Read the overview
Choose a useful date range before comparing results. Internal and QA traffic is excluded from analytics. Some breakdowns require a higher plan; the interface identifies features that need an upgrade.
Find the step to improve
The Step funnel shows views for each step and, when available, hotspot clicks, CTA clicks, and average time. Look for a meaningful change between steps, then view the demo yourself from the preceding step. Check whether the next action is visible, the instruction matches the screen, and the content answers the question that brought the viewer there. A drop is a clue, not a diagnosis: someone may leave because the next action is unclear, or because they already found what they needed. Branching paths and different entry points can also affect the shape of a funnel. Change one thing, publish it, and compare a later period with the earlier one. Treat small samples carefully and combine analytics with feedback from people who use or share the demo.Example: investigate a drop
Suppose a fictional demo has 100 views and 40 completed views in a period. Its completion rate is 40%. The funnel shows fewer viewers reaching the step after a filter selection. That tells you where to investigate; it does not tell you why people left. Open the published demo and try the preceding step on the devices your audience uses. Is the filter visible? Does the guide use the same label as the screen? Does the next step show the promised result? Ask a colleague to try it without coaching. If the instruction is unclear, revise that instruction and publish the change. Compare a later period using the same metric and a comparable audience. If traffic came from a different campaign, or only a few people viewed the demo, a change in completion may have other causes. Completion is also not the only outcome: an onboarding guide can be useful even when someone leaves to perform the task in their own account.Choose the next experiment
Write down the question, what you observed, and one change to try. For example:
These observations suggest what to investigate; they do not prove that the proposed change will work. Record the published change date, compare an appropriate later period, and include feedback from actual reviewers. Keep the audience and distribution channel in mind: a website visitor and a buyer receiving a follow-up may behave differently.