Customer Feedback Forms That Get Answered

Short answer: two questions — a rating and an open "why" — sent the same day the job finishes, through the channel the customer already uses. Feedback fails from length and timing far more often than from wording, and the open answer is worth more than the score.

Feedback surveys have a peculiar failure mode. The business that most wants to know how it is doing builds the longest form, sends it a week late, gets a four per cent response rate from the two most extreme customers, and concludes that customers do not fill in surveys.

Customers fill in surveys. They do not fill in fifteen-question satisfaction instruments sent eight days after a visit they have stopped thinking about.

Ask two questions

The whole of a post-job survey:

A rating. Any scale you will use consistently — one to five stars, or nought to ten if you want NPS. The scale matters less than never changing it, because the value is in the trend.

"What made it that?" One open box. This is where everything useful is. A four-star rating tells you nothing actionable; "great work but nobody told me you were coming" tells you exactly what to fix on Monday.

Optionally a third, but only if it is the specific thing you are working on this quarter — "was the arrival window accurate?" while you are trying to fix scheduling. Retire it when the question is answered.

That is the form. Resist the urge to add demographics, a matrix of sub-ratings, or "how did you hear about us" — the last one belongs on the enquiry form where the answer is accurate, not here where it is a guess.

Timing is most of the response rate

The window where someone will answer is much shorter than businesses assume.

Moment	Send when	What you learn
Job or visit completed	Same day	The experience, while it is vivid
Product delivered	1–3 days after arrival	First impressions and condition
Onboarding finished	End of the first week	Whether getting started worked
Support issue closed	Within hours of resolution	Whether the fix actually landed
Long-term relationship	Quarterly or twice a year	Drift, and reasons people leave
Lost quote or churn	Within a week of the decision	The most valuable and least collected

The last row is the one almost nobody does. The customer who chose someone else knows precisely why, will often tell you, and is never asked. One question — "we'd genuinely like to know what tipped it" — sent to lost quotes for a month will teach you more than a year of satisfaction scores.

Send it where they already are

The channel decides most of the rest of the response rate. An email to a customer who only ever texted you is a message into a void.

A short text with a link is the highest-response channel for consumer services, and the reason is simply that it is read within minutes. If you work by email, email; if the relationship happened over WhatsApp, the link goes there. The comparison of the two main automated channels, and their costs, is in speed to lead: SMS versus email. On BotForms, SMS is a separate add-on from $29 a month for 1,000 messages as of September 2026, while email notification is on every plan.

Two things to put in the message itself: how long it takes ("thirty seconds"), and a reason that is about them, not you. "So we can fix what slowed us down" beats "your feedback is important to us", which is the most ignored sentence in commerce.

One question first, then the rest

The technique that lifts response rate more than any wording change: make the first interaction a single click.

Put the rating in the message as tappable options. The click submits the rating and opens the form with the rest. You have captured the score from everybody who clicked, even the ones who abandon the follow-up, and the open box now looks like a small extra rather than the price of entry.

This works because the ask is honest — you said it was one question and it was. It fails if the click lands on a five-question page, which trains people not to click next time.

A conversational format suits feedback particularly well for the same reason: one question at a time, each answer shaping the next, so someone with a complaint gets asked about the complaint and a happy customer is not marched through it. The mechanics are in conversational forms that feel like chat, and what that kind of form bot solution is doing underneath is in the explainer.

NPS, honestly

Net Promoter Score is the nought-to-ten recommendation question, with the low scores subtracted from the high ones to produce a single number.

Its virtue is comparability over time and against an industry. Its limitation for a small business is sample size: a score built from nine responses moves several points when one customer has a bad week, and the movement means nothing. Treat it as a trend that needs a quarter of data before it says anything, and treat the open answers as the actual output.

If you use it, use the standard question and scale unchanged. The moment you reword it or switch to a five-point scale, you have a private metric that cannot be compared to anything, which is fine — just do not call it NPS.

Feedback in two languages

If a meaningful share of your customers speak Spanish, a survey in English is measuring your English-speaking customers. That is a sampling problem masquerading as a satisfaction score, and it usually hides the segment with the most to say about how the job went. Running both versions is covered in multi-language forms in Spanish and English.

Do something visible with it

The reason feedback programmes die is that nothing happens, people notice, and responses fall.

Reply to the bad ones today. From a person, not a template. A complaint answered within hours frequently ends with a more loyal customer than a job that went smoothly — the recovery is the memorable part.

Count themes monthly. One complaint is an anecdote. The same complaint five times is a process problem, and it is the only thing in the responses that should change how you operate.

Tell people what changed. "You told us our arrival windows were too vague, so they are now two hours instead of four." This is the single most effective thing you can do for future response rates, and almost nobody does it.

Route it to the person who can act. Feedback about a specific technician's work should reach their manager, not a marketing inbox. Same logic as routing form submissions to the right person.

Tracking whether responses are rising or falling, and which send times work, is the same discipline as any other form — the measurements worth keeping are in form analytics metrics that matter.

Reviews come after, never instead

Private feedback and public reviews are different instruments and the order matters.

Ask privately first. You hear the problems before they are on the internet, and you get the chance to fix them. Then invite the people who were happy to leave a public review — with a direct link, because friction kills this step.

What you must not do is build the survey as a filter that routes high scores to the review site and low scores to a private apology form. Several major review platforms explicitly prohibit that, and the sanction is applied to your listing. Ask everyone to review, or ask nobody; select on nothing but whether they replied.

Similarly, do not offer anything in exchange for a public review. A prize draw for completing a private survey is ordinary practice; the same offer attached to a public review breaches most platforms' terms.

Where the responses should end up

Two useful destinations beyond reading them.

Feedback responses tie back to the customer record, which makes the next conversation better — knowing someone rated the last job three stars is exactly what the person picking up the phone should see. And the contacts who said they were happy are the natural list for a referral ask or a case-study request later, which is a job for form submissions feeding an email audience.

Getting started

Pick one moment — job completion is the usual best — and build two questions. Send it by whatever channel that customer already used, on the same day, with the time cost stated.

Run it for a month without changing anything, then read the open answers in one sitting and write down the three themes. Fix one. Tell people you fixed it.

That loop, run four times a year, is worth more than any survey instrument.

BotForms builds the form from a description, runs it conversationally one question at a time, and keeps the responses in one dashboard alongside the rest of your submissions. Starter is $29/month, Pro $79, Business $199, with SMS as a separate add-on from $29 and a 30-day trial on every plan. See features, pricing, or start a trial.