How a museum trip can generate insights for online retail
New research by Ali Aouad and colleagues looks at how algorithms can enhance user experience in the customer journey
- Museums and galleries increasingly feature interactive guides as gatekeepers to the visitor experience
- But providing guides is an imprecise science and does not guarantee optimum visitor experience
- Novel approach uses objective data and analytics to create tools to personalise the digital experience
- A by-product of the study is the two-way flow of learning that can emerge, with insights for business and retail
Going to a museum or a gallery to see works of art is a common leisure activity. And it’s supposed to be enjoyable, educational and enriching. But how can you be sure you’ve seen the things that will really resonate with you when you visit London’s Tate Modern, say? How do you know if, in the throng of visitors, you haven’t missed some undiscovered masterpiece that you may have appreciated more? Curators and, increasingly, interactive guides function as gatekeepers to your experience, helping you navigate collections and select the best works for you. Even so, guides and layout can be hit and miss – an imprecise science.
That is, until now. A new body of work by London Business School’s Ali Aouad, in collaboration with IESE’s PhD student Abhishek Deshmane and Professor Victor Martínez de Albéniz, is shedding new light on how to structure visitor experience, and their findings are of relevance not only for the arts, but also for experience providers in commercial sectors.
Dr Aouad investigates how algorithms can be used in the cultural setting to understand, predict and ultimately enhance behaviours and experiences. The approach is novel both in its methodology and in its collaboration with a major cultural institution.
“Algorithms have been used to facilitate various operational decisions from retail platforms to ride-sharing by leveraging data on user preference,” says Dr Aouad. “But, so far, no one has looked at how to use algorithms in curatorial practice, so we’re really looking at a new frontier in this field.”
The use of objective data and analytics to drive engagement and diversify user experience is an unexplored area, and Dr Aouad and his colleagues believe the potential is enormous. In the museum or gallery setting, algorithms may have the ability to personalise user experience, giving organisations access to targeted strategies for managing congestion and optimising engagement – and thereby converting them from “mausoleums to living museums”.
“The visitor population, like any population, is highly diverse”, Dr Aouad points out. “Some people may visit the Tate or the Van Gogh Museum just to see the highlights. Others may want to venture further, while some might already know the collection and be looking for something specific. It’s hard to meet everyone’s goals. A multitude of pathways through the collection needs to be possible, but we are only scratching the surface of how curation and technology can complement each other.”
Discover fresh perspectives and research insights from LBS
'When visitors face an optimised layout, the simulations suggest they’ll be 5% more engaged'
Van Gogh Museum collaboration
The researchers wanted to see whether analytics could provide insights into decisions such as layout and help create tools to personalise the digital experience provided by multimedia applications such that individual users would remain engaged over time, relating more deeply to the content and to what Dr Aouad calls “the narrative of the experience”.
Museums now have access to large-scale, granular data on visitor experience, providing insights on how people interact with pieces and how they use multimedia, for instance.
Using similar types of data, Dr Aouad and his colleagues built an algorithm to predict visitor journeys at Amsterdam’s Van Gogh Museum – and made some interesting discoveries along the way. “The data shows that the distribution of artworks in space has a significant impact and users are very much guided by proximity – they will move from one piece to something nearby,” Dr Aouad explains. “When you dig a little deeper, you find other interesting patterns. Visually, we seem to like diversity. People are more likely to move from a large canvas to something smaller and vice versa. One plausible explanation is that individuals could be variety-seeking.”
Building the algorithm is also yielding prescriptive findings. The researchers are now using algorithms to optimise the physical layout of the museum. They have simulated hundreds of millions of trajectories through the space and assess which types of layout are most likely to maximise engagement with different types of visitor profiles.
“Typically, museums will cluster their most notorious pieces on one floor or room, so we will be monitoring the data to see how different permutations will encourage users to explore different spaces and whether this will impact engagement as we predict,” Dr Aouad says.
When visitors face an optimised layout, the simulations suggest that individuals would be 5% more engaged and the crowds less concentrated. Following these findings, the Van Gogh Museum has experimented with new designs for the multimedia guide. Like A/B testing in online retail, experimentation is a powerful mechanism for museums to test and validate new ideas.
Leveraging insights for business
For the arts sector, dependent on the frequency and volume of paying visitors, these initial findings are of enormous relevance. Maintaining interest and engagement is critical to visitor satisfaction and a steady flow of income in the long run. But, Dr Aouad says, there is food for thought here for other industries, too.
Retailers may want to think about the ways that the online setting mimics the layout of museums and galleries, and reframe the user experience as a journey over time instead of a sequence of incremental actions mediated by clicks.
“Right now, retailers are really good at using algorithms trying to predict behaviour and match product recommendation, but it’s done without looking at repeated interactions with the platform,” Dr Aouad explains. “Online platforms have tended to discard the idea of a journey solution for the consumer because it’s hard to do. But our study suggests that we can actually model sequences of choices and predict outcomes.”
“Algorithms have historically been quite limited in predicting whole customer journeys in the online retail setting. They’re good at forecasting a consumer’s interest in something specific right now, but less adept at prognosticating what will be interesting to that individual further down the road.”
Frontiers of AI
The study is pushing at the frontiers of what AI can do in terms of shifting our understanding of isolated purchase events to the entire customer experience.
“One of the interesting by-products of the study is the two-way flow of learning that may emerge from our work,” Dr Aouad reveals. “Museums have already started the heavy lifting in terms of adopting customer-centric metrics in their approach to curation. But their core expertise is in creating journeys and narratives. So, there are emerging learnings from the cultural sector for business and retail.”
He concludes, “As we learn more about AI and virtual experiences, there’s the potential that understanding will flow in both directions: from business to culture and back.”
Ali Aouad is Assistant Professor of Management Science and Operations at London Business School.
