Case study: testing engagement with online collections

This case study, from our Let’s Get Real action research program, looked at how people engage with online collections and blogs, and used these platforms to gauge reactions to themes around death to inform a planned exhibition on the subject.

About the participant:

Name: Fay Curtis

Organisation: Bristol Culture (Bristol Museums, Galleries & Archives)

Jump to:

What was the research question behind your experiment? | Why was this important to your organisation? | What did you do to implement this? | What happened? | What were the personal challenges you faced when carrying out this experiment? | What did YOU learn? | What did YOUR ORGANISATION learn? | What next?

What was the research question behind your experiment?

I wanted to see how people engaged with our online collection search via social media, and whether this differed from how people engaged with our blog. I also wanted to see how people engaged with content from an exhibition which opened at Bristol Museum & Art Gallery on 24 October 2015 called death: the human experience. This is a tricky subject and I wanted to introduce some of the themes in the exhibition to people before it opened, and gauge their reaction so that we could adjust content accordingly once it opened.

Why was this important to your organisation?

Two main reasons – firstly, the death exhibition is our flagship exhibition this year, and being on the subject it is, we wanted to measure audience reaction prior to opening to better prepare our communications strategy.

Secondly, we’re looking a lot at storytelling/narratives online at the moment and how we should be approaching this (and if there is a need for it). I wanted to use the Let’s Get Real program to inform Phase 3 or 4 of our website development, which we’re looking to base around digital stories and our online collections.

What did you do to implement this?

I worked with one of the lead curators, Amber, who set up a ‘Symbols of death’ narrative on our online collection search.

This worked well with what I had in mind: short pieces of content to share daily at different times of the day, alternating between objects/narratives when posting to see if there was a difference in engagement.

There were just over 20 symbols, so I chose 14 of these to highlight on the blog and posted one a day on social media in the two weeks in the run-up to the opening of the exhibition. I wanted to see how people (referrals from social media) navigated through our object/narrative pages on the online collection search – if they went onto further pages, if the hierarchy of information made a difference etc.

I set up some tracking on Google Analytics – I originally wanted to try Event Tracking but for a couple of reasons it wasn’t quite feasible to do this in time for the experiment on our online collections search. So, I set up campaign links, content grouping and segments on GA. Alongside this I created a simple ‘listicle’-style blog post to compare how people engaged with this style of content.

Lastly, and on a slightly separate note, I pulled together some data (mainly from Twitter) around people responding to our death content/talking about the death exhibition to run it through a sentiment analysis tool – Alchemy.

What happened?

Just before I started the experiment, we created a Facebook event for the exhibition, which gained huge traction very quickly. It became pretty clear to us that the exhibition theme wasn’t a put off! The event had more people attending it than actually like our Facebook page for Bristol Museum & Art Gallery(!) so I decided to post the content in the event instead of on the page. I posted the symbols on Facebook, Twitter and Instagram each day, alternating between linking to the ‘object record’ and the ‘object narrative’ on our online collection search, at different times of day.

What were the personal challenges you faced when carrying out this experiment?

The usual – time and trying to juggle multiple projects at once. We were going live with website Phase 2 just before the opening of the exhibition, and I would’ve liked to spend some more time focussing on our Let’s Get Real experiment in more depth.

What did YOU learn?

It was clear straight away that there were small changes we may be able to make which could make a difference to user experience on our online collections.

Across our collections, the field use isn’t consistent apart from a few mandatory fields, e.g. object number, which causes issues for user experience design.

This experiment helped me to really narrow down some specific content to test out of the hundreds of thousands of object records we have, showing that we may be able to change the order that the data fields display in to make it more user-friendly.

People didn’t navigate much to/from object records and object narratives – maybe this is okay and we don’t want to try to change this behaviour, but there could be ways for us to make it easier to navigate and test again to see if there is a need for this. So a change we made as a result of this is pulling through text from narratives onto object pages, instead of relying on people clicking on links to read narratives.

Overall, though, the blog post was a bigger driver to the symbols of death on the online collection search than the posts on social media, even though most visits to our blog come from social media anyway (meaning that people are going from social media > blog > online collections search). It seems that this format of content works better with social media audiences. It may be better to pull groups of collection items together on the blog as opposed to in a narrative on the online collection search, and I have done this since for the National Gallery’s #AngelTrail which proved really popular again.

I think this shows that we can learn from what other (probably non-cultural) organisations, with lots of resource, are doing. We don’t need to reinvent the wheel in terms of content – it seems obvious, but we just need to format it in a way that audiences are used to seeing (and will notice in the huge amount of content that gets thrown at them on a daily basis!).

I didn’t have much time to do a huge amount around sentiment analysis/Alchemy but putting the data through it initially showed a really positive response (even with terms normally seen as negative e.g. ‘death’) – this showed me we were on the right track and to carry on doing what we were doing once the exhibition opened.


That people ‘love’ death and that we maybe shouldn’t have been so worried about the audience response.

We’ve now made some changes to how the data for objects in the online collections search is displayed as a result of findings from the project. We prioritised the order of fields displayed to be more user friendly, for example replacing the object number with the object name as the title – on page titles, search results and thumbnail navigation.

It’s made us think about how we design and test methods for publishing content in order to successfully analyse it, for example having meaningful URLs for our online collections.

What next?

We’ve already worked on small improvements to the online collections search – order of lists/fields etc.

We’ll be working on Phase 4 of our website, part of which is around stories/narratives online and integrating online collections.

From my content experiment, I think we need to break this down into two sections:

  1. Stories/narrative content as the primary focus with object records integrated but secondary
  2. Object records as the primary focus with stories/narratives integrated but secondary

We’ll be testing what user needs are for both kinds of content – which we’re sure will bring up lots of lovely challenges around what digital assets we have, how we use these and how (or whether) the content we have in our physical spaces needs to change for digital platforms.

I also want to work more on the Alchemy sentiment analysis API to see if we can use this for analysing social media engagement in a simple and consistent way.