How to Take Charge of your IVR Surveys


IVR surveys can be a great way of collecting customer feedback, but often suffer from poor response rates. Our panel provides some great advice on how to design your IVR surveys better.

Ask questions the right way round

Ensure that the customer experience is a simple and positive one. For example, ask ‘How pleased were you with the service you received today?’ rather than ‘On a scale of 1 – 10, how disappointed were you with the service that you received today?’ Each selection that the customer makes should be confirmed back to them to ensure accuracy.

Test, test and test again!

All questions must be well thought-out and drill straight to the point so as not to waste customers’ time. Then test, test and test again! Make sure that there are no areas of confusion or opportunities for the survey to go wrong.

Act on the results

Laura Campos

Analyse and, most importantly, act upon your IVR survey results. There is no point doing a survey if your findings don’t positively impact customers going forward. They will be much more willing to complete a survey if they can trust that their opinions are valued and considered.

Ultra recommends that an IVR survey lasts no longer than 3 minutes, including an informative introduction and a thank-you message.

Laura Campos, Marketing Associate, Ultra Communications

Include hang-ups

No matter how well-designed your survey is, many respondents will hang up before completion; perhaps the kids are playing up, or perhaps they’ve just run out of time or patience. This has several implications for the design of IVR script flows:

  1. Make sure that ‘respondent hung up’ is a valid outcome for any question. If left unhandled, any gathered data could be rendered useless.
  2. Decide how respondent hang-ups should be followed up. Do you call back later? If so, best practice is to offer a way back in where they left off. If not, decide what to do with an incomplete survey. Dump it? Add it to the data pool? Perhaps it depends on the quality of the gathered data, which leads us to…
  3. Ask the important questions with closed answers early on. This gives a greater likelihood of any incomplete surveys still being statistically valid and useful.

Bottom line: maximise the usefulness of any and all captured data by designing flows carefully with the expectation of some hang-ups.

Garry Pearson, CTO, Sytel

Keep it simple

Design your IVR survey to be easy to complete and understand. Don’t try to capture too much information as you may see callers getting bored or running out of time. Consider asking the caller to ‘opt in’ to receiving a further call if they would like to offer additional feedback, as you can always ask for additional information later should you choose.

Involve your agents

Consider suggestions from your agents when designing your surveys. After all, it’s likely that they will be measured on their completion rate. Think about offering incentives for agents in order to encourage surveys to be completed by not just measuring the agent in terms of total satisfaction ratings for their calls, but also in terms of volume of feedback received. This assists in avoiding ‘cherry picking’ of feedback and can really help in terms of response rates. It can also drive your agents to achieve the best results.

Consider other channels

Sian Ciabattoni

Sian Ciabattoni

Remember that a customer may not wish to complete a survey at that specific time, so offer other potential channels for feedback such as a website or via text. If you try to adopt a single-channel approach then response rates are likely to suffer.

Sian Ciabattoni, Marketing Director, Noble Systems EMEA

Don’t ramble at the start

One of the basic rules in getting customers engaged is to interact with them fast. The longer they wait, the bigger the chance for them to lose patience and give up.

Don’t waste time at the start of the survey giving out too many details, save them for the relevant questions or for the end of the call. Give helpful hints and information when the caller is having problems, not at the start of the call. Bombarding him with facts and figures at the start will make him believe the journey ahead will be difficult.

Don’t give more than four options per question

The human short-term memory can typically manage four different items and might find it difficult to remember the initial options at the end of an enumeration of possibilities. Keeping your surveys brief and simple will avoid caller frustration, as it can be very irritating to try and remember first answer options in long questions.

In addition, being consistent is essential for making the process easier. For ease of use, keep all the questions of a similar type together. The caller does not want to go from range type questions to yes/no and back to range, it is too confusing.

