What drop-off measures
Drop-off is the share of people who reach a step in a conversion funnel and do not reach the next one. It is measured between two adjacent steps, never as a single site-wide figure, and it is the number that turns a vague sense that the site underperforms into a specific place to work.
Every funnel has drop-off at every step, because not everyone arriving intended to buy. What matters is where the loss concentrates. A steady thinning from step to step is the ordinary cost of doing business. A sudden collapse between two particular steps is a problem with something in between them: a form, a price reveal, a payment method, a page that will not load on a mid-range Android phone.
The measurement only means something if the two steps are genuinely sequential and both are tracked as real events. Comparing a page view against an inferred action gives you a number that moves for reasons unrelated to the visitor.
Why drop-off matters
It is the cheapest improvement available to most sites. Traffic already paid for is sitting at the failing step; recovering some of it costs nothing per extra visitor, while buying the same number of new conversions through ads costs money every month. For small businesses in Nepal running modest ad budgets, this is usually the difference between a campaign that works and one that does not.
It also prevents misdiagnosis. A campaign blamed for poor results is often fine at bringing the right people; the loss happens after the click, on a page nobody in the ad account can see. Looking at drop-off before touching bids stops you from cutting a channel that was never the problem.
Common mistakes with drop-off
The first is treating high drop-off at the top of a funnel as a failure. Early steps carry the most casual visitors and will always shed the most people. Judge each step against itself over time, and against the same step on other devices, rather than against the step below it.
The second is fixing the loudest step instead of the costliest one. A step that loses a modest share of a large audience can waste far more than a step that loses most of a tiny one. Weigh the loss by the number of people behind it.
The third is guessing at causes. Analytics tells you where people leave; it rarely tells you why. Session recordings, form field analysis and a few honest conversations with customers close that gap faster than another round of theories.
How to act on it
Find the single step with the largest weighted loss, then go and use it yourself on a phone, on a normal connection, as a first-time visitor would. Most serious drop-off has an obvious physical cause once you stop testing on your own laptop: a form that demands information the visitor does not have, a delivery cost revealed late, a payment option the local audience does not hold, a button below a fold nobody scrolls past.
Change one thing, let it run long enough to be believable, and re-measure the same two steps. Then move to the next largest loss. That loop is the practical core of conversion rate optimisation.