Below are the answers to the questions we have been asked most often when we have talked to people about the new segmentation research.
If you have a question not covered here, you can contact us at
- What is segmentation?
- How was the segmentation conducted?
- How robust is the Taking Part survey?
- Who did the analysis?
- How did you define ‘arts attendance’ and ‘arts participation’?
- What are the segment profiles based on?
- Aren’t the profiles just an exercise in crude stereotyping?
- Where do the segment names come from?
- Why did you update the profiles in 2011?
- Why have the sizes of the segments changed between 2008 and 2011?
- Why didn't you update the postcode modelling?
Using the segmentation
- What can the segmentation tell me about audiences for particular artforms?
- Can I relate the audience of my own arts venue to the segments?
- Can I found out the segment profile of my own region, Local Authority or catchment area?
- How does this research relate to existing segmentations – ACORN/Mosaic and box office specific segmentations?
- Isn’t it easier just to look at and target people by lifestage or social group, for instance young people?
- What about young people aged under 16?
Segmentation is a market research method where a given market is broken down into distinct groups of people that behave in similar ways or have similar needs. An organisation can use a segmentation to better understand its market, to identify groups that it would like to target and to develop products which anticipate their needs better.
Our new arts segmentation aims to do this in the ‘arts market’ by identifying distinct arts consumer types and describing their characteristics. This information can help us develop more effective and targeted audience development and marketing plans by giving us insight into the types of approaches that people in different segments are most likely to be responsive to.
Of course, as the segments are ideal consumer ‘types’, not every individual in a given segment has identical behavioural or attitudinal patterns and nor will they respond in exactly the same way to the same arts offer. But within each of the segments, definite trends can be observed: amongst each segment some socio-demographic and lifestyle features are particularly prevalent. The segment profiles draw out key features of actual past and current behaviour. They give a good indication of the types of initiatives or marketing routes that are more likely to be successful among a particular segment.
back to top
Statistical analysis of 2005/06 Taking Part survey data was used to break the population into a variety of groups who share similar characteristics. A number of different factors were used to determine membership of the different groups. These included:
- arts attendance – the first factor we examined was arts attendance – distinguishing between those who had attended an arts event at least once in the 12 months prior to interview and those who had not. Those who had attended were then divided into six groups using cluster analysis, based on frequency of attendance at different types of arts events. Therefore each of these groups attend similar types of arts events at a similar frequency.
- arts participation. Among those who had not attended an arts event, the next factor that was examined was arts participation. The sample was divided into two groups, those who had participated in at least one arts activity in the 12 months prior to interview and those who had not. The group who had participated in the arts was then divided into three groups, according to age (16–29, 30–59, 60+).
- barriers to arts engagement. Finally, the group who neither attended nor participated were divided into four groups using cluster analysis, based on the reasons they gave for not attending or participating in the arts. Therefore each of the unengaged groups have similar barriers to engaging with the arts.
The segment profiles were updated in 2011. The variables that determined membership of each segment in that update were slightly different from that in the original segmentation exercise. For more details see the technical note.
back to top
Taking Part survey is conducted by TNS, and is very robust in methodology. The survey sample is carefully designed to generate a representative sample of English adults aged 16 and above. Respondents are selected from the small-users Postcode Address File, with the sample restricted to private households (ie businesses and non-private accommodation such as student halls, care homes and prisons are not included in the sample).
In the first year of the survey (fieldwork period July 2005–October 2006) – the data used in the original segmentation analysis – the achieved sample size was around 28,000 (a half sample of around 14,000 respondents was used for the segmentation analysis). The sample was stratified by region (at least around 2,450 core sample interviews per region) and by population density and the proportion of residents classified as managerial/professional/full-time students. In addition to the core sample, to enable robust analysis by ethnic subgroups, a booster sample of around 1,500 non-white adults was carried out. The Taking Part survey from 2008/09 used for the update of the segment profiles in 2011 had an overall sample size of 14,452. It followed similar methodology as the 2005/06 year. Visit the DCMS research pages for more details about the Taking Part survey.
