On-site surveys
Job of the week
Featured threads
- How relevant do links need to be? 14 replies
- Tracking Online Response to Marketing/Communications Activities 8 replies
- Behavioural targeting software 4 replies
- Penalty avoidance on English-speaking foreign sites 5 replies
- 3 way linking - good or bad? 21 replies
Most viewed threads in last month
Most active threads in last month
- Entry level search function 0 replies
- Introduction 0 replies

Head of Digital Experience at Royal Mail group
23 June 2008 07:55am
We are looking at running an on-site survey to gather information from our users - usual sort of thing. However there is a big debate internally about where and how the survey should appear. The options are as follows:
1) Do the standard pop-under when a user exits the site
2) Serve a banner as an eye-blaster after the user has been on the site say 20 seconds, asking them to complete the survey. Clicking on the banner opens the survey. This then gets round the possible problem, that if a user keeps their browser open all day, they will still see the survey maybe hours after going on the website. But this will then affect the user-journey and take the user off our website.
Can anyone offer any advice / comments on these 2 options?
Many thanks!
Caroline
CEO at Econsultancy
23 June 2008 09:46am
Hi Caroline
I'd advise not quite either of your options. Pop-unders can easily get missed (and blocked) and I'm not sure I like the sound of a 'banner eye-blaster'!
I like the way Intuit / Quickbooks do it (see http://quickbooks.intuit.com) both on terms of *what* they ask (short, but the most important questions) and *how* they ask.
They have a top left drop-down / Javascript layer that first asks whether you're an existing customer or not and why you've come to the site. They then ask *permission* to survey you at the end of your visit. This courtesy makes you more likely to say yes and MUCH more likely to fill in the survey at the end of the visit, so it helps their completion rates as well putting the cusotmer completely in control.
Regards
Ashley Friedlein
CEO
E-consultancy.com
Technical Project Manager (MBA, MBCS, CITP, CEng) at Naxtech.com
23 June 2008 11:56am
I think Ashley's idea is not bad. An alternative option could be a link (either text or graphics) which would then either open a new page or a hidden area (div) within the same page which would contain the survey or information you want.
I hope this helps.
regards,
Denis
www.naxtech.com
Head of Digital Experience at Royal Mail group
24 June 2008 06:14am
Does anyone have any stats or research on why it is a really bad idea to take users off to a survey, part way through a visit - to an external survey tool. I know the reasons why, but have been told this is not 'concrete' enough!
People don't seem to understand it is general bad practice to take users off to another website part way through their visit!!
Web Project Manager at World Archipelago
24 June 2008 13:58pm
How about this.
http://4q.iperceptions.com/
4Q employs a two-stage invitation process. When visitors arrive at your site, they will be presented an invitation to participate in a survey after their session. If they accept, a second, minimized window, which contains the survey itself, will be launched and will wait in the background for the visitor to complete his or her session. 4Q surveys are designed to be collaborative brand building exercises, not annoying browsing interruptions.
hope that helps
On 07:55:50 23 June 2008 CWhyatt wrote:
CEO at Econsultancy
24 June 2008 14:10pm
Yes, I think the 4Q approach is great.
Ashley
CEO at Clicktools
25 June 2008 09:54am
Caroline,
Permission based is always best -i e ask them if they are willing to participate and then present the survey if they agree. You don't say if the users are registered, in which case you could send the survey via email after the visit.
A few other points to consider:
Don't present the invitation or survey at a critical stage in the on-line journey, eg checkout or during an application process. It will distract.
If you launch the survey from different pages, capture this information, it can be useful in subsequent analysis.
Keep surveys short - long surveys = high drop out and unreliable results.
Write the survey to reflect and test what is important to the customer/user and use language that is appropriate to them.
Cheers
Dave J
www.clicktools.com
On 07:55:50 23 June 2008 CWhyatt wrote:
Research Director/Head of Web research at SPA
30 June 2008 10:19am
Hi Caroline,
I agree with the general consensus that invite on entry and survey after the end of the visit is the best approach. This will ensure that the visitor can continue with the site journey they intended and therefore also provide you with considered feedback on their site experience. This used to be easy before the days of pop-up blockers but is more complex to achieve these days.
The way we implement on-site surveys for our clients is a dynamic DIV layer invite on site entry (shown on entry for 10 seconds to a sample of visitors, eg 1 in 10, visitors cookied so most visitors only invited once in 60 days). If they agree to take part in the survey then a new window is launched that contains a script that checks every few seconds if they are still on the same site - if they have left the site then the survey is launched - similar to an exit survey. If they are not interested then they will not be invited again until cookies are cleared.
Chris
On 07:55:50 23 June 2008 CWhyatt wrote:
Head of Digital Experience at Royal Mail group
30 June 2008 14:29pm
Thanks Chris, that sounds interesting, what is your website URL?
On 10:19:23 30 June 2008 ChrisHandford2 wrote:
Research Director/Head of Web research at SPA
14 July 2008 16:30pm
Hi Caroline,
Unfortunately this is not on our own site but something we implement for our clients and we're not able to mention clients without approval. I head web research at a market research agency.
We hope to have a demo site set up in the next month so we can share this approach. I'll post something when this is in place.
Cheers,
Chris
SPA Research
On 14:29:38 30 June 2008 CWhyatt wrote: