In the past we have converted a few telephone scripts to online studies (and also host the data-capture for telephone surveys) and there are often a few items we can change in a telephone survey script to make the survey easier to complete online. (not always). Some questions are also easier to answer as a written question, for example if you’re asking “how satisfied are you with …” and then present a list of attributes, say:

- Ambience inside clinic
- Punctuality of Osteopath
- Thoroughness of case health history
- Thoroughness of clinical examination
- Clarity / helpfulness of diagnosis

Then the respondent can race through these questions faster than an interviewer can read aloud. (not always). That doesn’t mean we should cram in more questions though.

Thanks for sharing!

There was one survey I still remember, about postage stamps. It opened up by asking where the respondent bought postage stamps – post office, supermarket, petrol station etc. I used to PRAY that they’d only say one or two places – if they said 5 or 6 then we’d have to go through the same hideous set of about 80 questions for each place, it was mind-bending, with respondents wanting to bail out etc.

So I agree, keep ‘em short…

http://www.peterbennett.net

Best wishes
Peter

Excellent info. I was introduced to a similar approach when I worked for TARP (another research company) at the end of the 90′s – albeit using just the 1 survey to examine both customers’ satisfaction and purchase intention, willingness to recommend, etc. This is common in a lot of customer surveys today – I imagine the very first time someone dared to ask both types of questions in the same survey they perhaps thought they had stumbled upon an amazing ability! Often, the simplest ideas…

I’m particularly intrigued by the cost of service levels and what it takes to convert a ‘mollified’ customer into a very satisfied customer. Definitely agree that a lot of the steps needn’t cost the organisation money.

Thanks for sharing.

Xerox did two separate surveys, one about customers’ perceptions of the service they received (this as part of an employee reward scheme) and one about customers’ intention to buy again (this as part of a sales forecast). Neither department knew of the other’s survey, but they happened to survey the same customers. Eventually, it was realised that they had asked the same people and someone suggested they overlay the data to see if there was a correlation between satisfaction and repurchase.

They found that the people who very, very dissatisfied (4/5) were very, very likely to buy again and likely to tell others about the great service. What was surprising was that customers who were quite satisfied were scarcely more likely to buy again than customers who were quite dissatisfied. Customers who rated service above 4 out of 5 were six times more likely to buy again than customers who rated service between 3 and 4.

We looked into this in more detail and used it as the basis of Lotus Development/IBM’s award winning Customer 2000 project. We also found that while total costs increase as companies improve service levels from 1 – 4, they fall significantly as companies climb from 4 towards 5. This seems to be because to get to level 4 the basic infrastructure must work, but at the highest levels service is about engagement, attitude, pride, friendliness, personability, which are all the outcome of peoples responses to good leading and as priceless as they are ‘free’.

Results of such customer responses’ include free disclosure of valuable information about feelings and intentions, customers’ willingness to make long commitments that reduce costs and customers effectively selling on behalf of the company.

Before and since then companies such as Apple, Amazon and Southwest Airlines have followed similar models, achieving the triple gold standard of increasing perceived service, revenue growth and greater profitability simultaneously year after year, while most other companies have found it impossible to do more than one (or even one) at a time.

I wasn’t really thinking of similar questions to the 2 audiences when I was writing this post, but this can be very useful to figure out whether employees (and the business) really understand what is important to their customers.

I think we are probably thinking along the same lines though…

We ask the engagement questions to our employees, and the feedback/experience questions to the customers. Then examine the results of the “higher-engaged” employees (by site/region) against the customer survey results also broken down by the same site/region. Do the same sites also have better customer sat. scores? Which sounds relatively simple!

Maybe a bit more complex statistically, multilevel models (http://en.wikipedia.org/wiki/Multilevel_model) may be a great way to overcome the limitations of correlating financial performance as a proxy for customer satisfaction with some measure of employee engagement because you can actually include an aggregated measure (maybe by site/ region) of employee engagement as an explanatory variable into a statistical model with customer satisfaction as the dependent variable.

You may also find the report “The Drivers of Engagement” published by the IES in 2004 of interest. It’s a study based on analysis of a survey of over 10,000 NHS employees (across 14 organisations in the NHS). This only looks at the drivers of engagement… no joined-up surveys in this study(!). The summary page http://www.employment-studies.co.uk/pubs/summary.php?id=408 lists some of the key findings such as engagement levels decline as length of service increases.