Best Practices for "Was this helpful?"
16 Mar 2010 - 11:42am
7 replies
1794 reads
Many sites that offer help conclude the information with a feedback question such as "Was this helpful?" This can be a valuable mechanism to tailor online or in product Help better.
Our team would like your help: We'd like to hear what everyone does. What question do you ask with what scale or input? How well does that work? What would you change? What are some best practices for "Was this helpful?" questions.
Josie Scott
TechSmith Corporation User Experience
Comments
We use a straight yes/no question: "Was this answer helpful to you?" When a customer is browsing FAQs, they may be annoyed or frustrated about something, and asking them to think about rating the value of an FAQ on a scale may increase that annoyance, so we keep it simple. We do supplement our yes/no question on occasion with surveys that get at more detailed answers, but just the yes/no answers have been extremely valuable in evaluating our FAQs.
We ask "Was this helpful" and allow the user to leave a comment, then push all the comments into a database, extract entities and datavis them based on rank/affinity with other keywords. This can be pretty painful (and occasionally humorous), but it really lets you know how people are feeling about it and pretty quickly when changing it works.
I think the comment field lets users feel like they can vent their frustration, and also provides us with the opportunity to learn what they expected from the page. This helps with SEO and related links as well.
Thanks, both of you. Nixkuroi, can you clarify for me: Do you ask the question followed by a yes/no input and a text field, or just the question and the text field? (Love your analysis idea.)
josie
We ask like this
Was this helpful? O Yes O No
Comment (optional)
_____________________________________
| |
| |
_____________________________________
Send
I realized I gave an incomplete answer earlier. If a customer rates a question as not helpful, we pop open a little window with a field for comments. Providing the extra feedback is optional, but a lot of customers do provide it.
Ronny Kohavi's group has published some work showing a dramatically increased response rate for Yes/No versus more complex questions, with inline presentation:
http://exp-platform.com/acmDMSig.aspx (see "Slides from Talk: Practical Guide to Controlled Experiments on the Web: Listen to Your Customers not to the HiPPO")
I'd also recommend capturing as much analytic state information as you can with the user response. At a minimum, the referring URL can help you mine this dataset for additional insights.
hth, Andy
On Tue, Mar 16, 2010 at 3:45 PM, Amy Silvers <contact@ixda.org> wrote:
What I've typically asked is:
- Was this helpful?
( ) yes
( ) no (What were you expecting instead?)
[ textarea]
Typically I've gotten about 10% of the "no" answers to enter useful freeform information.
If you have a product where you'd like to ask this question WITHOUT building anything,
I am the product manager for KISSinsights, a product that allows you to embed targeted surveys within your web application. I just started letting people into our beta this week; email me directly if you'd like to try it out.
Thanks!
Cindy Alvarez
--
The Experience is the Product: 5 “Seducible Moments” to Hook Web Application Users:
http://www.cindyalvarez.com/user-delight/5-seducible-moments-to-hook-web-application-users
On Tue, Mar 16, 2010 at 9:43 AM, Josephine Scott <contact@ixda.org> wrote: