Volume as a score

A volume score measures a theme by how many responses it drives, not by how people rated it.

The score's value is the number of responses in the period, so 300 responses is a score of 300. It's a total, not an average and not a rate. This puts your themes on a scale you can act on directly.

What it's for

Some datasets are made of events that cost something. In a complaints inbox, a contact centre or a support queue, every response is a real thing that happened and had to be handled.

On that kind of dataset if account problems make up a fifth of your calls, halving them takes about a tenth of your calls away. An average score or a per-customer rate can't be read that way as it tells you how bad something was, not how much of it there was.

Setting up a volume score

1. Go to Manage Lenses > Configure Lens > Scores.
2. On the Edit score screen, under Score Range, select Volume. A volume score counts responses, so it doesn't need a score column.
3. Under Select & Configure Datasets, select each dataset you want to count and tick Count responses from this dataset. Datasets you've included are marked Included.
4. Enter a Score Name, then click Save.

Where you see it

Select the volume score from the Score dropdown in the filter row. Two views change their meaning when you do.

Written summaries of a change count in responses rather than points.

The line reads as counts rather than as a rating between 1 and 5.

Underneath it, the waterfall breaks one period's change into themes and the bars account for the difference. The numbers may not be whole numbers as we still proportion themes according to how many responses they occur in, and a single response can have multiple themes.

Reading the numbers

It grows with the period

A quarter has more responses in it than a month, so the score is larger but nothing has got worse. This is because the score is a total, and it's why the two periods you compare should be the same length. Score Change already compares like with like, so this is worth watching when you set a date range by hand.

Theme counts overlap

A response can carry several themes, and some carry none, so the themes' own counts don't add up to the total. The waterfall scales its bars so they do add up to the real change. That means a bar is a theme's share of the change, not its own count. The theme list beside it shows each theme's actual count.

Weighting does not apply

If you turn weighting on, a volume score doesn't move.

Weighting reshapes a sample so it matches a population. That makes sense for a share or an average, but a count of real events is neither. Turning 300 calls into 450 would report calls that never happened. Revenue at Risk is left out of weighting for the same reason — it's a total, not an average.

Everything else on the page still weights normally. Only the score itself is left alone.

Surveys with more than one comment question

Volume counts comments, not rows. If your survey asks two open questions and someone answers both, that response contributes two. So a volume score can be larger than the number of responses you uploaded, and that's correct — there really were that many pieces of feedback.

Lenses

Across the sources in a lens, volume adds up. Two sources with 120 and 80 responses give a lens score of 200, not an average of 100. Themes add up the same way.

Things that can look odd

• The score is bigger than your row count. Your survey has more than one comment question — see above.
• A waterfall bar isn't the theme's own change. Bars are scaled to add up to the total change. The theme list has the raw counts.

Related articles

• Editing your lens — where the lens configuration lives
• Score Change
• Example scores or metrics
• Impact formulas