A School of Continuous Improvement paper audits 14,206 supermarket shelf labels and follows 88 shoppers through the aisle, finding near-total compliance with a price-transparency code alongside comparison accuracy at split-unit shelves that does not separate from chance.
In-aisle eye tracking of 1,406 comparison episodes finds the unit-price field drawing most attention exactly where the two figures cannot be placed on one scale
Ninety-eight point six per cent of the shelf labels in this study carried a compliant unit price. Fifty-eight point seven per cent of the shelves they sat on offered any way to use it. Displayed, Not Comparable, published today by the School of Continuous Improvement, reports the distance between those two figures.
The team transcribed 14,206 shelf labels across twelve stores, grouped them into the 1,038 runs of competing lines a shopper chooses between, and coded each for whether its members resolve to a shared unit. Eighty-eight participants were then fitted with head-mounted trackers and sent down the aisle with one instruction: find the cheaper line per unit and say so.
A shopper who cannot compare two figures is not failing at arithmetic. They are being handed two different questions and asked for one answer, and every part of that arrangement is compliant.
— Dr Joost Nwosu, Postdoctoral Fellow, School of Continuous Improvement
Where a common unit was available, participants named the cheaper line 88.4 per cent of the time. Where it was not, they named it 51.7 per cent, an interval that includes the value expected of a coin. Attention ran the other way: the unit-price field drew a fixation in 34.1 per cent of episodes at the unusable shelves against 18.9 per cent at the usable ones, and held roughly twice as long. Dr Ingeborg Nwachukwu, Lecturer and Convenor of Capability Sprints, said that pattern “changed what we thought we were measuring”, the field being “not so much ignored as attempted”.
The work was supported under the Indicator Stewardship Seed Fund, and adds a familiar retail surface to the School’s interest in indicators that report their own condition rather than the thing they stand for.
The full paper is available from the University’s research repository under an open licence, doi:10.5555/slop.19q87i.