FAQs

Ethnicity is one of the variables which comes from the Census and is used in the build of Acorn. Ethnicity is a key variable for many of our public sector clients.

We have also built a version of Acorn which does not use any ethnicity from the Census as an input variable, and so can be used by insurance companies without the worry that they could be making decisions related to protected characteristics.

As both the Scottish and Irish Census’s were delayed due to Covid the data had not been released when we built Acorn. We did however have access to all the data contained in our postcode spine for these areas.

We therefore built Acorn using our postcode spine combined with the current Census for England & Wales and the previous Scottish & Northern Irish Census. We have then applied the segmentation in Scotland & Northern Ireland using the information we know using the up to date postcode spine data. This means we have a good view of the Scottish & Northern Irish landscape before the Census data arrives.

We have also done extensive testing and reviewed postcodes in Scotland and Northern Ireland to make sure these areas look as expected.

When the Scottish& Northern Irish Census data is released, we will reclassify the postcodes within Scotland & Northern Ireland using the new census information; the underlying structure of Categories, Groups and Types will not change.

The Census took place on 21^st March 2021 when England was in the middle of a lockdown. The main effect this has had on the Census data is around students living at home and people’s normal place of work.

To identify daytime population, we are planning to use current mobile app data which will show the movement of people by Acorn code, to understand the residential Acorn segment and where they are likely to be located during office hours, as we return to the new normal patterns of work post Covid. We will use the BRES survey data as control figures for this piece of work.

Understanding the worker Acorn profile of a catchment helps our clients to provide the right services in a particular location based to the daytime demographics of an area.

To identify areas where students live, we will use the Postal Address File (PAF) to help us identify student halls as well as using information from property portals to identify properties which have been advertised as accommodation for students.

Acorn has been completely rebuilt by using innate data sources to reflect the demographic, social and economic changes in the UK.

There are now over 20 sources of data in addition to the Census data, accounting for over 1 billon data points, which have been used to build Acorn. This means that Acorn can be updated easily, allowing for an infill of data where the Census had gaps due to being undertaken during the Covid-19 pandemic.

Acorn combines 2021 Census data with CACI’s recently developed Postcode Spine. From the Census, variables such as age, property type, tenure, family structure, education, occupation, ethnicity and persons per household were extracted. The Postcode Spine comes from a range of sources, including Open Data, Government Data, commercial data and our own proprietary datasets as part of the build process. Geographically detailed data about house sales, rental properties, planning applications, new build and student housing is used along with proprietary information about income and disposable income and AI techniques were utilised to determine built-up areas from satellite imagery.

These two sources in tandem provide an optimal view of the UK population in relation to the demographics and affluence of all UK postcodes.

Yes, Acorn is GDPR compliant and is available as a postcode directory built from datasets at postcode or larger area level, including the Census. No individual level data has been used as a direct input to build the segmentation, only data aggregated and combined with other datasets at postcode level or above, such as with the Land Registry data or CACI’s Paycheck postcode income model. Acorn is only classified as personal data under GDPR when it is appended to a piece of personal data such as a customer name and address.

There is a vast range of supporting collateral available to assist users with their interpretation of Acorn segments:

This includes:

• Pen portraits – a visual summary of each Acorn segment showing at a glance their key demographic and behavioural characteristics.
• Acorn videos – a video version of the pen portraits, hosted on YouTube
• Acorn Explore Dashboard – our PowerBI dashboard enabling users to drill into the detail of each Acorn segments as well as comparing characteristics between segments
• Acorn business use cases – a dashboard which demonstrates how Acorn and its supporting collateral can be used to discover where your key customers are located, what their key characteristics are as well as identifying the best Acorn types to target based on the features which are important to you.
• Acorn user guide with a detailed explanation of how to use Acorn, how it has been built and client use cases
• Acorn Knowledge Sheets – an Excel template enabling users to look up the characteristics and behaviours across over 500 variables for each of the Acorn segments
• A microsite which hosts all the above information as well as enabling users and prospects to look up a postcode and bring back the associated information for the Acorn segment associated with that postcode.
• Acorn Profiler – a tool which enables you to upload your consumer’s postcodes to create a profile of how they compare to the UK or other cohorts of customers eg your highest spending customers compared to your total customer base.

