Video analytics

Video analytics explained: what is video analytics?

Video analytics is the measurement layer around a video: who opened it, how far they watched, what they clicked, and what happened next. In a dealership it covers two different things at once, the published marketing video on your site and the one-to-one videos your reps and advisors send to individual customers.

Those two use the same words for very different numbers, which is where most reporting arguments start.

What does video analytics actually measure?

Almost every platform reports some version of the same six things. The labels change, the underlying events do not.

• Delivery and open. The message arrived and the customer opened the page it points at.
• Play rate. The share of people who landed on the page and pressed play. On a page that autoplays, this number is close to meaningless.
• Watch time and completion. How many seconds were watched, and what percentage of the runtime that represents.
• Drop-off curve. Where viewers stop. On a two minute walkaround this is the most useful single chart you have.
• Clicks. Which buttons around the video were pressed. See video call to action.
• Response. Whether the customer replied, booked, approved, or called back.

How is video analytics different from web analytics?

Web analytics counts anonymous sessions in aggregate. A personalized video page has an audience of one, and you know exactly who it is.

That changes what the data is for. A 40 percent completion rate across 10,000 anonymous views is a content signal, useful for editing. A single named customer who opened your video three times and watched the last 30 seconds twice is a sales signal, useful this afternoon. Same metric, completely different action.

It also means small numbers. A rep who sent nine videos last week has nine data points, and one customer who left the tab open all day can move their average watch time by a third. Read individual behaviour at the rep level and only trust percentages at the store or group level.

Video analytics explained: a worked example

An aftersales manager pulls last week for a team of six advisors. They sent 84 inspection videos. 71 were opened, 66 played, average completion was 78 percent, and 41 customers pressed an approval or reply button.

Read in order: opens are healthy, so the messages are landing on the right channel. Completion is high, so the videos are the right length. The gap sits between watching and acting, at 41 out of 66. That points at the ask on the page, not at the camera work. Now compare it with the send rate and the video response rate to see whether the problem is the ask or simply that only six advisors are sending at all.

Which numbers a manager should read weekly

Most dashboards give you 20 metrics and no order of operations. Three, in sequence, cover almost everything:

• Are we sending? Volume per rep, per week. Nothing else matters if this is near zero.
• Are they watching? Open and completion. If opens are low the channel or the subject line is wrong, not the video.
• Are they acting? Clicks and replies. If watching is strong and acting is weak, fix the buttons and the ask.

Everything else, drop-off curves, device breakdowns, heatmaps, is diagnostic detail you go looking for once one of those three is off.

Where video analytics gets misread

• Counting a play as a view. Some platforms fire a view on page load, others after three seconds of playback. Two tools can report the same week 30 percent apart and both be correct.
• Comparing completion across lengths. Eighty percent of a 25 second video is 20 seconds of attention. Eighty percent of a four minute video is a customer who is genuinely deciding. Never rank reps on completion alone.
• Blending sales and service. A follow-up video to a cold internet lead and an inspection video to a customer whose car is on the ramp have nothing in common. Split the report by department.
• Treating an open notification as intent. It tells you the customer is available right now, which is a good reason to call. It is not a buying signal on its own.
• Averaging across a team with uneven adoption. Two heavy senders and four non-senders produce a team average that describes nobody. Look at the distribution, not the mean.
• Reporting monthly. The whole point of this data is that it is same-day. A month-end summary arrives long after the customer decided.

Turning the numbers into coaching

The useful move is to pair one behavioural metric with one outcome metric and look at them together per rep. Videos sent against reply rate. Average length against completion. Buttons offered against buttons pressed.

Patterns show up quickly. The advisor whose videos get watched to the end but never actioned is describing the problem well and asking for nothing. The rep with high send volume and low opens is probably emailing customers who only ever answer a text. Both are fixable in a ten minute conversation, and neither is visible in a single headline percentage. See video engagement for the wider view of watching behaviour.