As for the actual data collection, to maximise response rate, prior to the interview calling at the address a letter and an information leaflet about the survey are sent to the selected household. If more than one adult live in the household, the interviewer will make a random selection. After selection, no substitutions are permitted. The survey itself is conducted in computer assisted face-to-face interviews in the respondents’ own homes, lasting an average of 40 minutes. All questions in the survey have gone through careful piloting to ensure that people understand clearly what is being asked.
Finally, all the data are weighted to account for the unequal probability of selection (based on number of adults living in a household, over sampling in smaller regions and the selection of multi-household addresses) and to adjust for different rates of non-response.
However, as with all population estimates based on survey samples, Taking Part estimates are associated with a degree of statistical error. In addition, even in the case of a large-scale survey with a good response rate, it should be borne in mind that there is a possibility that the respondents or their responses are not representative of the actual patterns among the whole population.
back to top
The original segmentation analysis and the segment profile update in 2011 were both undertaken by our contracted supplier TGI Insights & Integration (Kantar Media), a leading market research agency, in consultation with researchers at Arts Council England. The profiling of the segments was conducted by TGI, with some additional Taking Part survey analysis conducted by Arts Council England researchers.
back to top
‘Arts attendance’ refers to going to any of the following 17 categories of events:
- exhibition or collection of art, photography or sculpture
- craft exhibition
- video or electronic art event
- event connected with books or writing
- street arts or circus (not animals)
- culturally-specific festival (eg Mela, Baisakhi, Navratri)
- play or drama
- other theatre performance (such as musicals, pantomime)
- opera or operetta
- classical music performance
- jazz performance
- other live music event
- contemporary dance
- African people’s dance, South Asian and Chinese dance
- other live dance event
‘Arts participation’ refers to doing any of the following 17 activities:
- other dance not for fitness
- rehearsing/singing to an audience
- rehearsing/playing a musical instrument to an audience
- playing an instrument for own pleasure
- writing any music
- rehearsing/performing in a play or drama
- rehearsing/performing in an opera
- painting, drawing, printmaking or sculpture
- photography as an artistic activity (not holiday snaps)
- making films or videos as an artistic activity
- creating original artworks or animation using a computer
- textile crafts such as embroidery, crocheting or knitting
- wood crafts such as wood turning, carving or furniture making
- other crafts such as calligraphy, pottery or jewellery making
- writing stories or plays
- writing poetry
Arts attendance does not include going to the cinema and participation excludes reading, because they are very popular activities and adding them in the segmentation analysis does not help us to further differentiate between different people. But we do recognise that these activities are also good indicators of an individual’s level of cultural activity. For this reason, we have examined the levels of cinema attendance and reading across the resulting segments and drawn on this data for the segment profiles.
back to top
The segment profiles draw on a vast range of data from the Taking Part and TGI (Target Group Index) surveys. The data include information on each respondent's socio-demographic characteristics; free-time activities; media consumption; consumer preferences; and the extent to which they agree or disagree with a broad range of attitudinal statements covering such things as attitudes towards the arts, creativity, other cultures, their job and career, family, eating, cooking, drinking, money, the environment, trying new products, fashion, cosmetics, availability of time, current level of happiness, satisfaction with life, ambition, ability to effect change.
As the segments were defined based on their arts-related behaviour and attitudes, the segment profiles are the most distinct in these areas, but some clear distinguishing trends can also be observed in terms of all these other features. As the range of source data in the two surveys is immense, the segment profiling is necessarily selective: the profiles pick out features that are characteristic of people in a given segment, instead of providing a comprehensive description. Please note that 'characteristic features' in many cases only apply to a small overall proportion of a segment - they are characteristic in the sense of being more prevalent among the segment than average.
back to top
To a certain extent segmentation is always a kind of stereotyping, because the objective of the exercise is to spot patterns across individuals and make some generalisations. But this is the trade-off we make in order to gain insight that allows us better to make sense of the world and develop more effective, targeted strategies.
From previous research we know that the patterns of arts consumption are not entirely random: people’s arts engagement is strongly patterned and those with similar lifestyles, upbringing and social circumstances are also more likely to have similar levels of involvement with the arts. Thus, while recognising that on a micro level each individual is unique and that there is an infinite number of factors that influence an individual’s level and nature of arts consumption in a given time period, it is possible and useful to examine the broad patterns of arts engagement and how they relate to other aspects of people’s lives.
