Building an ATS tool taught me that resume scores aren't the biggest problem

When I started building ATSAlign, I thought the hardest part would be coming up with an accurate ATS score.

After talking to users and testing different approaches, I realized the score itself wasn't what people were looking for.

Most of the questions I received were things like:

"Why doesn't my resume match this role?"

"Which skills am I actually missing?"

"Is this keyword really important?"

Very few people asked how the score was calculated.

That completely changed how I approached the product. Instead of focusing on making the score more sophisticated, I spent more time helping users understand why their resume matched or didn't match a specific job description and what they could do to improve it.

Looking back, I think explanations are far more valuable than a single number.

I'm curious whether other founders have had a similar experience. Have you ever built a feature that you thought users would care about, only to discover they were trying to solve a completely different problem?