Losing a featured snippet can raise traffic because the snippet was answering the question so completely that searchers never needed to click, and once it disappears, those users have to open a result to get the answer. The mechanism is straightforward once you stop assuming a snippet is always a win: a snippet that fully satisfies the query cannibalizes your own click. This is observed behavior reported across many sites rather than a rule Google publishes, so treat it as a pattern worth confirming against your own data.
The dynamic depends on what kind of query you hold. When a snippet gives a short, self-contained answer (a definition, a date, a yes or no), the searcher reads it on the results page and leaves. You earn the visibility but not the visit. When that snippet drops and your page falls back to a standard blue listing, the same searcher now has to click to read the answer, and the click returns to you. The traffic rise is not Google rewarding you for losing the snippet, it is the click that the snippet was quietly taking back.
The opposite case is just as real, which is why this is not a reason to chase snippet loss. A snippet on a complex or commercial query often acts as a hook, the brief answer leaves the reader wanting the full picture, and the prominent placement drives more clicks than a standard listing would. Lose that one and traffic falls. So the same feature helps on some queries and cannibalizes on others, and the difference is whether the snippet finishes the job or only starts it.
To judge your own case, pull the query in Search Console and compare impressions, clicks, and average position before and after the snippet changed. If clicks rose while you dropped from the snippet to an ordinary listing, the snippet was answering for you and the loss recovered the visit. If clicks fell, the snippet was earning the click and you want it back. Decide per query rather than treating every snippet as worth holding.