Developing a ‘free’ customer classification system – The Alternative Approach!

Like many other Local Authorities, here in Hull we too have wrestled with the decision of whether to purchase expensive customer profiling software. The localism bill, creating a ‘bigger’ society and the need to save lots of money is certainly our immediate challenge, all to happen alongside looking after local people, meeting their needs and ensuring that our core services continue to support vulnerable people.

A deep understanding of citizens is paramount then. That was certainly the message we received at a recent customer insight conference held in London. But how?

Off the shelve products and profiling software are certainly tempting and seductive, with clever use of large national datasets and GIS mapping platform producing endless customer groupings mapped down to the lowest geographies. But after careful consideration it just wasn’t for us.

You see, Hull is unique and so are our citizens, whether it’s the Port influenced local economy (thanks to the Vikings), our ‘end of the road’ isolated location or maybe our strange love for Patties, national statistical models never seem to describe the city in the way we know it to be.

Therefore our journey of customer insight would need to be bespoke, built upon what we know already and most certainly be centred on ‘Hull’ data.

We set out on a journey to develop a customer insight data hub, collaborating mountains of data and threading them to a classification system built upon Hull’s own census data. Armed with a highly skilled team of analysts we began to re-model 45 census variables that include age, ethnicity, car ownership, housing tenure, health, employment and many more using ‘Cluster’ analysis to find natural groupings within the 250,000 census dataset. Not before long clear customer groupings were beginning to emerge as SPSS (statistical software package) did its job in differentiating people who owned their own home, lived in a terraced house, had children, worked and had 2 cars, from people who rented homes from the Council, single parents, lived in a flat or were high income earners.

10 groups were finalised, groups that at the highest level seemed to accurately describe the city into 3 hierarchies; owner occupiers, private renters and public sector renters. With the Census file for the city now ‘flagged’ with a customer group, this opened up a plethora of opportunities to learn more about these 10 groups. Mapped via ArcGIS software to census output areas, any dataset with a Hull postcode could now be added to the datafile.

Hull’s unique customer groups, mapped via ArcGIS to census output area

It took some time before the realms of possibilities kicked in. We started with matching 1 million records from our CRM system, linking every service request within the last 2 years to each customer group. We now knew for each unique customer group what they had contacted the council about (their service needs) and of course how often. We linked our property database to show where all the publicly owned assets reside in the city and mapped this to highlight where services were being over subscribed or totally under utilised. Health data was next, showing which groups were most likely to suffer from certain health problems, smoking, cancers, life expectancy. Crime data could show us which customer groups suffered which types of crime. Finally and maybe most importantly survey research data, all though smaller sample sizes showed how the different groups differ attitudinally, how they rated their neighbourhoods, local problems, satisfactions with services, and more complex issues such as their aspirations and multiple needs.

Detailed Description of one of the unique groups ‘A3’

All this based entirely for free, easily available data which most importantly is collected from Hull people, in Hull about Hull.

The work is rapidly taking off, as we present this data using inspiration from ‘information is beautiful’ which uses graphic design to present complex data and make it easy to understand. This is paramount given our audiences, whether they be community development workers, planners, senior managers or indeed members of the public, we must assume people don’t know how to interpret what we produce.

Analysis of ‘matched data’ to unique customer group A3

Initial presentations in Hull of our findings are being used to plan our long term capital and asset strategy, area and neighbourhood plans and projects, commercialise our leisure centres and theatres and most importantly provide the basis for key business intelligence to the Corporate Strategy Team, helping to underpin high level strategic decisions.

There is always another way as long as you know how!

If you would like any further info or want to see some of the work we have produced then please get in touch:

Andy Parkinson

Kingston Upon Hull City Council

e-mail: andrew.parkinson@hullcc.gov.uk

tel: 01482 613336