End Screens and Cards: The Twenty Seconds Most Creators Waste

The last twenty seconds of a video are the most valuable and most wasted real estate on YouTube. The viewer has already decided they like you — they watched to the end — and they are about to make a decision about what happens next. Most creators fill that moment with a logo animation and a wave.

This guide covers what end screens and cards actually do, how to use them without damaging retention, and why the most common configuration is close to the worst one.

Cards and end screens are different tools

The essential difference: a card is an offer made during the journey, and an end screen is a signpost at the destination. Using either for the other's job is where most of the value is lost.

The retention cost nobody accounts for

End screens require the last five to twenty seconds of your video, and here is the part creators miss: those seconds count in your retention graph.

If your video ends properly at 9:40 and you add twenty seconds of end screen over a logo loop, most viewers leave during that outro. Your average view duration falls, your retention graph shows a cliff, and that weaker signal affects how the video is distributed.

The fix is not to skip end screens. It is to make the final seconds worth watching rather than dead air. Keep talking over the end screen. Deliver a genuine final thought, a summary, a specific reason to watch the next video. The moment a viewer perceives the video as over, they leave — so do not signal that it is over while you still need twenty seconds.

Choosing what to put there

You have limited slots and they are not equally valuable. In rough order of usefulness:

1. A specific next video. Not "best for viewer" — choose deliberately. The right choice is the video that continues the thought this one started, because that is what the viewer is currently interested in.
2. A playlist, if the video belongs to a series. This is frequently better than a single video, because it chains rather than ending after one more.
3. Subscribe. Worth one slot, though its practical effect is smaller than creators assume. People who intend to subscribe generally do so without prompting.
4. An external link, if eligible and if you genuinely have somewhere valuable to send people.

What to avoid: filling every available slot. Four options is a decision, and a decision at the end of a video frequently resolves as "none of these". Two well-chosen options outperform four generic ones consistently.

Also avoid the automatic "most recent upload" setting as a default. Your most recent video is rarely the most relevant follow-on, and relevance is what drives the click.

Designing the ending around it

The end screen is a design constraint on your edit, and it works best when planned rather than bolted on:

• Leave visual space. If you know panels will occupy the right side, do not put important content there in the final seconds. Frame yourself to one side.
• Script the handoff. "If you want the version of this for beginners, that is on the left" outperforms silence by a wide margin. Point at it, verbally and physically.
• Give a reason, not an instruction. "Watch this next" is weak. "This is the part that trips most people up, and I cover it there" is a reason.
• Keep it short. Five to ten seconds is usually enough. Twenty seconds of end screen is twenty seconds of retention decline for most channels.
• Stay consistent. Same position, same structure every video. Returning viewers learn where to look.

Using cards well

Cards are subtle and easy to misuse. The failure mode is scattering them through a video on the theory that more chances to click is better. In practice an irrelevant card is ignored, and a card that pulls someone away mid-video costs you the rest of the watch time.

Cards work when they answer a question the viewer is having right now. If you reference a technique you explained in another video, a card at that exact moment is genuinely helpful. If you mention a tool you reviewed, same. The test is whether a viewer at that second might be thinking "wait, what is that?" — if so, the card serves them; if not, it is clutter.

Two or three well-placed cards per video is plenty. Place them at moments of genuine relevance rather than at regular intervals, and avoid the first thirty seconds, where you are still establishing that the video is worth watching.

Measuring whether it worked

YouTube Studio reports end screen and card performance — impressions, clicks and click rate — and the numbers are worth checking, because assumptions here are frequently wrong.

What to look for. Click rate by element: if one video consistently outperforms as a next step, use it more widely. Whether cards are clicked at all: many are not, and an unclicked card is a signal that it was not relevant where you placed it. Average view duration against your end screen length: if retention collapses exactly where your outro begins, the outro is the problem.

The comparison worth making is between videos that send viewers onward and videos that end flat. Channels that treat the ending as a handoff rather than a farewell accumulate substantially longer sessions, and session length is one of the things the recommendation system most consistently rewards.

Frequently asked questions