A demo with many views can still be failing. If most viewers leave at step 3, the rest of the demo, including the result you built it for, is seen by few people. Drop-off by step shows you exactly where that happens. This post is about finding and fixing that one step. For the wider set of numbers worth tracking, see interactive demo analytics that matter.
Short Answer
Track completion rate for the whole demo and drop-off for each step, then fix the step with the sharpest fall.
- Completion rate tells you whether the demo works overall.
- Drop-off by step tells you which step to fix.
- Fix one step at a time, republish, and compare.
Why Demo Analytics Are Harder than They Look
Most teams look at views. Views tell you the demo loaded, not that anyone understood it.
- A popular page inflates views without telling you about the demo.
- Bots can inflate counts if they are not filtered.
- Changes made all at once make it impossible to tell which one helped.
The Numbers That Matter
| Metric | What it means | Use it to |
|---|---|---|
| Views | Times the demo was opened | Check visibility |
| Unique viewers | Distinct people, approximately | Compare reach between placements |
| Completion rate | Share of viewers who reached the end | Judge the demo overall |
| Drop-off by step | How many viewers each step lost | Find the step to fix |
Read the Drop-Off Shape
- A steady slope. Every step loses a few viewers. The demo is long overall. Cut steps.
- A cliff at one step. That step is the problem. Read its callout and its screen.
- A cliff at step 1. The opening does not match what the viewer expected. Check the intro card and the page around the embed.
- A flat line to the end. The demo is working. Look at what happens after the end card.
Diagnose the Step
When one step loses viewers, it is usually one of these.
- The callout is unclear, so the viewer does not know what to click.
- The hotspot is small or off target.
- The screen is dense and the important part is hard to see.
- The step does not move the story forward.
Fixes include rewriting the callout, moving the hotspot, zooming the step, highlighting the area that matters, or cutting the step. The post on writing hotspot callouts has examples.
Change One Thing, Then Compare
Change one step, republish, and give it time to collect views. If you change three steps at once, you will not know which change mattered. Because the link stays the same when you republish, every embed picks up the new version.
An Illustrative Worked Example
This example is hypothetical. A homepage demo has nine steps. The drop-off chart shows a cliff at step 4, a filter panel.
- The callout reads "Use advanced filters to narrow your results."
- The team rewrites it to "Only clients who owe you money."
- They zoom the step on the filter so it is readable.
- They republish and check the chart again after it gathers views.
The Old Way Versus a Better Way
| Old way | Better way |
|---|---|
| Report total views | Report completion and drop-off by step |
| Change several things at once | Change one step, then compare |
| New link for each version | Republish to the same link |
| Look at analytics once after launch | Review on a regular schedule |
A Review Routine
Set a regular time to look at your demos, such as once a month.
- List each demo with its completion rate.
- For the lowest, open the drop-off chart and find the sharpest fall.
- Write down the one change you will make and the date you made it.
- At the next review, compare.
Keeping a short log of changes and dates is what turns analytics into learning.
Questions to Ask Before You Start
Answer these before recording. They take a few minutes and prevent most rework.
- Which demos matter most, and which are reviewed first?
- Is each placement on its own demo, so the numbers are not mixed?
- Who keeps the log of changes and dates?
- How long will each change run before you compare?
- What happens after the end card, and how will you track it on your site?
Where Proovyt Fits (Full Disclosure)
Proovyt is an interactive demo tool, and this is our blog.
- Creator, Pro and Team report views, unique viewers, completion rate and drop-off by step for hosted links and embeds. Free has basic view counts.
- Known bots are filtered out.
- Hosted pages and embeds set no cookies. Analytics store no IP addresses, and unique viewers come from a hash that changes every day.
- Pro and Team add per-viewer analytics with CSV export for viewers who send a lead form.
Proovyt does not sync analytics to a CRM yet. See pricing and the product page.
Common Mistakes
- Judging a demo by views.
- Changing many steps at once.
- Ignoring a cliff at step 1, which usually means a mismatch with the page.
- Publishing a new link for each version and splitting the data.
The Bottom Line
Drop-off by step turns "the demo is not working" into "step 4 is not working." Fix that step, republish to the same link, and check again.
Frequently Asked Questions
What is a good completion rate? There is no universal benchmark we can point to. Compare your demo with its own earlier versions.
Why are unique viewers approximate? Proovyt counts them with a hash that changes every day and keeps no IP addresses, so the same person on two days can count twice.
Are bots counted? Known bots are filtered out.
Can I see individual viewers? Only those who send a lead form, on Pro and Team. Everyone else stays anonymous.
Does republishing reset analytics? Republishing keeps the link, so the demo's numbers stay in one place. Note the date you republished, and compare the period before it with the period after.