This is what the segmentation aims to do, by identifying and profiling distinct arts consumer ‘types’. As the segments are ideal consumer ‘types’, not everyone in a given segment will have an identical pattern of behaviour and attitudes and will respond in the same way to a particular arts offer. But as the profiles draw out key features of actual past and current behaviour, they give a very good indication of the types of initiatives or marketing routes that are more likely to be successful among this segment in particular, as compared with other approaches. Consider for example yourself, or your existing/target audience group, in terms of the 13 segments. While you may identify relevant aspects in a number of the segments, there is likely to be one segment that clearly offers the best fit. You or your existing/target audience will not have all of the same characteristics as the segment, but there will be many features that you can relate to, and the rich profiles offer a lot of useful ideas and information for appropriate activities, venues, timings, tone and approach, methods of communication etc.
back to top
We wanted to give each segment a short descriptive label to act as an illustrative shorthand for referencing and quickly recalling the segments in communications. The names are based on the data we have on the segments: they highlight some characteristic features of their arts engagement, other leisure-time activities or lifestyle. While the naming process of course involved lots of subjective choice also, care was taken to ensure that the labels are not misleading or contrary to any data available. Like the segmentation profiles overall (see the previous FAQ), a given segment label is likely to fit some respondents assigned to a segment better than others.
back to top
People's consumption habits change over time, as do culture and commerce. After supporting the use of the Arts audiences: Insight segmentation for three years the Arts Council felt that the profiles of the segments should be refreshed and updated in order that they better reflect people's contemporary lives. The segment profiles had become out of date in a number of specific ways: some brands were no longer fashionable among certain segments, a number of retailers had gone out of business and new products and services were missing from the old profiles. Therefore we commissioned TGI Insights & Integration (Kantar Media) to update the segment profiles using more recent survey data (the 2008/09 Taking Part and the GB TGI NET 2010 Q3 surveys). For more informationabout how this was undertaken see the technical note.
The focus of the profile update was to ensure that new products and services were reflected in the segment profiles, this included the digital behaviour of the segments, therefore ensuring that cultural policymakers, programmers and marketers would have access to the latest data on the consumption patterns of the segments.
When attempting to recreate the 2008 segments using more recent data from the Taking Part survey, there was a change in the size of the some of the segments (for example Older and home-bound went from 6% to 11%). It is likely that a combination of three factors is at work: that there has been a small change in the level and character of arts engagement by respondents in the Taking Part survey between the years 2005/06 and 2008/09; the changes to the survey between those years (especially the removal of variables used to determine segment membership in the original segmentation) will have changed the way in which respondents can report their arts engagement; the use of an algorithm to recreate the segments in a subset of Taking Part respondents means that respondents from the 2008/09 Taking Part sample were not perfectly allocated to segments in the same way as those from the 2005/06 sample. Therefore it is unlikely that these changes reflect real substantial shifts in the sizes of the segments between 2005/06 and 2008/09. In light of this we have not updated the regional and local analysis spreadsheets that denote a modelled distribution of the segments in specific geographic areas. However, we felt that it was important to report the changes in sample size in the new booklet since they reflect a change in the size of the respondent pool for each segment in the Taking Part data.
The decision was taken not to update the postcode modelling at the same time as updating the segment profiles. The focus for the update to the segmentation in 2011 was to refresh the segment profiles. Because the 2011 update did not undertake a fresh segmentation of the English adult population (but rather attempted to recreate the segments identified in 2008) we were conscious that a re-modelling of the data based upon the 2011 refresh would add a further level of error into the area profiles. Further reasons for not undertaking an update to the local level modelling were that we did not believe that the benefits to ourselves and the arts sector would outweigh the costs of the analysis; another consideration was that the character of postcodes used in original the calculation of the geodemographic modelling had not changed to the same degree as the consumption behaviour of the segments, and hence the original postcode modelling had a longer 'shelf life' than the segment profiles.
The segmentation can tell you which segments are currently the most likely to attend various events or activities, and which segments show some interest in attending and participating at present. This information can be accessed either by studying the individual segment profiles or from the 'arts engagement' data tables.
