Monday morning. The campaign ran all week. You open the dashboard and you see something like this.
Campaign overview — Week 1
10,247
Views
38%
Avg completion
47
Sales
0.46%
View-to-sale CVR
Is 0.46% good or bad? Should you scale the budget, kill the campaign, or recut the video? The numbers don't tell you — because you're looking at the wrong level. Every answer to "what to do next" lives one layer deeper than what a standard dashboard shows.
This isn't a problem with your campaign. It's a problem with your data. The five questions below each point to a specific number that most hosting platforms don't surface. Get those numbers, and the 10,000 / 47 story becomes actionable in under an hour — if your stack exposes them. Video analytics built for funnels (not vanity views) is the usual missing piece.
~73%
of video marketers can't connect watch depth to downstream conversion in their current setup
2×
higher ROAS on campaigns targeting 75%+ completion viewers vs. all video viewers
Question 01
Did the right people watch — or just a lot of people?
10,000 views means nothing if 8,000 came from an audience that was never going to buy. Before touching the video, split your completion rate by traffic source. This is the single move that changes the diagnosis most dramatically.
Completion rate by source — same campaign, same week
Email list
71%warm
Retargeting
58%warm
Paid search
42%mid
Paid social
17%cold
If paid social drove 6,000 of your 10,247 views at 17% completion, your real engaged audience is ~4,000 people — and your 0.46% view-to-sale becomes closer to 1.2% when calculated against channels where people actually watched. Same 47 sales, completely different picture.
What this tells you
Paid social completion below 25% is a targeting problem, not a script problem. Pause that channel and reallocate budget before you spend a week recutting the video.
Question 02
Where exactly are people leaving — and is there a spike?
Average completion rate is the least useful number in video analytics. It smooths together someone who watched 2 seconds with someone who watched 80%. What you need is the retention curve — and more specifically, whether there's a cliff at a specific timestamp.
Viewer retention — per-second breakdown
A smooth average hides the cliff. Here: 38% of still-watching viewers leave in a 30-second window at 6:12. That's the moment a transition falls flat, or a price hint lands without buildup. Without timestamp resolution, you'd never find it — you'd just "try a shorter script."
What this tells you
Watch the 60 seconds before the spike yourself. That specific section is your edit target. A single recut of that segment — nothing else — can recover 10–20 points of completion rate without rerecording anything.
Question 03
How many viewers actually reached your offer?
Your CTA lives at 9:32. The only conversion denominator that matters is: how many people were still watching at 9:32?
Looking at the retention curve above, roughly 1,800 viewers reached that timestamp. Your "0.46% conversion rate" is actually:
Real CVR calculation
47 sales ÷ 1,800 viewers who saw the offer = 2.6%. That's not a conversion problem — that's a retention problem. The offer is working. The script is losing people before they reach it.
This is the most misdiagnosed problem in VSL marketing. Teams rewrite the offer, change the price, add bonuses — when the actual fix is 90 seconds of editing at 6:12 that doubles the number of people who ever see the offer at all.
The diagnostic rule
Under 20% reach the CTA → retention problem (script, pacing, distractions). Over 20% reach it but don't buy → conversion problem (price, trust, offer framing). These need completely different fixes.
Question 04
Are you retargeting intent — or just retargeting views?
Most retargeting campaigns use one audience: everyone who "watched this video." That pools together someone who bounced at 8 seconds with someone who watched 11 minutes, clicked the CTA, and got distracted before checkout. Same ad, very different people.
What each retargeting segment actually needs
0–25% watchedcold → awareness ad
25–75% watchedmid → objection ad
75%+ watchedhot → close / deadline
Clicked CTAbuyer intent → FAQ / bonus
The person who clicked your CTA button already sat through your pitch. Showing them the same top-of-funnel hook ad you run to cold audiences is wasted spend and wasted signal. They need something that removes the last objection — a testimonial, a guarantee, a specific FAQ answer.
What this requires technically
Your player needs to fire Meta or Google pixel events at specific completion milestones. That's not a YouTube feature. It needs a hosting platform that exposes its event stream to your ad accounts directly.
Question 05
Is mobile destroying your funnel silently?
Depending on traffic mix, 50–70% of your viewers are on mobile. They watch on worse connections, with more interruptions, often with sound off for the first 10 seconds, frequently in a vertical orientation your player wasn't designed for. If you're only looking at overall metrics, mobile behavior is buried inside your desktop numbers.
