Voice of the Customer – Good, Better, Best

Earlier this year Microsoft included new Voice of the Customer functionality in its Dynamics CRM 2016 Online product.  Having previously worked in Customer Insight for several years the thought of having Research fully integrated into CRM was truly exciting.  So, in advance of my presentation with Andrew Bibby at the CRM User Group, I thought I’d share top tips for successful research.


1) Start with the end in mind

For each question you should be able to answer ‘what will I do when I have the answer?’. Be really strict to separate the truly insightful from the merely interesting; this will help keep your questionnaire from being too long. Having research capability built in to your CRM mean that you can source demographic and transactional data directly from within CRM itself rather than having to discover these via a questionnaire.

2) Target a representative group

If you’re not inviting everyone within a target group then make sure that those invited are representative of the whole. For instance, if your intention is to find the opinion of customers attending recent marketing events then you could invite every attendee. However, things get more complex should we wish to exclude people who have been recently invited to partake in other surveys.  When these are excluded from this survey then will those invited still be representative of those attending recent marketing events?

3) Make it interesting

With so many auto-generated surveys it’s all too easy to ignore them; and as companies increasingly drive research from CRM we are likely to be invited to partake in even more research. To help prevent our research invitations from being ignored we must deliver them with a compelling message and design; explain who you are, why you want their input and what you will do when you have the results, and above all make the message clear to understand. These things help achieve a higher response rate.

4) Review your respondents

Take a good look at the profile of people who completed the survey; are they representative of the target group? Check that you have a good spread of participants, from different demographic groups and people who have roles with diverse experiences. And if not, then consider the impact of this when reviewing your findings.

5) Send the invite at the right time

Much like email marketing, the timing of a research invitation is important. Should the invitation be send shortly after the event, such as closing the Service Case, or should these be batched and sent at a specific time? Think about your customers, when will they be most responsive to receiving the invitation?

6) Be impartial

Be balanced and non-judgemental in how you phrase your questions. Don’t “lead the witness”, for instance “Some people think CRMUGUK is the best CRM conference, what do you think?” is not an impartial question, instead ask “What do you think of the CRMUGUK conference?”, or “In your experience how does the CRMUGUK conference compare to other CRM conferences”.

7) Make it easy to respond

People should be allowed to respond to your question with an answer that represents their opinion. Consider all the possible answers, not just the most common responses, not forgetting responses such as “I don’t know” or “I don’t want to answer”. And ask open questions such as “why do you say that?”.  While such open-ended responses are harder to translate into percentages or charts it will ensure that you have the breadth of opinion. Responses to open questions will reveal opinions you have not previously considered and are therefore a great help in designing the available responses for your next survey.

8) Odd or even scale

Where you have a question with a scaled response, such as a Bad-Good, Sad-Happy, etc then you need to decide whether an odd or even number of response categories is more appropriate. Odd numbers are good when respondents could be truly neutral or indifferent, although this may encourage lazy neutral responses instead of directing the responder to consider whether they’re slightly more positive or negative towards the statement in the question. Even numbers are required if you want to eliminate that kind of fence sitting.

9) Keep it simple

A weakness of conducting research without an interviewer is that questions are more likely to be misinterpreted. Plain and unambiguous language is important as we need to minimise that occurring. If the question is misinterpreted then responses aren’t useful.

10) Don’t over analyse

Presentation of results as charts and statistics needs to be done with care, by all means be selective in highlighting results but do not distort the overall findings. Research that disagrees with existing business assumptions can be far more insightful than that which supports those assumptions. And be careful in the conclusions you make: correlation does not imply causation, the popular press would do well to remember this.