I am working on a project that provides a service (landscaping) to members of the community. When the work for that client is finished we always encourage the client to go to our website to fill out a survey to review the work that we have done.

The problem that we have been having is simply that we are only getting around 20% of the customers to fill out a survey, and only about 10% of those have anything useful on them.

In the past we have tried to convince customers to give use feedback on these surveys by offering a sort of e-coupon sent to their email. This only generated about 20% more surveys with only 5% being useful and the rest being something like "Good service guys".

My question is, how do I convince customers to fill out our surveys, and fill them out with useful comments? Do we need to offer a coupon again, but only for "useful" surveys (how do you judge something like that)? Or should we offer a certain % off their final payment? Or are there others ways of convincing a customer to fill one out?

  • 2
    Well the most important question here is why do you want them to fill in a survey? What do you use this feedback for? For gathering testimonials to put on the website? (If so then you can probably gather that more effectively by just asking the landscaper to prompt the customer for a quote for you to use on the website - that takes far less effort on the clients part).
    – JonW
    Commented Jun 14, 2012 at 19:56
  • Coupons are useful if you expect repeat business. Do you? There's a spectrum there; I sure hope a coupon from the guy who replaced my furnace would be unhelpful, while a coupon from the restaurant I visit once a week would be. What's your customer profile like in that regard? Commented Jun 14, 2012 at 19:57
  • 1
    @JonW Believe it or not we actually use the surveys ways to improve on what we are doing. We have used a couple for testimonials though.
    – PL3X
    Commented Jun 14, 2012 at 20:13
  • 2
    @Yallow, so the workflow supports doing the survey before the reckoning of the final bill? That sounds like your best bet, then; it's hard to argue with cold, hard cash. :-) Commented Jun 14, 2012 at 20:24
  • 1
    how about a tiny gift? Like a small gift certificate for a local business?
    – DA01
    Commented Jun 14, 2012 at 20:41

6 Answers 6


Are you asking users to fill in text boxes and write long narratives? If so, this is a high commitment task. Instead, make answering your survey a low commitment task by:

  1. Shortening the length of the survey overall--could you limit it to one or two questions, varying questions by user? This would increase participation overall and you could still get all of your questions answered.
  2. Making each question quick and easy to answer--instead of asking users to describe something, give them a description and ask them to agree/disagree on a Likert scale (e.g., 1 = totally disagree, 5 = totally agree).
  3. Providing an optional open field at the end (e.g., Anything else we should know?) for freeform answers.
  • 5
    This raises a good question: are people currently visiting the survey and bailing out without completing it, or are they not bothering to even go and look at it?
    – JonW
    Commented Jun 14, 2012 at 20:54
  • As it stands right now we have a three question survey with a comment box at the end for any additional thoughts. I do like the 1-5 scale idea!
    – PL3X
    Commented Jun 15, 2012 at 3:24
  • I'd put the comment box at the beginning. Then if they want to say anything they can without having to engage with your preset questions at all.
    – PhillipW
    Commented Jun 15, 2012 at 19:43

Request their email onsite, so that you can send the survey link directly to their inbox.

Why should they bother? Make sure to preface the survey with a paragraph about how much you value their feedback, stating explicitly that you want to learn how to deliver a better service to them personally

Make all questions optional and make sure the user is aware of this

Users often abandon surveys half way through. Save changes periodically or paginate

Allow users to build momentum by offering simple questions first.

Question phrasing can often be overlooked. Be clear and direct do not allow for ambiguity. Get rid of the guess work.

A longer list of questions, each requiring a brief response is much more effective than essay type questions

  • 3
    This is some truly good advice. Not only covers how to get them to the survey in the first place, but also gives some great advice about how to keep them there. 'Allow users to build momentum by offering simple questions first' is an inspired idea, I'll be using that myself next time I come to spec a form / wizard.
    – JonW
    Commented Jun 15, 2012 at 20:14
  • I agree! This is excellent. It is so common sense to build momentum with really anything you do. So why not with surveys too? Thanks!
    – PL3X
    Commented Jun 16, 2012 at 1:59

Make it fun to fill out the survey.

Most surveys are DEATHLY DULL. I don't know why, but somehow people think this is acceptable. Meanwhile, there are a ton of "What Mad Men Character Are You?" quiz sites out there that are super popular.

Why not combine the two? Gather data about your audience while providing entertainment.

The NY Times did this last year, categorizing and surveying its users with only the promise of telling them what kind of person they were. I don't have the link, but it looked up front like a 'quiz' site--and told them a hell of a lot about what products you purchase, what income level you were, and so forth.

Alternately, just make your survey a little more fun. Graphics and messaging might be enough.


One of the core difficulties is that if you offer a gift or discount for doing the survey, then more people will fill them out, but it doesn't necessarily improve the quality of the feedback. So you don't really gain.

As it is, those who fill out the survey will be a portion of your customer base who a) like filling out surveys, b)loved your service and/or c)hated your service. You don't tend to get the ones who could make some small suggestions for improvements that could make a big difference - people who generally liked your service, but had an issue that means they wouldn't use you again or recommend you. These are the important ones.

TBH, 20% response, with 10% of those being insightful is not bad. Most inducements are liable to increase the 18% who respond with platitudes, rather than the 2% who give you good feedback.

Is there any way that someone unconnected with the work being done could call at the customers property after the work is complete, to check that everything is satisfactory and settling in well, and also do the questionnaire with them? That would probably produce more useful insight, although I accept that it is an expensive option.


I guess one of the problem is that there may not be a clear incentive for the users to fill out a survey. People will not require certain types of services more than once, and so you can't really entice them to fill it out 'for their own good'. This is where testimonies (which can just add to the perception of quality, but may be made-up) or ratings can do more than what a survey will. Of course, if you can demonstrate how filling a survey out will help them in the future, then it is going to be more effective because it is not a bribe but a genuine incentive. And the rule of thumb is to be short and sweet like what everyone else has said.


Are you also considering how many people fill the survey on desktop vs mobile?

One reason could be just that it is difficult for one to fill it out on mobile and that's why they drop off.

Also, as Alex mentioned, forms are extremely dull and if you can engage the user in some manner, they are going to continue till the end. It could be through quiz, gamification or making it a casual chat. I wouldn't want to tell you my name if you ask "Name" as a field label. But if you ask me, "Dude, I didn't catch your name?" - it's highly likely I'll fill that out.

My clear assumption here is that the user has atleast opened the form.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.