What remedial evaluation is
Remedial evaluation takes place after an exhibit has opened, in the early weeks or at any later point, with the purpose of finding problems that can be fixed in place. It sits between testing rough versions before the build and the formal judging of the finished exhibit against its goals. The central question is practical: what is not working well for real visitors, and what is the smallest change that would help?
Even carefully tested exhibits meet conditions in the gallery that no test fully reproduced. Crowds, noise, daylight, school groups and visitors who behave in unexpected ways all show up only once the exhibit is open. Remedial work accepts that and builds in a chance to adjust.
How it differs from other stages
It helps to keep the stages distinct, because each has a different aim:
- Early audience research, before design, asks what people know and want.
- Testing rough versions asks whether a design idea works, while it is still cheap to change.
- Remedial evaluation asks what to fix in a live exhibit.
- Summative evaluation asks whether the exhibit met its goals overall.
Remedial work is narrower and more immediate than a full assessment. It usually focuses on specific problems and produces specific changes. It does not try to settle whether the project as a whole was a success.
Finding the problems
The same methods used elsewhere in evaluation apply, used lightly:
Watching. A few hours of visitor observation at the exhibit often reveals obvious problems: a control people miss, a label that is blocked, a queue forming at a single screen.
Staff and volunteer reports. Gallery staff see the exhibit all day. They hear questions, notice repeated complaints and know when something stops working. A simple way for them to report issues, such as a shared log or a short form, is among the most valuable and least expensive sources. Staff training that includes what to watch for makes this more reliable.
Usage logs. Data from the system can point to sticking points, such as many sessions that end on the same screen. Small numbers are easy to over-read, so treat exhibit analytics with care.
Short conversations. A few quick questions at the exit (“Was there anything confusing?”) can show what people could not work out.
Comment cards or feedback forms. Useful for specific complaints, but the people who write comments are not typical of all visitors.
Technical checks. Not all problems are about design. A miscalibrated screen, a speaker that is too quiet, or a restart that fails can look like poor design to a visitor. Touch calibration and sound bleed are common examples of faults that are physical rather than conceptual.
Sorting what you find
Not everything found needs fixing. A rough way to sort issues:
- Safety, access or serious failure. Anything that prevents visitors using the exhibit, puts them at risk, or excludes a group of people. Fix soon.
- Barriers to the main goal. Problems that stop visitors reaching the central experience.
- Friction. Things that annoy or slow people but do not stop them.
- Wishes. Ideas that would be nice but are not fixing a problem.
Consider two things for each: how many people it affects, and how hard it is to fix. A change that helps many visitors and takes an hour is an easy win. A change that affects a few people and needs a software rebuild may have to wait, or be handled differently, such as with a staff note.
Small samples can mislead. A single complaint may or may not represent a wider problem. Look for the same problem arising from more than one source, such as an observation and a staff comment.
Typical targeted fixes
Many remedial changes are small and inexpensive:
- Rewording a label or on-screen instruction, perhaps using plain language principles.
- Adding a simple sign or arrow to show where to stand or touch.
- Enlarging or spacing buttons so touch targets are easier to hit.
- Adjusting screen brightness, angle or lighting to cut glare.
- Changing the volume or moving a speaker to reduce sound bleed.
- Shortening an idle timeout or improving the attract screen so people understand what to do.
- Adjusting a screen height so it is easier to reach, for example for wheelchair users.
- Fixing a broken link, a typo or a mistaken caption.
- Cleaning, repairing or replacing worn hardware.
Some fixes are larger, such as restructuring a menu, but the same principle applies: change the least amount necessary to address the problem.
Changing one thing at a time
When possible, make one change and then check its effect before making another. If you change five things at once, you cannot tell which helped or whether one made things worse. This is not always possible, particularly when a screen is failing, but it is a good habit for design changes.
Before making a change to a running exhibit, consider:
- Does the change risk breaking something that works? Test it away from the public first, or in a quiet period.
- Can it be reversed? Keep a copy of the earlier version.
- Does it affect other material, such as translations, audio descriptions or captions that must be updated too?
- Who needs to be told, from front-of-house staff to the people maintaining the exhibit?
For software changes, a short test with a few visitors can catch a mistake before it is rolled out fully.
Checking whether it worked
After a fix, look again with the same method that found the problem. If you noticed visitors missing a button, return to watch whether they now find it. Compare roughly equivalent conditions, such as similar crowd levels and times of day, and keep in mind that a small number of observations is only indicative.
If the problem remains, reconsider your explanation. The first fix is sometimes aimed at the wrong cause.
Recording what you did
Keep a change log: what the problem was, how it was found, what was changed, the date, who did it, and what happened afterward. This helps future staff, supports the formal review of the exhibit and prevents the same problem being reintroduced. It also fits with project handover documentation and the maintenance plan, which should say who is authorized to make changes and how they are approved.
Budget and planning
Remedial work is easier when it is planned. Include time and a modest amount of money for post-opening fixes in the exhibit budget, and keep the team available for at least the first few weeks. If the contract ends at opening, with no one able to change the software, even simple fixes can become impossible. Asking about this when choosing outside suppliers can save trouble later.
Not every issue is fixable, and some are best accepted. Part of good practice is recording what was left alone and why, so that findings carry into the next project.