A Eurovision Song Contest ServQual

ServQual is a system for measuring things which cannot be measured.

Actually, it’s for measuring things which are difficult to measure because they do not fit neatly onto an ordinal scale. For example, it’s not easy to say that “this school uniform policy scores 7 out of 10”, or to score the standard of the service in your local supermarket, or to judge the quality of a song in the Eurovision Song Contest. What one person thinks is “7 out of 10” is not always the same as what somebody else thinks.

If only there was a consistent way to measure this stuff . . .

Firstly, let’s explore what ServQual does, and then work on the example of the Eurovision Song Contest.

How ServQual Works

The common element of evaluation systems that try to measure these things will often have tick boxes like this:

strongly disagree disagree slightly disagree neutral slightly agree agree strongly agree

And, what’s the difference between an “agree” and a “strongly agree”? Is a “strongly agree” worth twice the value of an “agree”?

There are seven boxes at the top of the sheet, should they be numbered 1 to 7 in turn? Is that 1 to 7 measurement really a fair representation of the sentiment? Is “neutral” really worth 4? What’s the solution?

The ServQual methodology provides a way in which users can evaluate services and products when the “value” is hard to measure

During the 1990s a joint venture (backed by UK Government NHS funding) involved PhD students in pharmacology at the Lahore Institute of Quality & Technology Management and PhD students at the University of Stirling. They built an evaluation and measurement system for things which do not fit an ordinal scale. Representations of ticks or crosses on evaluation forms provided a simple heuristic review, what Nielsen calls a “does it look right” test.

Taking the “does it look right” test one step further, and by using the X2 test (the Chi squared test – measures of association on a non ordinal scale) the PhD students’ more detailed findings were relayed back to the governing bodies (in support of requests for more funding). For example, the researchers found that users valued sympathetic human contact, stating:

“we identified a wide gap in the perception of a ‘perfect’ receptionist and a ‘current’ receptionist, and now we need to do more research on how to make receptionists smile more.”

The PhD thesis added:

“Improving this single factor will lead to the biggest improvement in the ServQual score, and hence the biggest improvement in stress reduction (i.e. in patient happiness) in clinical encounters.”

If you’re interested in A Level statistics look up the X2 test (the Chi squared test), it is truly fascinating and enlightening for number geeks. The rest of us will just settle for Nielsen’s “does it look right” test.

The Eurovision Song Contest

Your job now, is to work out how to evaluate a song in the Eurovision Song Contest. The “public vote” is a gut reaction thing, but as an official judge you and your fellow judges will want to have a robust system for evaluating the quality of a song. Something that is true and fair and can stand up to scrutiny from your critics. Something that assigns the right degree of importance across a number of attributes. ServQual fits the bill.

What you need are a few “statements” which the judges can use as a basis for making their decisions. Something in the order of 8 to 12 statements is about right, and here are 4 to get you started. If you want to add yours to a word doc then you can download this basic proforma and add more detail.

In a perfect Eurovision Song Contest a perfect song would have:

1. a memorable melody, or tune, which you can remember and hum
2. a captivating first 10 seconds
3. a proportionate amount of canon fire
4. a touch of horror or shock tactics

In advance of the contest, each item is discussed by the panel of judges, and they may also be working with older ServQual docs form previous years. The judges collectively decide (for example) that “a captivating first 10 seconds” is rated a strongly agree and that “a touch of horror or shock tactics” is rated a disagree. We are talking here about a “perfect song”.

strongly disagree disagree slightly disagree neutral slightly agree agree strongly agree
1. a captivating first 10 seconds  




2. a touch of horror or shock tactics  




Half the job is done. They now have an indication of a perfect song in a perfect world (on a sheet with 8 to 12 statements). The next task is to score an individual song according to these statements, and they have a new, clean sheet with a blank matrix to help them do that.

In this particular case, this song has . . .

strongly disagree disagree slightly disagree neutral slightly agree agree strongly agree
1. a captivating first 10 seconds  




2. a touch of horror or shock tactics  




Once the “perfect song” and “this song” matrices have both been completed, a comparison can be made. The judges want to see a complete correlation in order to get a perfect song. With 8 to 12 statements it’s unlikely that anything will get to “perfection” but it could be close. The one with the closest correlation is the judges’ choice. Clearly with the examples above, the use of shock horror has moved this song away from the perfect song.

Number geeks may have already worked out that the judging sheets for Eurovision are then subjected to the X2 test (the Chi squared test) in order to get a scientific answer to the “correlation” issue.

And guess what? You can do this in business. Is your receptionist close to being a perfect receptionist, or way off target? Does your sales team get wonderful figures at the expense of upsetting customers? Do your managers manage effectively or simply annoy the staff? Is that why there are so many regular resignations and recruitment costs?

As Nielsen said “does it look right”? If you’re always getting one score way over on one side when it should be way over on the other side, then this is the statement that you should focus on most. Improving that one attribute will make the biggest improvement to your ServQual score. It’s basically the same whether you do the complex stats, or simply ask “does it look right”?

And that one thing wouldn’t happen to be time management would it?