Proposal for Audience Insight segmentation update 2011 – Lot 1 only Date: 21-02-2011 Reference number: ACE-2010-014 Josie Farnsworth, Russell Webb and Sandy Livingstone 020 8433 4090 [email protected] Project context In 2010, the Coalition government announced £6.2bn in spending cuts. This meant a reduction of central, and local, government funding for Arts Council England. The total reduction in 2010/11 is £23 million. A challenging economic climate does not change the central mission for Arts Council England, “Great art for everyone”. The goals of the organisation remain the same, with the core priority still to support and create artistic experiences “for as many people as possible across the country”. It is therefore still key for the Arts Council, and arts organisations , to continue to increase engagement and participation in the arts. Having information on where to focus effort and funding to achieve this aim is important. When funding is limited, it is ever more so. 2 Project context Even without the current economic and policy changes, it would be time to update as 4 to 5 years has passed since the original segmentation was completed. Having a segmentation that reflects current trends is vital. Over the last five years, the economic downturn, increased use of smart phones and social media, more large arts venues outside London and increased funding of digital arts are just some trends which may impact individual attitudes and behaviour. An updated profiling will provide current detail of audiences for the arts reflecting the impact of England's current fast changing social, economic and political environment. For all public funded bodies, the challenge is to deliver more for less. This project would provide valuable information to aid the Arts Council in understanding how best to use funds, by identifying and confirming targets and the communications with the most potential to engage them. 3 Project objectives The overall aims of the project are to: • retain the popularity and develop the use of the Arts Council segmentation • provide actionable profiles of arts audiences, those currently engaged and potential new audiences • ensure cost efficient use of resources is made • support arts organisations in driving further audience engagement Specifically, Lot 1 Segmentation and Analysis needs to deliver: • a refreshed segmentation for England o to re-size and understand changes in arts engagement for each segment • an updated profile of each segment o to include general characteristics, lifestyle, arts & culture engagement, media habits to include digital and arts broadcasting 4 Proposed approach Lot 1 – Segmentation and Analysis Stage 1: project inception and familiarisation using briefing and previous data Stage 2a: re-fresh segmentation to understand changes in size and engagement using Taking Part Year 4 (2009/10) Stage 2b: appending updated segments to latest TGI database ( TGI Net Q4 2010, July 09 – June 10) statistical linking of Taking Part & TGI Stage 2c: review algorithm with Arts Council and stakeholders statistical linking of Taking Part & TGI Stage 3: in-depth profiling of the segments using TGI, TGI Net and Taking Part variables 5 Benefits of our approach Unique access to TGI data • Kantar Media runs TGI, giving the unique ability to analyse and process the data at respondent level for statistical linking with other datasets • Enlightenment is Kantar Media’s specialist consulting arm and are highly experienced in using advanced analytical techniques to develop insightful consumer profiles from TGI data Impactful output for non-research audiences • Experience in presenting data and findings to wide audience base across organisations and their stakeholders, making information accessible and engaging. Knowledge of the sector and project • Previous experience working with the Arts Council and other public bodies means we are familiar with both the organisation and wider policy context. Creative rigour • We operate a transparent and open relationship with clients, and our solution is, as always, as creatively rigorous as possible. 6 Proposed approach: Stage 1 Stage 1: project inception and familiarisation using briefing and previous data Our approach is to work in close consultation with our clients. We will hold meetings at key milestones, send through weekly status updates via email and also suggest regular telephone contact with the Arts Council project team throughout the project. An initial kick-off meeting to understand the full context of the project and confirm detailed objectives, process and timings would be the first stage of the project. At the meeting we would also want to finalise the specification for the profiles. Following this meeting we would send through a contact reporting outlining agreed action points. 7 Proposed approach: Stage 2 Stage 2a: re-fresh segmentation to understand changes in size and engagement using Taking Part Year 4 (2009/10) We would re-fresh the segmentation* using the latest large sample Taking Part data. Rather than beginning from scratch (i.e. using the same input variables to a new cluster analysis) we propose to take the statistical definition of the cluster “centres” into the new Taking Part data and to put the new respondents into the new clusters according to their statistical proximity to these points. In this way we can better control the similarity of the new clusters to the old, ensuring the same type and number of segments as previously. Note that the cluster sizes may change in size and profile due to evolving arts behaviours but even so it would still be logical to apply, for example, the same cluster names. *Arts-based segmentation England, Arts Council England 2008 8 Proposed approach: Stage 2 (cont.) Stage 2b: appending updated segments to latest TGI database ( TGI Net Q4 2010, July 09 – June 10) statistical linking of Taking Part & TGI There are two possible statistical linking approaches, an algorithm or data fusion. In the original project there was a requirement to conduct TGI analysis against other Taking Part data than just the segments themselves so fusion was the preferred method. However, in this case our entire focus is on refreshing the understanding only of the segments themselves. A linking algorithm can be optimised for this purpose and so is arguably more robust. Being less complex it is more cost effective as well. Our proposal is based, and cost, on Option 1 (Linking Algorithm). Further details are given on the following slides. 