I built a SaaS for the past 6 months and just made the most uncomfortable change yet:
I removed every claim about "perfect" or "faithful" conversion from my landing page.
The reasoning:
1. CSS has fundamental limits.
For my SaaS (a Figma to HTML/CSS converter), gradient strokes don't work cleanly with border-radius. Variable-width strokes can't be reproduced (Figma's API doesn't even expose the width profile). Some blend modes behave differently across browsers. Any tool promising "1px-perfect" reproduction is either over-promising or limiting itself drastically.
2. Web developers see through inflated claims immediately.
My target users — freelance web developers and small studio implementers — have coded enough sites to know what their tools can and can't do. When marketing says "perfect," they get suspicious, not excited. They've been burned by tools that promised the moon before.
3. Honest expectations build trust.
When users try a product with realistic expectations and the experience matches, they keep using it. When they expected "perfect" and got "approximated," they churn and warn others. The expectation gap is the real killer.
So I rewrote the LP:
Counterintuitively, this feels stronger as marketing — because it sets expectations that the product can actually meet.
Curious about other indie hackers' experiences:
Would love to hear how this has played out for others.
I think this is especially true when the target users are experienced enough to understand the tradeoffs underneath the product.
Overpromising can work briefly, but once users detect the gap between marketing language and system reality, trust drops very quickly.
What stood out to me is that your change was not really about weaker positioning. It was about better expectation calibration.
In a strange way, honest constraints can actually make a product feel more credible, because the user starts believing the claims that remain.
I have been noticing something similar in the WordPress space as well. Experienced users often trust products more when they clearly acknowledge limitations instead of pretending edge cases do not exist.