Martin CallinghamVisiting ProfessorBirkbeck College, University of London

Market Research Society

People and Geography

The use and further development of OACThe Society for Chemical Industry 27th November 2006

Contents

Naming the sub-groups

Mapping properties of OAC

Geographical properties of OAC

Customer profiling, market modelling, targeting people

The discrimination of OAC

Extending OAC to 10,400 levels

Greater discrimination

Bespoke systems

Access to the data

Super Groups 7 Groups 21 Sub-groups 521: Blue Collar Communities 1a: Terraced Blue Collar 3

1b: Younger Blue Collar 21c: Older Blue Collar 3

2: City Living 2a: Transient Communities 22b: Settled in the City 2

3: Countryside 3a: Village Life 23b: Agricultural 23c: Accessible Countryside 2

4: Prospering Suburbs 4a: Prospering Younger Families 24b: Prospering Older Families 44c: Prospering Semis 34d: Thriving Suburbs 2

5: Constrained by Circumstances 5a: Senior Communities 25b: Older Workers 4

5c: Public Housing 36: Typical Traits 6a: Settled Households 2

6b: Least Divergent 36c: Young Families in Terraced Homes 2

6d: Aspiring Households 27: Multicultural 7a: Asian Communities 3

7b: Afro-Caribbean Communities 2

Naming the sub-groups

Naming is difficult – virtually impossible

Use the census variables that were used to make OAC

Use geographical properties of the OAC sub-groups

Use survey data where available

Use common sense and ground truth

BUT the names describe the areas not everyone in an area is the same

Representation of activities and OAC groups

Multi-dimensional scaling on lifestyle activity indices

Grand Children

DIY

Sewing

Knitting

Fishing

Crosswords

Wildlife

Golf

Hiking

Garden

Mail order Wines

National Trust

Charity

Current Affairs

Gourmet meals

Theatre

Foreign Travel

Art

Skiing

Active sport

Records

Pub

Religion

Self improvement

Fashion

Grand Children

DIY

Sewing

Knitting

Fishing

Crosswords

Wildlife

Golf

Hiking

Garden

Mail order Wines

National Trust

Charity

Current Affairs

Gourmet meals

Theatre

Foreign Travel

Art

Skiing

Active sport

Records

Pub

Religion

Self improvement

Fashion

Student BedsitsYoung Cosmopolitan ProfessionalsGenteel Flatland Professional Town FamiliesWorking villagesPoorer CountrysideNot tied to the LandIn Contact with NatureWell-off Rural ManufacturingPleasant Rural Retirement

Hard Working but DullBetter-off Hard Working but Dull

Fading Middle Class

Bed-sit LandDeprived Bed-sit LandWell-off StartersRedeveloped established area

Asian GhettosImpoverished Melting PotSuccessful Second GenerationGentrificationDeprived Tower Blocks

Thoughts of some sub-group names

Mean Household Income

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

OAC 52

Inco

me

Student BedsitsYoung Cosmopolitan ProfessionalsGenteel Flatland Professional Town Families

Working villagesPoorer CountrysideNot tied to the LandIn Contact with NatureWell-off Rural ManufacturingPleasant Rural Retirement

Hard Working but DullBetter-off Hard Working but Dull

Fading Middle Class

Bed-sit LandDeprived Bed-sit LandWell-off StartersRedeveloped established area

Asian GhettosImpoverished Melting PotSuccessful Second GenerationGentrificationDeprived Tower Blocks

Maps from OAC

Map makes very good maps

Top level is 7 – and 7 colours can be perceptually resolved

The urban groups are distinct – so maps look sensible

AND geodemographics spatially autocorrelate

OAC Super Groups for southern Britain

The OAC types have some interesting geographical properties

They vary by:

•Population density

•Worker to resident ratio

•Settlement size

•Relationship to the centre of the settlement

•Coastal index

•Elevation

Geographical properties of OAC

Population density

0

20

40

60

80

100

Den

sity

Worker to resident ratio

0

0.5

1

1.5

2

2.5

Rat

io

Mean population of locations containing the sub-group

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

Loca

tion

popu

latio

n

Distance index indexed on the mean for the relevant population size

0.55

0.65

0.75

0.85

0.95

1.05

1.15

1.25

Inde

xed

Dis

tanc

e In

dex

Coastal Index

0.00

1.00

2.00

3.00

4.00

5.00

6.00

Coa

stal

Inde

x

Elevation

30

40

50

60

70

80

90

100

110

Elev

atio

n

Centre of gravity of each OAC 52 sub-group

52 Deprived Tower Blocks

48 Asian Ghettos

50 Successful Second Generation

51 Gentrification

49 Impoverished Melting-pot

Measuring propensities

People can be OAC coded from their postcode

This enables:

People to be profiled

Markets to be modelled

Groups to be targeted.