9 The two possible linking approaches Features 1 2 Discriminant analysis to create a linking algorithm Data fusion Application The algorithm predicts a single variable (i.e. segment membership) using only TGI data Most often used approach where the need is to recreate segments only - derived from another quant survey Creates a modelled match between similar respondents on each survey so that all variables can be donated from one survey to another More rarely used: gives greater scope and control to the analyst, allowing profiling to be carried out on selections defined on either survey N.B. Our proposal is based and cost on Option 1 (Linking Algorithm) 10 Re-cap of the original clustering method High-level groups Clusters Group 1 Attend/ participate cluster analysis based on types of events attended, frequency of attendance, participation in arts activities, reasons for engaging and desire to do more – resulting in 6 clusters Group 2 Participate only Group 3 Neither attend nor participate divided into 3 subgroups based on age (16–29, 30–59, 60+) cluster analysis based on background and barriers towards attendance and participation – resulting in 4 clusters 11 The link to TGI explained An initial approximation is made of the three high-level groups according to participation and attendance. Group 2 (the participate only group) is simply subdivided on age criteria in TGI To achieve the best possible link, two algorithms will be created: one for each set of clusters under Group 1 and Group 3 This involves a discriminant analysis which identifies a) which TGI variables best classify TGI respondents into TP defined segments and b) a mathematical model to compute “best fit” for every TGI respondent. These algorithms are applied at a respondent level (i.e. line by line) with every respondent being allocated to the best fit segment. A customised TGI database is created with the segment flags (only, so no other variables) appended for ease of analysis. 12 Proposed approach: Stage 2 (cont.) Stage 2c: review algorithm with Arts Council and stakeholders statistical linking of Taking Part & TGI We would meet with Arts Council once the linking of the refreshed segmentation and TGI has been done, prior to profiling. This would be to review the success of the link and confirm that all parties are happy to proceed to the full profiling stage. At this point, a review of the Taking Part Year 4 questionnaire shows that all but 1 of the input variables for statistical linking used in the last project are available. Therefore the link will be statistically as viable as in the original project. 13 Proposed approach: Stage 3 Stage 3: in depth profiling of the segments using TGI, TGI Net and Taking Part variables We would profile all 13 segments – providing key insight on who, if any, have changed behaviour and why We will deliver: • a refreshed arts-based segmentation of the population of England – adults 16+ • profiles of each segment • recommendations as to: • how the Arts Council might communicate and engage with different audiences • potential communication routes to increase arts attendance and participation • the potential for the Arts Council to reach new audiences – with a greater focus on digital and social media than previously. 14 Proposed approach: Stage 3 Stage 3: in depth profiling of the segments using TGI, TGI Net and Taking Part variables We would repeat the analysis of the Taking Part variables to understand arts engagement, incorporating this into the profiling outputs. For example, this provided rich insight into drivers /barriers to participation and new attitudinal data on Taking Part Y4 will enhance this. Using TGI we would profile to give a rounded picture of each segment, and we will be looking at demographics, lifestyle and behaviours, leisure time use, attitudes and motivations, and media consumption. In addition we propose to also use an additional data source, TGI Net, providing greater understanding of digital behaviour to enrich media consumption profiling. As this would be a clear enhancement of the original profiles, we have provided an option to remove this component cognisant of budget constraints. 15 Project outputs • We will provide informative outputs to cover all 13 segments • This will include • Excel data tables – including TGI Net • Profile of all 13 segments in Powerpoint • This will cover for each segment: Activities and outings summary, Demographics, Key attitudes, Media summary, Art Behaviour summary, Internet behaviour summary, Charitable giving & volunteering summary,Drivers / Barriers • Will include an executive summary and outline of the research process • Recommendations Costs for deliverables assume use of the 2008 presentation template (using the Arts Council house style) and imagery, plus 2 new slides to incorporate findings from TGI Net. If any further change is required, costs would need to be amended accordingly. 16 Proposed approach: TGI data profiling specification by data source 330 attitude statements Leisure Cultural Eating out Outings Hobbies Holidays Socialising Word of Mouth Media TV Newspapers Magazines Radio Cinema Outdoor Promotions Internet Frequency Purchases (e.g. Tickets) Top line uses (e.g. Study, Social Networking) Visiting sites on (e.g. Listings) Specific brand sites (300+) Blogging Video content viewed (music/concert footage) Social Network services used Mobile access Internet attitudes (c40) TGI x TGI Net x x x x x x x x x x x x x x x x x x x x x x x x 17 Additional Project Output: Segments will be released on a new TGI dataset (GB TGI Q4 2010*) available in Choices *NB The segments will be appended to a single database (Q4) which excludes TGI Net as this is not part of ACE’s subscription 18 Golden Questions • The final stage of the process would be to provide Arts Council England with ‘golden questions’ – the key questions to give a reasonably accurate prediction of a respondent's segment for use on future research. • At this point we aren’t able to give parameters (e.g. this will result in 10 questions) but can break it down into its two constituent parts. • There is the initial discriminant analysis (or similar technique) which would be done on all 13 segments. • The final step would be to build a model for re-creation, and test its accuracy. 