The first should be pretty obvious, but the second is a little less obvious. But think about it this way: if it worked for me, I wouldn't complain about it. Agree?
So there are two possible choices: either I'm wrong or your code is wrong (assuming you're an engineer). As a member of a startup, it is generally in your best interest to assume that the complainant is right and your code is wrong.
Assume that your code is wrong. Investigate it. Ask for more information if you need it. Track down the problem. If it works for you, then think of all the reasons it might not work for someone who doesn't know how it works. Think of all the edge cases. Clear the cookies. Bust the cache. Make sure it's not Wednesday (http://gyrovague.com/2015/07/29/crashes-only-on-wednesdays/). Whatever it takes to figure out why it's not working for the user (which may or may not mean it works for you!).
Maybe the code is wrong -- if so, fix it. Maybe it's a feature -- if so, explain that. Maybe the user made a mistake -- if so, identify why and fix that.
Only after identifying the problem and potential solution do you go back to your complainant and say, "Here's what happened and here's what we're doing about it." If the complainant made a mistake, you can explain "it works", but you should also follow up with some way that you're going to make sure this doesn't happen again.
This is basically the single best way I know to resolve complaints. People respect you significantly more when you can always present a solution -- even if it turns out they were wrong all along!