They sit at the top of the range on the share of questions that reach a single cell, at the bottom on almost everything else, and they are worth a page because the simplest case is where the structure is easiest to see entire.

Readers interested in the reporting side can review how reports are structured and note which controls are available to employees and administrators. For a separate technical reference, see Metabase.

All five, completely

Five rows, and the whole of what the record can say about one-facet boards. Every figure is a small whole number and the share is the facet size over the next number up.

What a reader of such a board can do

Set the filter, or not.

Setting it returns one cell — one kind of listing. Not setting it returns everything. There is no middle position, because a middle position requires a second filter to be partly set and there is no second filter.

So the grain count is two for all five: results come in exactly two sizes, one and the whole board. That is the floor of the record, and the five one-facet schemas are the only ones that reach it.

The highest share of pinning questions in the record belongs to boards that offer a reader two coarsenesses of answer and nothing between.

Why the high share is not a recommendation

Because the share is a ratio, and both its parts are tiny.

Six sevenths sounds excellent. It describes a board that tells six kinds of listing apart. If the field has more than six kinds — and fields usually do — then the board merges everything else, and the reader's high chance of pinning a cell is a high chance of pinning a very coarse one.

This is the clearest case in the record of a figure being accurate and useless on its own. The share is exactly six sevenths. What it means depends entirely on the cell count sitting next to it, which is six.

What they cannot do

Collide. Each of the five has a cell count no other schema in the record reaches with that question count — but more than that, the cell counts two, three, four and five are reached by no other schema at all, since every multi-facet schema has at least four cells and most have far more.

The exception is six: six cells is also reached by nothing else, because the only other way to make six from facet sizes would be two times three, which is two facets, giving twelve questions rather than seven. Different pair, different schema.

So the five one-facet boards are each identified by their cell count alone. They are the only schemas in the record of which that is true without qualification, and they are identified by the smallest numbers in it.

The two-facet step

Adding a second facet to any of the five changes the character completely.

A facet of six alone: two reachable result sizes, share six sevenths. Facets of six and four: four reachable sizes, share twenty-four thirty-fifths, and twenty-four cells instead of six.

One control added, and the board goes from a two-position switch to something with intermediate states. That is the largest qualitative change anywhere in the record, and it happens at the step from one filter to two.

Everything after it is more of the same: more resolution, more questions, lower share, denser ladder. The step from one to two is the one that introduces the structure at all.

The lesson the small case teaches

A board is not described by whether its figures are high or low. It is described by the figures.

One facet of six: share nearly eighty-six per cent, resolution six. Six facets of six: share about forty per cent, resolution forty-six thousand six hundred and fifty-six.

On the share, the first board wins by a wide margin. On anything a reader cares about, it is not in the same category. Both facts come from the same five numbers and neither is wrong — they answer different questions, and quoting one without the other is how the record gets misused.

Near this one

• What The Share Is Not
  A proportion of the questions a board permits, not of the searches anybody performs. Four things the figure is
• The Questions That Return Groups
  Most of what a board permits leaves some filter alone. That is where nearly all searching happens, in a few fi
• The Questions That Reach One Answer
  Only queries that set every filter reach a single cell, and there are as many of those as there are cells. The
• Sixty-Four In Two Hundred And Twenty-Five
  The median schema reaches a single listing with about twenty-eight per cent of its questions. Three schemas si