For example, people who:

Read the Telegraph or the Sport

Go Skiing or play Bingo or Knit

Or have treatment for migraine

Propensity profiles by OAC

Daily Telegraph

0.00

0.50

1.00

1.50

2.00

2.50

3.00R

elat

ive

Inde

x

Daily Sport

0.00

0.50

1.00

1.50

2.00

2.50

Rel

ativ

e In

dex

Skiing

0.00

0.50

1.00

1.50

2.00

2.50

3.00

Rel

ativ

e In

dex

Rel

ativ

e In

dex

Bingo

0.00

0.50

1.00

1.50

2.00

2.50

Knitting

0.2

0.4

0.6

0.8

1

1.2

1.4

Rel

ativ

e In

dex

Take treatment for Migraine

0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

1.80

I calculated the gain for 14 eating out markets for two levels of propriety systems A and M – 56 calculations

Discrimination of OAC

50

Area % for eating out markets with Mosaic 61

French

Indonesian/Thai

Spanish

Greek/Turkish

American Style Restaurants

Italian

Pizza Restaurants

Hamburger Bars

Steak Houses

Other Type

Indian

English

Pub Restaurants

Chinese

0 5 10 15 20Area %

25 30 35 40 45

Adj Area %

Hamburger 23.07Steak Houses 23.18Pub Res 22.17American Style 32.90Pizza Res 29.38Chinese 16.44English 20.16French 49.98Greek/Turkish 38.57Indian 25.57Indonesian/Thai 44.50Italian 32.06Spanish 41.58Other Type 22.81

Mean 30.17

Gain achieved in different markets by M 61

Area % by market of A 17 against M 11

y = 0.8768x + 0.4633

R squared2 = 0.96

0.00

5

10

15.

20

25

30

35

40

45

0.00 5 10 15 20 25 30 35 40 45 50

M 1

1

A17

The systems discriminate in a similar way across markets

35

A 6+

M 11+

A 17+

M 61+

Area % for A and M systems across 14 eating out markets

0 5 10 15 20 25 30

Mean Area % across 14 markets

System Area %A 6+ 17.61M 11+ 21.15A 17+ 23.59M 61+ 30.17

Average gain by each system

Relationship between the levels in the system and mean Area %

R Squared =0.99

10

15

20

25

30

35

0 10 20 30 40 50 60 70

Number of Levels

Are

a %

Effectiveness of a system depends only upon the number of levels

OAC Of Mosaic

OAC 52 0.98OAC 21 0.82OAC 7 0.62

Geodem Weighteddeviation

Mosaic 61 40.7OAC 52 39.9Bespoke 36.6OAC 21 35.7OAC 7 29.7Mosaic 11 28.2Bespoke 27.1IMD 24.0Social Class 19.0

GISRUK poster 2006

Geodem Weighted Relative otdeviation Mosaic 61

Mosaic 61 40.7 1.00OAC 52 39.9 0.98Bespoke 36.6 0.90OAC 21 35.7 0.88OAC 7 29.7 0.73Mosaic 11 28.2 0.69Bespoke 27.1 0.67IMD 24.0 0.59Social Class 19.0 0.47

Geodem Weighted Relative ot Predicteddeviation Mosaic 61

Mosaic 61 40.7 1.00 1.00OAC 52 39.9 0.98 0.98Bespoke 36.6 0.90OAC 21 35.7 0.88 0.82OAC 7 29.7 0.73 0.62Mosaic 11 28.2 0.69Bespoke 27.1 0.67IMD 24.0 0.59Social Class 19.0 0.47

OAC is as discriminatory as other geodemographic systems

I calculated the gain for 14 eating out markets for two levels of A and M – 56 calculations

This showed, in these markets, discrimination of the two systems differed only by the level of clustering. As M had the higher number, it discriminated the best.

Comparison of the OAC to Mosaic at UCL with adjustment for cluster numbers showed that OAC performed comparably.

Summary of discrimination

OAC was further sub-divided into 260, 1300 and 10,400 clusters

At 10,400 there are on average only 20 OAs per cluster.

