← All Issues
Behind the Label

When the Museum Hands the Reins to Visitors

Osama Elemary · October 24, 2026

In 2009, the Cleveland Museum of Art asked a dangerous question.

Is anyone reading what we write?

The museum commissioned a full visitor study in its reinstalled galleries of European and American art. Not a satisfaction survey. An observation of behavior.

The answer was uncomfortable.

Visitors were browsers. They walked past the introductory texts at the entrance of each gallery, the paragraphs curators and educators had labored over. They stopped only in front of the works that pulled them in, and read a label only when a specific work made them want to know more.

The museum did not respond by writing more. It responded by redesigning interpretation around the browser: no preferred path, no front-loaded narrative, interpretation offered at the moment of curiosity. When Gallery One opened in 2012, it was built to honor browsing behavior, documented in the museum’s own evaluation.

I worked on an exhibition documenting the legacy of a Saudi monarch.

Every photograph carried collective memory. Every image held a story already alive in the minds of those who came to see it. In the early version, we wrote beneath each photograph: the occasion, the context, the moment.

Then we stopped.

In coordination with the commissioning authority, we replaced individual captions with broad thematic headings: the human dimension, the political, the social. And we left the photographs without explanation.

What followed was not planned.

Visitors who had lived through those years stopped in front of the images and began to speak, to the strangers beside them, pointing, remembering. Younger visitors listened. The exhibition had turned some of its audience into guides.

That did not come from our design.

It came from our restraint.

There is a persistent assumption in interpretive work: that the visitor needs us to tell them what they are seeing.

But visitors arrive carrying their own histories, their own questions, their own silences. When we fill every surface with our words, we do not enrich their experience. We replace it.

A label that explains everything gives the visitor an answer to a question they have not yet asked.

And a label offered before the question is formed does not teach. It closes.

Cleveland’s lesson was not that text is the enemy.

It was that text offered before curiosity is noise.

The museum kept its labels, and rebuilt everything around the moment a visitor leans in.

The next time you review interpretive text you have written, ask one question: does this answer a question the visitor has already asked?

If the answer is no, cut it.

Not out of laziness. Out of respect for how visitors actually behave.

The English framework and downloadable professional guide: link on the website.

Next issue: Intentional Absence — when an architect builds a wall that is never opened, never used, and never explained.

Osama Elemary | Museum & Heritage Strategy Consultant | Riyadh, Calgary

Questions about this issue

Is interpretive restraint only relevant for art museums?
No. The principle applies wherever visitors arrive with prior knowledge or emotional connection to the subject. History museums, heritage sites, and memorial spaces often benefit most from restraint, precisely because the visitor brings more to the encounter than any label can match.
How do you decide which labels to keep and which to cut?
One test: read the label, then look at the object. If the label tells you something the object cannot tell you on its own, keep it. If it tells you what to feel, or summarizes what is already visible, cut it. The label earns its place by answering a question the visitor is already asking, not by anticipating every possible question in advance.
What is the risk of removing too much interpretation?
Visitors who lack prior context can feel excluded or disoriented. Restraint is not the same as abandonment. The Cleveland model kept labels — it restructured when and where they appeared. The goal is interpretation on demand, not interpretation by default.