OPINION7 January 2021

The future of geodemographics

Covid-19 Innovations Opinion Public Sector Trends UK

Geodemographics has been a useful approach for researchers for more than 40 years. Yet changes to how the census operates could have a profound impact on the industry. By Barry Leventhal.

Blurred people

Geodemographics – the analysis of people by where they live – took the UK research industry by storm at the 1979 Market Research Society (MRS) conference and emerged as an industry in its own right during the 1980s. Arguably, it was the ‘big data’ analysis of the time. However, where is geodemographics now, how much has it changed and where is it heading?

We hear little about geodemographics nowadays because its value has long been proven and its uses are well-established. The systems are embedded in sample survey operations and in business applications such as site location and customer marketing. 

Geodemographics has evolved over the last four decades by retaining the essential concept and taking advantage of beneficial developments: what has stayed the same and what has changed?

The essence that has remained unchanged is to classify or segment small areas into types, according to the characteristics of their residents. Techniques such as cluster analysis are employed and the end result is a set of segments to which all areas in the UK are allocated. The successful changes in geodemographics come down to advances over the decades in two main areas – data and technology.

The primary data source used in geodemographic systems is the census of population, in particular the results for output areas, the smallest geographical units available from the census. For the early systems in the 1980s and 1990s, each output area typically contained around 180 households and so could include a diversity of residents. From the 2001 census onwards, output areas were redesigned and reduced to around 125 households, thus ‘sharpening’ the profile. 

By combining with sources, such as the Postcode Address File, the Land Registry and other open data, classifications could reflect measures of consumer activity and would discriminate at unit postcode level. Thus, geodemographics became more targeted, while still including neighbourhood characteristics obtained from the census.

Another significant change, also made from 2001, was that the payment of royalty charges on census output was discontinued and access to most outputs became free and unrestricted. This widened the use of census data in the private sector, increasing the benefits gained. Smaller agencies were able to enter the geodemographics marketplace, which helped to grow range and creativity of the approaches available. Currently there are no less than eight general-purpose classifications in the UK together with approaches for specific industry sectors, all largely driven by the 2011 census.

The obvious advance in technology – the exponential growth in processing power – means that a classification that once took many hours to build on a large mainframe machine can now be created in a fraction of the time on a personal computer. With some analytical tools, the entire development process may be saved as a workflow and rerun at will, and therefore users can create more relevant classifications for themselves. For example, users could finetune a product to employ the most relevant variables or customise the geographical area it covers.

As we start 2021, where is the geodemographics industry heading?  The industry seems secure from a data perspective for the coming years as, despite the pandemic, the Office for National Statistics (ONS) intends to conduct a census in England and Wales this March. The questionnaires will be completed online where possible and contingency plans will be in place to mitigate for Covid-19 in situations where face-to-face fieldwork is required.

Similarly, in Northern Ireland the census will also take place in March. However, in Scotland, the census has been delayed by a year to March 2022 due to the impact that the pandemic has already had. Users such as geodemographics agencies and their clients have yet to understand the implications of this split timing for UK-wide census outputs.

Of greater concern, the forthcoming census could be our last ever decennial population survey. For the last few years, the ONS has been researching the feasibility of gathering and processing administrative data – information from operational systems run by government bodies such as the Department for Work and Pensions, HM Revenue & Customs and the NHS – in order to create an administrative data census.

This would enable census-type results to be produced more frequently but much of the richness of information collected in the traditional census would be lost. This would threaten the very existence of geodemographics as they are currently understood, with profound implications for the products described above. It could also, through necessity, drive radical innovation in the solutions.

A decision on the future of the 2031 census will need to be made in the next three years and it remains to be seen whether the Census Offices will be ready to make the significant change to an administrative data census. We can all be certain that geodemographics users will be keenly awaiting the decision and taking part in the debate.

 Barry Leventhal is director of BarryAnalytics.

A MRS webinar on geodemographics will take place on 28th January. Details can be found here