Before any analysis was
carried out, we cleaned the database to maximise
the accuracy of subsequent profiling. This involved
identifying duplicates, goneaways and deceased
records, updating records matched to the NCOA
file and enhancing addresses to Postal Address
File standard.
We then began by examining the current makeup
of the database.
Each customer was assigned a Mosaic Code, and
reports were produced analysing the overall distribution
of Mosaic Groups and Types within the database,
as well as penetration of each Group and Type
as compared to UK Base figures. Click
here for a Mosaic Sample Report.
We took a flat file of customer records and normalised
it in a way that enabled us carry out recency,
frequency and monetary analysis. We also created
hierarchies to enable segmentation on attributes
specific to the client’s data, such as destination
country, holiday type and resort.
Additionally, we used a GIS mapping tool to
locate each existing customer at street level
on a UK map. Click here for a
map of customers.
During a half day consultancy, we were able to
carry out further bespoke analysis on the trends
that were important to the client.
Mosaic Groups in which the company was strongly
represented (had high penetration) were singled
out for further analysis. Using the mapping tool
we highlighted exactly where these customers live,
and were able to drill down to identify additional
characteristics, for example growth or reduction
in spend, recency and frequency of transactions,
as well as trends specific to the company. Targeted Mosaic Type Map.
As a result of the consultancy, a selection was
made of Mosaic groups which were both well represented
on the existing database and had promising trends
in transaction history. Using the mapping tool,
counts for each of the Mosaic Groups were produced
broken down by Postcode District, Area and Sector.
Counts and prices were then obtained for the purchase
of additional Mosaic Data in these areas.
End result:
A highly focused selection was made of Mosaic
Groups in a handful of postcode areas. Data was
purchased for each selected Group in each selected
area, at a greatly reduced cost to the client
when compared to purchasing blanket, unprofiled
data.