Completion rate — same video, same week, by device
Desktop
62%
Tablet
47%
Mobile
21%
A 41-point gap between desktop and mobile isn't unusual. It means your mobile funnel is effectively broken at stage 2 — before viewers ever reach your offer. The fix isn't a shorter script. It's a mobile-specific cut with a tighter hook and properly handled muted autoplay on iOS Safari.
What this tells you
If your campaign is more than 40% mobile and you don't have device-split completion data, you've been optimising based on desktop behaviour and applying those decisions to an audience that behaves completely differently.
Why standard platforms can't answer these questions
YouTube Analytics gives you aggregate views and a smoothed retention curve. No source split, no device split by completion, no pixel events, no CTA timestamp events. Vimeo Pro improves the UX but has the same gap on pixel integration. These platforms were designed for content distribution — the metrics they show are audience metrics, not funnel metrics. A conversion stack usually pairs video hosting, a video player you can instrument, and analytics that slice by source, device, and milestone.
Here's what the same campaign looks like depending on which platform you're on — toggle between them.
10,247
Views
38%
Avg completion
—
Reached CTA
0.46%
View-to-sale CVR
Completion by source
Diagnosis
You see 0.46% CVR and don't know if the problem is the script, the audience, the price, or something else. You're optimising blind.
Answering all five questions requires a player that treats every viewer as an event stream, not a view count. Here's what that infrastructure looks like in Kinescope:
How Kinescope answers each question
01
Source-segmented completion via viewer ID stitching
Pass UTM parameters as viewer metadata. Every TimeUpdate event carries that context — completion depth becomes sliceable by source, campaign, and medium through the analytics API.
02
Per-second retention via live event stream
The TimeUpdate event fires continuously with { currentTime, percent } — timestamp-level resolution, not a smoothed average.
03
CTA at exact timestamp + CallAction event
Configure timePoints: [572] to show your offer button at exactly 9:32. When the viewer clicks, CallAction { id } fires.
04
Pixel events at completion milestones for intent-based retargeting
Native GA4 integration via gtag.js or a pre-built GTM container JSON. Fires events at 25 / 50 / 75 / 100% completion — wire directly to Meta or Google custom audiences.
05
Device + geography breakdown from day one
Built-in device split and viewer geography in the dashboard — available from the free plan without any additional setup.
Per video: Analytics → Overview — views, loads, view rate, devices, browsers, countries, OS. Switch the date range (This month, Last 24h, etc.) to match how you read the questions above.
This month (example video)
Platform comparison
Can you answer…
YouTube
Vimeo Pro
Wistia
Kinescope
Completion by traffic source?
✗
✗
Partial
✓
Drop-off at a specific timestamp?
Aggregate only
Aggregate only
✓
✓
How many viewers reached the CTA?
✗
✗
✓
✓
Pixel events at % watched?
✗
✗
Add-on
✓ Included
Mobile vs desktop completion split?
Basic
Basic
✓
✓
No ads / competitor suggestions?
✗
✓
✓
✓
Price for full analytics
Free
$65+/mo
$99+/mo
From $19/mo
Back to Monday morning — what to do with 10,247 and 47
Here's the exact sequence. Work through it before touching the ad budget or reopening the script doc.
01
Split completion by traffic source
If any source is below 25%, pause it immediately. You're paying to inflate your view count with people who aren't watching. Fix the audience before the script.
02
Find the retention cliff — exact timestamp
Open the per-second curve. Find the steepest single drop. Watch the 60 seconds before it. That section is your edit target — not the whole video.
03
Calculate real CVR: sales ÷ viewers who reached CTA timestamp
Above 3% → you have a reach problem, keep the offer. Below 1% → you have a conversion problem, the script isn't the issue. These require completely different next steps.
04
Build three retargeting audiences: 0–25%, 25–75%, 75%+
Run a different ad to each. The 75%+ audience already heard your pitch — they need social proof or urgency, not another hook.
05
Check mobile completion separately
If mobile is below 30%, your next production task is a mobile-specific cut with a tighter open and proper muted-autoplay handling — not a new script.
One thing to do today
Open your analytics for your last VSL campaign and find how many viewers reached your CTA timestamp. If you can't answer that in 60 seconds, your current setup isn't built for conversion work.
Host, protect, and stream your videos in one place.
