Does building requested features actually fix low conversion?

Whenever a prospective customer said "I’ll buy if you add X feature, we built it. Six months later our product was twice as complex, but our trial to paid conversion rate hadn't moved an inch.

Turned out, the feature requests were just polite excuses. The real barrier was that users didn't understand our core value proposition within the first 30 seconds.

At what point did you stop building features and focus entirely on positioning?