Apple TV’s Slow Horses is not an equestrian comedy but rather a genre-bending series about a ragtag group of post-modern British spies. Here is the description on the streaming service:

This darkly funny espionage drama follows a team of British intelligence agents who serve in a dumping ground department of MI5 due to their career-ending mistakes. Led by their brilliant but irascible leader, the notorious Jackson Lamb, they navigate the espionage world’s smoke and mirrors to defend England from sinister forces.”

Slow Horses lives up to its name almost too well in the first couple of episodes, and there's a point where many viewers might well be tempted to move on. But patience pays off amply over the next four episodes through the end of Season 1.

Much of Lumiere's content testing necessarily deals with reactions to the surface of content – to what's audibly and visibly happening in a scene or across an episode. However, underlying the viewer reactions are all sorts of invisible narrative structures, like sliding Tectonic plates – character development, mystery and suspense, questions and expectations about where the story might be headed, elisions and so on. Slow Horses succeeds by using narrative density to build deeper connections with in the plot and, ultimately, with the audience. Let's look at a high level how this happens, without diving into details of the show.

What in fact is narrative density? For a start, consider network science. We’re all familiar with the concept now used in domains ranging from the social graph to infectious disease models to route planning: nodes are points (often people), and links connect them. We intuitively understand that some networks exhibit a greater number of connections (total and/or on average) between nodes than others, as well as different types of connections.

Slow Horses develops network density in three interrelated ways:

1. Adding relationships between the characters as the story moves forward (that is, new links)

2. Uncovering pre-existing, unexpected relationships in each episode (revealing hidden links)

3. Creating depth within the relationships (thickening the links)

In the first case, the pleasure for viewers is to see the story progress in ways that we generally expect from well-written series — density and complexity in narrative are inherently compelling. They heighten our cognitive activity around the storyline and offer additional doors for character empathy and identification. In the second case, the pleasure is slightly different — that of surprise and unexpected trickery at the narrative level (all the better in a spy story) and of the intellectual satisfaction we receive when intriguing loops and back connections are formed. Similar in some ways to the usually pleasant and interesting personal experience of learning that relatives or friends have secondary connections with other relatives and friends.

Finally in the third case, one with which we’re all familiar — Slow Horses, in elegantly understated British ways — takes time to allow the relationships of which we were originally aware, in plain sight all along, to gradually evolve and thus give us the pleasure not only of subtle and unexpected character arcs (node changes) but also relationship arcs (link changes).

The overall effect of these three interlocking density generators is what we might refer to in colloquial terms as ‘the plot thickens’. That doesn’t quite do justice to the concept but it’s not too far off. Slow Horses, now has an audience waiting to see not only what will happen next season but just as importantly what might be underneath the surface that will provide complex and intriguing new connections.

As for the semi-obvious and relevant question — how to translate all this into useful metrics for content creators — connecting in-the-moment viewer reactions with underlying story analytics is Lumiere’s work in progress.