Use a consistent scale

Don’t ask for a rating between 1 and 4 for one question and then for a rating between 1 and 6 for the next, as this makes it difficult for callers to remember how they should score each question. Also make sure you keep the same scale indicating satisfaction.

Always offer the option for the information to be repeated in case the caller missed any important details. Also, guide the caller to a person if they are struggling or offer the option of a call-back.

Give the number after the option

Give the number after the option – for example ‘For this press 1’ NOT ‘press 1 for this’. By mentioning the answer/option first, you can grab the listeners’ attention and offer them a chance to process the information before offering the solution and pressing the number.

Pleasant voices increase conversion rates

As in a regular conversation, a pleasant voice on an automated survey can influence the rate of responses received and the number of abandoned calls. Changes in pitch and timbre are distracting, and text-to-speech technology should only be used if absolutely necessary. On the other hand, don’t try to sound like a person as ‘Hi, I’m Melanie and I’m going to ask you to press some buttons’ might sound too surreal.

Always allow callers to use verbatim comments

To make sure that all customers, regardless of their experience with the company, agree to be part of the survey, make sure you mention the importance of varied opinions for future customer service developments. To maximise the number of responses, ensure that key questions about overall satisfaction and other important issues are asked at the beginning of the survey. Also, always allow callers to leave comments, as this helps to satisfy those who have an issue they are eager to express a view on.

Don’t allow middle ground

An odd number of answers might not be a good idea. Human nature makes us pick the middle ground. In other words, don’t let the caller select ‘average’ for every answer. Make them go towards ‘good’ or ‘bad’ by taking out the neutral answer option i.e. 1- very good, 2- good, 3 – not very good, 4 – poor.

Tony Norman

Mention it’s not a sales call

It may be important, depending on the context of the survey, to state clearly at the beginning that it is not a sales call, that participation is voluntary and that you are NOT trying to sell the caller something, but simply seeking their input. Since an IVR survey is automated, there is a greater chance for drop-off than with an interviewer-administered survey.

Tony Norman, Solutions Architect, Netcall

Don’t allow agents to skew the results

It’s important not to let agents choose which customers get to take an IVR survey, as the results will inevitably be skewed. Instead put everyone into the survey – providing you’ve let customers know there’s going to be a short survey at the end of the interaction. It’s better to alert customers at the end of their query – at the start it can annoy customers, particularly if it gets in the way of resolving their queries.

Paul White

Paul White

Always be clear with your customers about what’s going to happen. Keep your IVR surveys short – limit to three questions only – and make sure they don’t last more than around 45 seconds. Don’t subject regular customers with the same surveys – you can vary questions using random number selection. And don’t forget to reward your customers by incentivising them to take part – you can do this by entering them in a free draw.

Paul White, CEO, mplsystems

Timing is critical

The key to achieving high response rates, as well as accurate and actionable insight, is requesting feedback as close to the point of the interaction as possible (and ideally in real-time) whilst the memory of the experience is still fresh in the mind of the customer.

Use their preferred channel of communication

Tom Lynam

Tom Lynam

It is also crucial to request feedback via their preferred channel of communication whether this is SMS, email, web chat, phone or social media. For this reason, many organisations are moving away from traditional IVR surveys in favour of more comprehensive Voice of the Customer (VoC) solutions.

Tom Lynam, Marketing Manager VoC Solutions, NICE Systems

Do you have any advice on increasing IVR survey success?

Please leave your comments in an email to Call Centre Helper

Published On: 9th Oct 2013 - Last modified: 2nd Jun 2017
Read more about - Call Centre Management, , , , , , , ,


1 Comment
  1. Our clients normally achieve around a 92% survey completion rate and a 20-25% survey acceptance rate, which is among the highest response rate of ANY survey method that does not have an incentive. IVR can often achieve a significantly higher participation rate than email, web and especially SMS surveys.

    Alan 10 Oct at 11:56 am
Get the latest exciting call centre reports, specialist whitepapers, interesting case-studies and industry events straight to your inbox.