Guide | An introduction to indexes
This toolkit explains how indexes are useful in allowing us to compare data against a baseline average...
An index figure is commonly included in Area Profile Reports (APRs) and in geo-demographic profiles. Indexes are a useful way for us to compare a specific group of people (often bookers) to a larger set of the population, the people from a local or regional area for example.
This article explains how indexes have been used to compare the relationship between the same subsets or segments, (in this case Audience Spectrum segments) of two different populations, (in this case a cluster of arts organisation’s bookers and their catchment population) relative to the overall relationship between the two different populations. The indexes help us see the different dynamics between each of the different segments compared with the overall relationship between bookers and catchment population.
The table shows a fictional Audience Spectrum profile based on an Audience Finder geographic cluster. The table compares bookers in the cluster with the population of the wider catchment area.
An index value of 100 indicates that a result exactly matches the baseline average. The baseline average is the overall relationship between the bookers and catchment area population. The index shows how that relationship differs, compared to the overall base, for people in each different Audience Spectrum group. An index of 200 shows that the result is twice the average, and an index of 50 that it is half the average. Broadly speaking, an index of less than 90 or more than 110 would be considered different enough from the average to take notice of.
|Audience Spectrum segment (adults 15+)||Bookers||Catchment Area||Index|
|Trips & Treats||13,472||25||216,072||26||95|
|Home & Heritage||5,621||10||72,897||9||117|
|Up Our Street||3,625||7||92,205||11||60|
In the table, 27% of bookers in the cluster belong to the Audience Spectrum group Dormitory Dependables. However, only 19% of all the people who live in the catchment area belong to this segment. This result illustrates that the Dormitory Dependables group are over-represented in the cluster’s bookers compared to the catchment area population. The index figure allows us to understand this in more detail. So, an index of 144 means that people from Dormitory Dependables are 44% more likely to engage than the population of the catchment area as a whole.
25% of bookers belong to the Audience Spectrum group Trips and Treats, but 26% of people living in the catchment area belong to the Trips and Treats group. This tells us that the proportion of the Trips and Treats segment in the cluster and the proportion of Trips and Treats living in the catchment area is broadly the same. An index of 95 means that the cluster booker’s group is just 5% less likely to contain Trips and Treats people than a random sample of the catchment area population.
10% of bookers belong to the Audience Spectrum group Facebook Families. However, 18% of all the people who live in the catchment area belong to this group. Facebook Families are under-represented in the cluster’s audience compared to the catchment area population. An index of 55 means that the cluster group is 45% less likely to contain Facebook Families people than a random sample of the catchment area population.
Extremely high indexes
The highest index in this example is for Metroculturals at 4,805, however 1% of the bookers are Metroculturals, and 0% of the local population are Metroculturals. This index does indicate correctly that Metroculturals are over-represented in the booker data set when compared to the base population but, as the percentage of people from this group is so small in both the catchment area and the booker data sets, the index becomes less useful. Here, although the percentage for both groups appears as zero, when looking at the numbers in both data sets about one in several 100 bookers are Metroculturals compared to one in several thousand of the base population. This is why the index is so high, but this only occurs when the percentages are less than one.
A note on small sample sizes
Similarly, as indexes are calculated to compare the percentages of each group in the booker and population set, the smaller the data the larger the differences in indexes as each booker counts for a greater % of the overall number. So in a very small data set, the difference of 4 or 5 people in one group compared to another may drastically affect the index. This is useful to keep in mind if you have a small amount of data.
Indexes allow you to benchmark yourself and establish where your data set sits in the context of other artform attenders within your organisation or regionally or against the population as a whole. This way you can work out whether your identified segments are over or under-represented compared to each other or to benchmarks and therefore whether they are more or less likely to be a prospect. Where you are over-represented it would indicate that you are already successful in attracting that group and you would need to gauge the scope for further penetration, if you are under-represented it would suggest an area of potential.
By placing data in context, you can feel more confident that you have identified a workable segment. This context can also come from previous or other types of information you have collected yourself – effectively your own internal benchmarking; regional or national benchmarks, existing omnibus surveys, general research undertaken by other organisations or agencies and as a final check your own knowledge and experience of the market place.