What to fix first, decided by the people who answered.
Most engagement surveys ask everyone to rate every question, then work out what matters from how the
ratings move together. Ranx asks each person first: they put the categories in order of importance, and
that order weights their own answers. What comes out is one ordered list of what to fix, built from what
your people said matters to them.
Each person puts the categories in order of importance to them. The list starts in a different random order for each person, and ties are allowed.
Then they answer the questions, category by category in their own order, with the questions shuffled inside each one.
Their order becomes their weights: first place counts most, and each person's weights add up to 100%.
Each category's priority is how far scores fall short of 100, weighted by how much each person said it matters.
From a ranking to weights
First place counts most. Every person's weights add up to 100%.
Weights for a ranking of 4 categories
1st52.1%
2nd27.1%
3rd14.6%
4th6.3%
Rank-order centroid weights (Barron & Barrett, 1996). Tied categories share the average weight of the places they hold.
Two people give Career Growth 40 out of 100
Ana ranked it 1st of 460 points short × 52.1% weight31.3
Ben ranked it 4th of 460 points short × 6.3% weight3.8
An average counts both scores of 40 the same. In Ranx, Ana's counts about 8 times
as much, because Ana said it matters most. An illustration: two people, 4 categories.
Why it's clearer
One list, in order, that your people set.
Everything can't come first.
A rating scale lets someone call every area important. A ranking asks them to choose. Ties are allowed, and tied categories share their weight.
A low score counts by how much it matters.
A low score on something a person ranked first moves the list far more than the same score on something they ranked last.
Disagreement shows.
When scores on a category are spread widely, Ranx lowers its priority and shows the spread, so a split doesn't hide inside an average.
Every team gets its own list.
The weights come from each person, so the same math gives each team, department or manager its own priorities, for any group with enough responses to meet your minimum group size, which starts at 5.
A second opinion
The ranking leads. The statistics check it.
Drivers sets each priority beside two independent reads: whether it moves together with outcome questions,
such as job satisfaction or plans to stay, tested once at least 30 people have answered both; and AI's
read of the results. Where they agree, act. Where a priority ranks high but the link isn't confirmed yet,
look further. Neither one changes the ranking.
