@Irish_Elk and escoville: I think Irish_Elk is right. Asking open-ended questions is exactly what is done in qualitative surveys or interviews. If you provide a scale or multiple-choice boxes (which - correct me if I'm wrong - would be a quantitative approach, asking for "how much importance is attached to the aspects we defined"), the interviewees are unable to give any answers other than those that you expected (and therefore listed in your answer scheme).
But if you don't want to assess the importance or relevance of preset aspects, but rather want to identify the aspects themselves, you do a qualitative survey. In this case, you wouldn't want to influence the interviewees with answer schemes that you provide, but rather leave everything open so they can come up with aspects that you maybe hadn't even thought of.
"How could we improve our website?" is a perfect example for this. The interviewees can write anything here. If you had asked "Should we improve our website by: a) enlarging the pictures b) simplifying the dropdown menus or c) offering a German version" you'd rule out options like "increasing the font size", "inserting videos" etc... and you could only do a quantitative analysis ("How many interviewees ticked a, how many ticked b, etc.)