Look for Where It Narrows Sharply
A funnel narrows by definition, so it's easy to glance at one and see nothing but expected shape. The more useful question isn't whether the funnel narrows — it's where it narrows the most sharply relative to the stage before it, since that's usually where the most fixable problem is hiding, not necessarily at the widest or most visible stage.
Cohorts, Not Snapshots
Looking at a funnel as a single snapshot also hides how it behaves for different groups of people over time. Two cohorts that entered a month apart can have very different conversion patterns, and averaging them together can make a real improvement — or a real problem — invisible. Tracking cohorts separately, even roughly, tends to reveal more than a single blended number ever will.
More Traffic Won't Fix a Downstream Leak
It's tempting to respond to a weak funnel by pouring more people into the top, since that's usually the easiest lever to pull. But if the real leak is in the middle or bottom of the funnel, more top-of-funnel volume just means more people hitting the same leak — the numbers get bigger, but the underlying rate that actually matters doesn't improve.