19 Project plan Timings are tight. Should more time be needed to take account of internal sign off procedures these timings will need to be reviewed. Action Action Owner Date Inception meeting ACE 28-Feb-11 Consultation meeting with stakeholders ACE 03-Mar-11 Segmentation Refresh TGI By 18th Mar Weekly status to ACE TGI 04-Mar-11 Profiling of segments TGI 08-Mar – 24 Mar-11 Weekly status to ACE TGI 11-Mar -11 Weekly status to ACE TGI 18-Mar-11 Interim debrief TGI 24-Mar-11 Weekly status to ACE TGI 25-Mar-11 Report to ACE TGI 29-Mar-11 Comments on report to TGI ACE 31-Mar-11 Golden questions – TGI w/c 28th March Data Source Taking Part Year 4, TGI Q4 2010, TGI Net Q4 2010 Taking Part Year 4, TGI Q4 2010, TGI Net Q4 2010 20 Project Team • • • Please note that some information has been withheld from this section. This information is being withheld under s40 (2) of the Freedom of Information Act 2000. This is because it is personal information relating to another individual and therefore constitutes personal data. We do not have consent to release this information and I do not believe there is a legitimate public interest in this information being disclosed. The Arts Council has a responsibility to hold personal data securely in line with the principles of the Data Protection Act 1998 and releasing this information would contravene the first principle of fair and lawful processing. 21 Project Team • • • Please note that some information has been withheld from this section. This information is being withheld under s40 (2) of the Freedom of Information Act 2000. This is because it is personal information relating to another individual and therefore constitutes personal data. We do not have consent to release this information and I do not believe there is a legitimate public interest in this information being disclosed. The Arts Council has a responsibility to hold personal data securely in line with the principles of the Data Protection Act 1998 and releasing this information would contravene the first principle of fair and lawful processing. 22 Quality Assurance • • • We comply fully with the MRS Code of Conduct as company members of MRS. As a WPP company we are compliant with Sarbanes Oxley, and ISO quality standards. Project managers quality check all analysis and outputs, and are regularly audited to maintain standards. All outputs are checked by project leads systematically to ensure compliance with MRS and ISO. Project Quality Checklist Job Name: Job Number: Author: Items marked (QP) are mandatory as part of quality procedure Action Evidence required Proposal (QP) Document & sent to client Timings (QP) Document & sent to client Cost sheet (QP) Document Commissionning form with T&C (QP) Document & client signoff Approval of targets created/recreated (QP) Document & client signoff (where applicable) Deliverables contain checking stamp (QP) Any client concerns are addressed immediately. Escalations should first be addressed to the project lead and to the Unit Director– Sandy Livingstone [email protected] OK None - visual check Dates + version # on deliverable name (QP) None - visual check Deliverable placed in a zipped archive (QP) Document Deliverables (QP) Document & Sent to client Project quality review (QP) Completed form 23 Costs The total cost for all three stages is £48,875 +VAT (£58,650). The costs are based on estimated timings consisting of – Director level – 7 days Associate Director level – 7.5 days Sub associate level – 19 days An option is available without inclusion of TGI Net data at a reduced cost of £42,875 +VAT (£51,450). Assumptions have been made and considered in the above costs for the following: The Arts council template and house style has not changed since the original segmentation project in 2008 and where appropriate the style and imagery used for profiling will remain the same. Any format/style change will incur an additional cost. The final written report will be in the form of PowerPoint presentations as indicated on Slide 16. All above costs are inclusive of all expenses that may be incurred during the project 24 Previous experience: Department of Health Client Issue: Client had a qualitative attitudinal segmentation with multiple groups based on health, but little knowledge of beyond this and basic demographics etc Solution: • • • GfK quantified the clients initial project using an ad-hoc study (5,000 sample) asking about attitudes and behaviour towards health related issues. TGI respondents were fused to the Health Foundations respondents, to allow profiling the segments on wider lifestyle attitudes and behaviour. To conduct an equivalent large scale survey on all these customers would have had far greater time and cost implications Profiles were presented at 10 workshops throughout the England, to embed throughout the organisation Impact: Greatly enhanced the behavioural segment profiles – bringing them to life and describing them in a language that is common to marketing, media and creative agencies alike. Allowed SHAs, PHOs and PCTs to make regional assessments of the national picture. 25 Previous experience: Financial Services Client Client Issue: Client had hypothesis of a brand targets for two new product launches. They wanted to verify and quantify these groups across its main European markets. Solution: • Rigorous statistical analysis using respondent level TGI data helped prove whether these targets really existed as a relatively homogenous group within the population. • Using factor analysis of the key attitudes we presented a number of solutions to try to draw out the discriminating statements. Following agreement on dimensions we ran two parallel cluster analyses to plot both the brand targets on TGI. • We were able objectively to prove what opportunities existed in each market they operated in and make this available in a currency such as TGI to allow sizing and profiling opportunities. Impact: There are specific targets in each market embedded in the client’s brand strategy. They are now able to market their products directly to banking partners and potential partners as well as plan and buy media spend directly aimed at each group. 26
© Copyright 2026 Paperzz