Retail Analytics Micro Study

Matt Duffy FCCA

Matt Duffy takes a dive into Jersey’s retail sector data, see what we found even with just a short data model and prototype…

Stimulated by recent concerns around the local economy in respect of the retail market and the, seemingly, ever-increasing shift to online shopping, we thought we would see what we could learn from some of the States Statistics Unit’s datasets (a whole host of them published on One in particular caught our eye and seemed like it might be useful: the quarterly retail sales survey. As it happens, this survey ceased at the end of 2015, superseded by the Business Tendency Survey.

However, for the purposes of illustrating some of what we could do with Power BI and OData sources (the sources are OData), we connected to the Retail Sales Survey data, undertook some exploration, and then developed some visuals in Power BI to illustrate some of Power BI’s capability/features and offer viewers some food for thought in respect of what’s happening to retail sales in Jersey.

To help provide a little more context, we also pulled down the RPI dataset (again, OData). The actual content of the surveys are unknown to us – we are only working with the index output compiled by the States Statistics Unit. For the retail sales data, index 100 was at end of Q1 2007 (with data ranging quarterly from 2005 through 2015). It was necessary to rebase the RPI dataset to synch up (its base 100 was June 2000).

The outcome is a summary report illustrating:

  • Overall, retail sales through the 11 year period 2005-2015 grew at 2.7% per annum - although 2007 to 2015 not so much (1.5% per annum)
  • Predominantly* Food – grew at 3.8%/annum in whole period
  • Predominantly Non-Food – grew at 1.5% - although from 2007 onwards it shrunk by 0.6% per annum

RPI comparators are included, but we are cognisant that they are not directly comparable – Volume can drive Value of course.

  • However, the data indicates that Food has grown in sales value slightly ahead of RPI for Food and this compares nicely on the Volume page – so helping reassure that the data is sound (i.e. Predominantly Food is, really, predominantly food)
  • With the relatively level Volume of Food sales, we can fairly confidently say that the Value of Food sales is increasing due to price through the period
  • Food sales Value increase is offsetting a decline in Non-Food sales

So our conclusions (not answers, only hypotheses) are that Non-Food value is falling (fact) probably due to increasing trend of internet sales (hypothesis). We would like to explore some data, if available (and we will search for it) on internet sales levels. But, for the time being … it’s a hypothesis, not a fact.

* Predominantly meaning the survey’s precision in respect of sales was not perfect – as you would imagine, retailers selling goods of different types (food/non-food) would need to work very hard to provide perfect data.

By clicking “Accept”, you agree to the storing of cookies on your device to enhance site navigation, analyse site usage, and assist in our marketing efforts. View our Privacy Policy for more information.