Pricing page error that no one ever tells you about

In the past month, I've been working on revamping our pricing page copy, tightening the feature bullets, adding the comparison table. The result was zero effect on conversions – not even in a round-off error way.

Then, after doing literally nothing to the copy, I removed our cheapest plan and saw almost 20% lift in conversion rates for the remaining plans. What we've done before was not getting more customers, it was giving potential buyers a refuge from commitment to the plan they needed.

Before I've heard about decoy pricing and never thought it would work for my five-man start-up without a dedicated data science team behind.

Have you tried testing removal of the plan, rather than adding a new one? And how does it work for usage-based pricing?