In addition to the segments’ relative propensity to attend different events and participate in different activities, when estimating the overall importance of the segments among your audience consider also the segments’ relative size and the relative frequency at which they attend. For example, even at moderate levels of engagement, a relatively large segment (eg Dinner and a show) might make up a considerable proportion of all those people who go to that type of event, and segments that tend to attend particular events more frequently (eg Traditional culture vultures for opera and ballet) are likely to account for a relatively larger proportion of overall ticket sales in those events.
back to top
You can gain some insight on the likely segment profile of your audiences from the national engagement patterns among the segments – see the previous FAQ. You can also gain insight on the likely segment composition of your audiences by relating any additional information you may have on your own audience (eg age profile, family structure, other interests) to the characteristics of the 13 segments.
In addition, we have developed a geodemographic modelling of the segment data. This modelling allows organisations with postcode data on their audiences to gain some broad insight on the segment composition of their audience, by analysing the characteristics of their neighbourhoods (postcodes).If you are interested in this analysis (minimum 1,000 postcodes required) please contact Audiences UK or your nearest audience development agency. There is a standard charge for this analysis.
back to top
Breakdown of the populations in each English region by the segments can be accessed from the Understanding regional populations data table
We have also worked with CACI Ltd to develop a local level modelling of the segments. This modelling allows us to estimate the segment profile in a given local area, for example a Local Authority or a 30-minute drivetime area. It also allows us to estimate the relative distribution of a given segment across a selected area. For more information on this see the Local area analysis page of the website.
back to top
How does this research relate to existing segmentations – ACORN/ Mosaic and box office specific segmentations?
Different segmentations are suited for different purposes and have different strengths and weaknesses. This arts-based segmentation is therefore not intended to replace existing tools that are already being used successfully. But it can be used an additional tool, alongside other sources of audience intelligence, wherever relevant.
It might be particularly suited for:
- strategic positioning – putting the audiences of individual projects, organisations or artforms in context
- developing initiatives designed to increase public engagement – it provides detailed information on the qualities and needs of different audience groups, including current non-engagers and audiences of non-ticketed events
But it doesn’t have all the answers. For example, it can’t:
- tell organisations who exactly is and isn’t engaging with their work, or provide other detailed information such as booking history, average spend, etc – box office specific segmentations are the best placed to provide this type of detailed information
- be used as a robust tool for detailed monitoring of changes in audience patterns over short periods of time
- yield data suitable for generating direct mailing lists – to access actual names and addresses to target, you should use box office data or ACORN/Mosaic services
Contact Audiences UK or your nearest audience development agency if you would like to know more about the limits and possibilities of this research and how you might be able to use it.
If you wished particularly to focus your activities on for example a particular age group or those with young children, the best way to profile your target audience would indeed be looking at raw data (eg Taking Part) on characteristics and cultural participation patterns of these demographic groups.
However, making strategic decisions based on such analysis assumes that people’s behaviour is predicted to a large degree by a single demographic feature. In reality, a number of socio-demographic factors, attitudes and interests all play a part. This is why we developed the bespoke arts-based segmentation of consumer types. The segments, while less ‘clear-cut’ in demographic profile, provide a tool for more effective and tailored strategies for engagement that take into account people’s socio-demographic profile as well as their arts preferences, overall lifestyle and attitudes. Such strategies are more likely to be effective in engaging a target segment than those that only consider the characteristics of their demographic peer-group as a whole. For example, Dinner and a show and Mature explorers segments have similar socio-demographic profiles, but very different arts engagement and broader lifestyle profiles and are likely to respond to very different kinds of engagement strategies. Or, the vastly different profiles of three segments each with high proportion of parents – Family and community focused, Mid-life hobbyists and Time-poor dreamers – make it evident that not all parents will respond to similar motivational messages and opportunities.
back to top
Due to the sample definition of the Taking Part survey, our segmentation covers only English adults aged 16 and over.
It might be possible in the future to explore the consumer types among children aged 11–15, however, using data from the Taking Part Child Survey which has been running alongside the adult survey since 2006.
back to top