Cutting Room Floor
Everything the filters threw away, on reels anyone can audit and splice back.
The job
Shows a sceptic exactly what was left out of a headline number, why, and how often each exclusion rule gets it wrong, so the number is believed. Customer comments that were wrongly removed are put back in.
The moment
Everything the filters removed from a headline number is laid out like film reels on a cutting-room floor, one reel for each exclusion rule.
Sorting blind, an auditor pulls a clip from the spam reel and marks it to keep: it is a genuine bribery allegation in the customer's own words.
A search for similar clips finds 212 more. The owner previews ten and approves putting them back.
The spam rule's measured accuracy drops on screen from 97% to 89%. The headline number grows by 3%, labelled as including 212 reinstated comments.
What it does
- Rule a clip rightly or wrongly cut
- Splice one clip back
- Reinstate a similar set in bulk
- Send wrongly cut examples to the rule owner
- Open a reel
What you see
Film editor's cutting-room floor: reels of cut clips per exclusion rule, a blind sorting round and a splice-back room
What it moves
How accurate each exclusion rule is, in points with a range, and the number of genuine customer comments put back into headline figures each month.
Built for
- Marketing, brand & comms
- Compliance, complaints & legal
- Analysts & data