Extending OAC – increasing its discrimination

Propensity (index) by OAC 52

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

OAC 52

Inde

x

Cumulative index versus cumulative population

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

4.00

0 10,000,000 20,000,000 30,000,000 40,000,000 50,000,000 60,000,000

Cumulative population

Cum

ulat

ive

inde

x

Analysis of Gain OAC 52

0.00

20.00

40.00

60.00

80.00

100.00

120.00

20.00 40.00 60.00 80.00 100.00 120.00

Cumulative population %

Cum

ulat

ive

Cus

tom

er %

Gain is 47.77%

3.4

Index OAC 260

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

Level Gain 2 Million index

52 48% 3.38260 53% 4.26

Index for OAC 1300

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

OAC 1300

Inde

x

Level Gain 2 Million index

52 48% 3.38

260 53% 4.26

1,300 56% 4.85

OAC 10400 index

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

OAC 10400

Inde

x

Level Gain 2 Million index

52 48% 3.38

260 53% 4.26

1,300 56% 4.8510,400 62% 5.57

OAC levelsCode included

%1 100

52 100260 98

1300 8110400 44

Gain % 2 million Advantage Cum over OAC 52Index

0.00 0.0047.7 3.38 152.6 4.26 1.2356.31 4.85 1.4061.61 5.57 1.61

Gain versus OAC level

0

10

20

30

40

50

60

70

0 2000 4000 6000 8000 10000

OAC level

Gai

n %

Making a bespoke classification

Because geodemographics types concentrate, it means that in any area there is likely to be only a sub-set of types that apply

This can occur in a Local Authority

For example Newham only has 6 types out of OAC 52

This means that the use of current systems of geodemographics isseverely limited in such areas.

The same is true for niche markets.

Percentage spread nationally (pop)

0.000.501.001.502.002.503.003.504.00

1 5 9 13 17 21 25 29 33 37 41 45 49

OAC 52

%

Percentage spread in Newham

0

5

10

15

20

25

30

35

OAC 52

% o

f pop

ulat

ion

Newham OAC 1300

0

5

10

15

20

25

OAC 260

% of p

opulation

Newham OAC 1300

0.00 5.00 10.00 15.00 20.00 25.00

Deprived Tower Blocks

Deprived Tower Blocks

Deprived Tower Blocks

Deprived Tower Blocks

Deprived Tower Blocks

Gentrification

Gentrification

Gentrification

Gentrification

Gentrification

Successful Second Generation

Successful Second Generation

Successful Second Generation

Successful Second Generation

Impoverished Melting Pot

Impoverished Melting Pot

Impoverished Melting Pot

Impoverished Melting Pot

Impoverished Melting Pot

Asian Ghettos

Asian Ghettos

Asian Ghettos

Young Cosmopolitan Professional

Young Cosmopolitan Professional

Young Cosmopolitan Professional

% o

f pop

ulat

ion

OAC 1300

Bespoke geodemographics

It is possible to make a bespoke geodemographic with sensible sizes and number of groups:

Combine the the smaller groups within the hierarchy

Split the larger groups by going up the hierarchy

BUT the bespoke system would still fall within OAC

Summary

OAC:Maps well

Has interesting geographical properties

Can be used to customer profile from a list of postcodes

Can be used to attribute market/behavioural values to an area from surveys

Can be used to target areas for resource allocation and marketing

Can be extended to give enhanced discrimination

Can be extended to give a bespoke system within the OAC framework

Is a publicly available and free government system

AccessOACA description of OAC is given at:

http://www.statistics.gov.uk/census2001/cn_139.asp

The classification can be downloaded free as regional files or obtained on disc (2001 Census Area Classification of Output Areas) from their customer services people (01329 813800) or by e mailing info@statistics.gov.uk).

It is better to get the disk, as it has the click use license and lots of additional information including files of the variables used in OACs making and other supporting information. The files comprise an output area - to – areal classification look-up table at the GOR and (as a txt file – Oaclass.txt) national level.

ONS also provide a ‘postcode to output area look-up table’ (also available free for ONS customer services on a separate disc).

A more up-to-date file post coded file can be obtained for modest cost from their customer services as an extract from the National Statistics Postcode Directory (NSPD) – although it has a rather strange (and irritating) postcode format. This is how you can get the most up to date version of OAC.

A combination of the two files therefore enables customer or respondent files to be given the appropriate classification of the output area in which they live. This opens up the opportunity to profile customers on this system and to model market propensity at thecensus output area level.

Detailed spider diagrams are published to explain the meaning of each of the classifications as a series of PDF files at:

http://www.statistics.gov.uk/about/methodology_by_theme/area_classification/oa/cluster_summaries.asp

Convenient names have been generated for the classifications OAC 7 and OAC 21 only at:

http://www.geog.leeds.ac.uk/people/d.vickers/OAclassinfo.html

This web site contains many other things of interest

Martin Callingham is contactable at martincallingham @ yahoo.com