Most software is designed for someone paying attention. They opened it to do a thing, they will do the thing, and they will leave.
Media products are not used that way. They are used in the evening, on a sofa, with a phone in the other hand, by someone who is half watching and does not want to think about the interface at all.
Friction compounds
An extra tap to resume is nothing once. Encountered every evening for a year it becomes the reason someone drifts to a competitor without ever articulating why.
This changes the economics of small decisions. A control that tests as merely acceptable in a usability session — where the participant is concentrating, in a quiet room, being observed — may be genuinely bad in the context it will actually be used in. First-use testing systematically under-weights the things that matter most here.
What long sessions demand
Resume that is right. Getting back to the right point, on the right device, without asking. This is the single most-used feature in most media products and the one most often subtly wrong.
Controls that survive imprecision. Large targets, forgiving hit areas, tolerance for a thumb rather than a cursor — people are not aiming carefully.
Nothing that moves. Content shifting position between sessions destroys the muscle memory habitual users build, which is the thing keeping them fast. A personalised rail that reorders every visit feels responsive to the team and disorienting to the user.
Undo over confirmation. Confirmation dialogs interrupt every user to protect against a rare mistake. A reversible action is better in almost every case here.
Quiet failure handling. Buffering, a dropped connection, an expired session — recover without ceremony. An error message on a television is a dead end; there is no support article to read and nobody wants to fetch a laptop.
The remote control problem
Anything on a television is operated with four arrows and a select button, from three metres away, sometimes in the dark.
That constraint is severe and gets designed around late, usually by adapting a mobile layout. Navigation order becomes the primary design decision rather than an accessibility afterthought — and everything that makes it work for a remote also makes it work for keyboard-only users, which is a rare case of the accessible design simply being the better one.
Measure the right sessions
Aggregate engagement hides this. The numbers to watch are the habitual users: whether resume rates are falling, where returning viewers abandon, how long from opening to playing something.
That last one is the number worth defending. Everything between opening the application and content starting is overhead, and in a habitual product it is paid every single day.
Where this comes from
We have designed and built interfaces at this scale, including work on STV Player — products where the same small friction is encountered by a large audience, repeatedly, and where the cost of getting a control slightly wrong is paid every evening rather than once.
Working on a product people use for hours? Get in touch, or read about our entertainment and media work.