7 tips to get the most out of online surveys

Example of an online survey used on
Blue Latitude was using an unobtrusive popup in the bottom right hand corner to get contacts

Industries which have disconnect between the digital activity of their customers and the sales they generate have difficulty proving return on investment. Part of our job at Blue Latitude is working with them to extract Key Performance Indicators (KPIs) from their Business Objectives; and helping them to enact those on their digital activity. KPIs are often a leap of faith where we think that a user doing something makes them more likely to buy or, in the case of websites aimed at health care professionals, more likely to prescribe.

A popular recent trend has been to validate those KPIs through running online surveys. We ask people who have just arrived at the site how likely they are to buy (or prescribe, or go to their doctor or whatever your business objective is) and then compare it to those who are just leaving and have seen our key content. The online survey has the added bonus that you can ask questions that will give you insight about your users for future projects, and can allow you to collect contact details.

The trouble with surveys is that they are now ubiquitous across the web. Too many businesses have been told by smartly dressed Heads of Optimisation that they need to ask their users to get insight and satisfaction scores; now there is barely a website out there that doesn’t ask you to participate. The reason is simple: it is cheap (in some cases free), easy to implement and not too intrusive (if done the right way).

However this has led to a problem of decreasing response rates, as users no longer see the survey as something special to them. As a website user, my time is limited and unless I have high levels of interest in the brand or I’m going to be suitably compensated I will generally ignore the survey.

So here are some tips on how to make sure that you get the appropriate response rate:

  1. Up your sample rate of who gets the survey. But remember, intrusive surveys will impact the performance of your website, so if you are going to show it to everyone, find a way of making it noticeable without obstructing the content that users arrive to see
  2. Incentivise your users to complete the survey. Making my favourite brand site better by giving my opinions will make me fill in data. Making some random website I’ve just visited better won’t, but a free iPad draw may do. Be wary though, if you are requesting personally identifiable information (to tell them whether they’ve won an iPad) then you need to store it safely and remind the users why you are collecting it and what you are doing with it. Watch out for those who are just after prizes.
  3. Segment your visitors and only give the survey to those who are relevant to you. One of the greatest benefits of A/B testing and personalisation tools is that they’ve made online survey tools start offering this functionality. If your tool doesn’t do this through the front end interface, then your developer will be able to make it do it through JavaScript and cookies. These segments might be returning visitors, search engine visitors, visitors who’ve viewed the site for more than 10 seconds, visitors who have viewed two or more pages, visitors who have viewed a page. Remember this when you run your analysis though, because you will have inherent bias.
  4. Make your surveys more noticeable and clickable. If you don’t state on the pop up or overlay how many questions your survey is, I won’t start it because I don’t want to be stuck there for 15 minutes. If it looks boring, I won’t fill it in. If it pops up over the content I’m trying to look at, I will close it. If it pops up somewhere obscure I won’t see it. Each site is different, so each survey should be too. Apply the same duty of care to your survey as you would with a web page – test it with some users and see what they think.
  5. Make your surveys short. ‘No’ should be your favourite phrase for anyone asking to add questions. Can I add a question asking what colour the logo should be? ‘No’. Can I add a question to see if people like the fluffy toy I gave away outside the train station? ‘No’. Work out what questions you can only get from an online survey and run them. Use other methods to answer the other questions (email surveys, hiring a panel, or stand in the street asking random people as they walk).
  6. Run your survey during a time of high traffic. The person who doesn’t want to run the survey whilst all those people are looking at the site is wrong. Run it at a time of highest traffic will allow you to get the required responses quickly and then you can turn it off. If you run it at a time of low traffic you may end up running it for so long that the responses from the people at the beginning are living in a different world to those at the end.
  7. Link up the survey data with your Analytics data. This will give you the most value out of your results because you’ll add extra dimensions to be able to segment by. You have different conversion rates for different campaigns, so you will likely have different satisfaction too. Don’t lump all site visits together.

Remember that online survey data is just another data set for your armoury when you do optimisation. Don’t run it once and forget about it. Make it part of your continuous improvement process. If you’re looking for optimisation advice, please feel free to get in touch.

This post was written for Blue Latitude by optimisation expert Alec Cochrane.


Free guide:

How to avoid data paralysis

Find out how you can better manage your multichannel data, in our guide to avoiding data paralysis, by Alec Cochrane, Blue Latitude’s Head of Optimisation.

Stop drowning in a sea of data – learn how to overcome your multichannel data challenge.

Leave a Reply