Household Acorn classifies each household in the country so is available at a more granular level than our traditional postcode Acorn product. It was designed to enhance the advantage of small area segmentation by recognising that not all households conform to the dominant or average lifestage or family structure in a street or postcode. It therefore places more emphasis on discriminating by lifestage so can, for example, identify empty nesters and families living side by side in a wealthy street.

There are a number of things which have been improved to make this version of Acorn our best yet. These include redeveloping all the code in Python and using the latest AI techniques to build the segmentation algorithm for Acorn that decides how to allocate the postcode data into segments.

We have also tested and licensed new data sources to improve how we identify areas of gentrification, as well as exploring ways to identify day time populations. We have developed a Postcode Spine of updateable sources of information as well as an AI solution for identifying built up areas from satellite imagery.

There are now over 20 sources of data accounting for over 1 billon data points which we use to build Acorn in addition to the Census data. This means we can update Acorn more easily as well as being able to infill data where the Census had gaps having been undertaken in a lockdown.

We will still call the three hierarchical levels Categories, Groups and Types. It is likely there will be different numbers of these at each level compared to the current version of Acorn, with more variety in urban areas to improve discrimination, particularly in city centres.

We have established an end-to-end testing process to ensure the build works smoothly and to speed up product development once the complete 2021 Census data is released by the ONS.

We are improving the Acorn data visualisation with a new microsite and dashboards to enable customers to drill into the detail of each Category, Group and Type.

In the 1950’s, the National Readership Society developed a way of classifying people based on the occupation of the head of the household. The system is known as the NRS social grade. It classifies people as being in Groups ABC1 (middle class) or C2DE (working class). It has been widely used by advertising and market research organisations over the years.

It differs from Acorn in that Acorn can segment the population into up to XX distinct behaviour Types enabling micro targeting. It can also identify where in the country the different segments of the population reside as well as understanding the key segments that make your consumer base. With vast amounts of supporting information about each of the Acorn segments available, marketers can really get under the skin of the motivations of each consumer group.

Acorn is refreshed annually using the latest version of the postal geography containing all new postcodes available from the Royal Mail at that time.

If postcodes have been created since March 2021 when the census was undertaken then we use our other, more recent and frequently updated datasets alongside the AI and regression models to inform how the postcode is allocated to the most appropriate Acorn segment. This is also how we will continue to keep Acorn updated for the next 10 years.

CACI have unparalleled expertise in the building of representative demographic segmentations. Acorn has a 40-year track record of providing powerful discrimination at a postcode level, and our large customer base using and gaining value from it is testament to that fact.

Demographic segmentation by its very nature can never get every postcode “correct” because there is no “actual” segmentation to compare it to. To that end we have gone through an extensive process of testing postcodes and area profiles, using real world data from clients and surveys to test the accuracy of the current Acorn as well as rerunning internal projects to ensure the classification remains intuitively discriminatory.

In addition to manual checking of postcodes, new visualisation techniques have enabled us to interactively examine the classification of postcodes, postcode sectors and local authorities on a map.

AI/ML techniques have been used throughout the Acorn build this year, from data processing all the way through to generation of new segments (Categories, Groups and Types).

In the data processing stage, more advance statistical techniques have been used to refine the driver variable values, as well as Deep Learning techniques to extract a measure of rurality from satellite imagery.

In the generation of new Acorn segments a number of leading-edge ML techniques have been employed both to improve the segmentation quality and improve the algorithm speed. For segmentation, a cutting-edge Deep Learning technique has been used to extract the best representation of the driver variable values, ready for segment creation. Thereafter, ML techniques have been used to automatically interpret the segments leading to a much faster turnaround in segment outputs from driver variable inputs.

Acorn is updated annually when the new postcode geography is built which adds in all new postcodes and flags deleted postcodes. All the information which goes into building Acorn is updated and a directory is built. This means that some postcodes will change their Acorn segment over time for instance as an area gentrifies and house prices start to rise.