Feedback form: the five questions worth asking, and where to put it

A feedback form is the cheapest research a business can run and the easiest to waste. The wasted version asks ten questions, appears on a page nobody is finished with, and feeds a dashboard that nobody has agreed to act on. The useful version is short, appears right after something happened, and has one owner reading the answers.

Here is the version that works, and the placement decisions that matter more than the wording.

The five questions

• What were you trying to do? — the context question. Without it, every other answer is ambiguous.
• Did it work? — a simple yes, partly, no. Three options beat a seven-point scale for a form people fill in on a phone.
• What was harder than it should have been? — the open question that produces most of the value. Leave it optional and give it room.
• One rating you will track over time — satisfaction with this specific interaction, not with the company in general.
• May we follow up, and at what address? — optional, and the reason people give you the detail that makes the rest actionable.

Ask one thing per question and put the open box in the middle rather than at the end. By the time someone reaches the last field they are finishing, not thinking — and the open answer is the part you actually needed.

Placement decides response rate

A feedback form on a contact page collects complaints from people motivated enough to hunt for it. The same form shown immediately after a completed order, a resolved support ticket, or a first successful use collects the ordinary experience — which is the one you are trying to improve.

Three placements cover most needs: after a completed action, inside the confirmation email, and as a persistent but unobtrusive link in the footer for anyone who wants to volunteer something. Pop-ups that interrupt an unfinished task collect irritation, accurately measured.

Building one that gets completed

1. Decide what you will do with each possible answer before writing the questions. Anything you would not act on does not belong on the form.
2. Keep it to one screen, and say how long it takes — honestly.
3. Use branching so people only see what applies: someone who said "it worked" does not need the troubleshooting question.
4. Make everything optional except the first question. Required fields on a voluntary form cost you completed responses.
5. Send the confirmation with a sentence about what happens next. Silence teaches people that feedback goes nowhere.
6. Report back publicly once in a while — "you said, we changed" is what makes the second round of feedback arrive.

Ettex Forms covers this end to end: a drag-and-drop builder from question blocks, logic and branching so people only see relevant questions, choice, scale, date and file-upload question types, multi-step sections with progress, validation for required fields and formats, custom themes to match your brand, quotas and limits to close a form by date or response count, and publishing by link or embedded on your site. Responses land in a timestamped, searchable inbox with a live summary, the team can review and process them together, and everything exports to CSV or XLS. Feedback, survey and registration templates ship with it.

What makes feedback unusable

• Leading questions. "How much did you enjoy the new checkout?" and "What did you think of the new checkout?" produce different data, and only one of them is worth having.
• Asking for a rating with no open box, so you learn the temperature and never the reason.
• Collecting responses nobody owns. An inbox without a reader is a form that wastes goodwill.
• Running the form permanently and never comparing periods, so you cannot tell whether anything improved.
• Ignoring the quiet majority: forms over-sample the delighted and the furious, and that skew is structural, not fixable by wording.

A good feedback form asks what someone was trying to do, whether it worked, and what got in the way — then someone reads the answers and changes something. Everything beyond that is optional.