PersonaLive comes with extensive metadata to learn about the segments and make them actionable. Customers have access to our metadata website which provides a online interface to learn about each segment. We often refer to the website that contains metadata for all of our segments as "The Taxonomy".

For our entire metadata on each segment, including thousands of brands, interests, influencers, and more, customers can use our segment and attribute sorters which can be found in your cloud account. You can also learn more about the extended metadata below.

Metadata is broken down into the following sections:

• Overview
• Live Data
• Geography
• In-Store Visitation
• Brand Influence
• Channels
• Social Topics
• Influencers
• Web Visitation
• Demographics

Before we visit these sections, it's important that we explain a very common mechanic that is common in our metadata, the index score.

### Index Scoring¶

In our taxonomy, you will regularly see scores integer scores next to various brands, influencers, attributes, and more. For example:

These score are indexed to the national average, where the national average is represented by a score of 100. So, in the above example, this segment (#RoaringRetirees) is 1.54x as likely be interested in Golf Digest. The general formula is:

$Index \, Score = \frac{Segment \, Percentage}{National \, Percentage} * 100$

Let's say that, on average, 1% of Americans engage with Golf Digest on social media. In that case, a segment that has 1.54% of it's members interested in Golf Digest would have an index score of 154.

$\frac{1.54 \%}{1.0\%} * 100 = 154$

Other than index scores, most other numbers in the metadata should be pretty intuitive. Let's dive into the various sections of the taxonomy.

## Overview¶

This is your summary page for any given segment, providing top level insights from each section of the metadata. All of the data that you see on this page can be further explored in each section of the taxonomy, so we won't spend too much time describing the specific data this section.

What is unique to this page is that it provides each segment's name, image, code, typical household age and income, and percentage of the population.

## Live Data¶

One of the most exciting parts of the PersonaLive metadata is our live data. You can see what each segment is talking about on social media, how in-store visitation is changing, and how their web visitation is shifting. Let's check it out.

### Live Social Topics¶

Live social topics reflect the hashtags being used by this segment right now. These are hashtags that this segment is overindexing on relative to other segments as well as relative to time (this hashtag is being used more now than in the past).

Live topics are calculated using Twitter and Instagram data. Use this section to find opportunities for engaging with customers and more.

### Live In-Store Visitation¶

In this section, we highlight the the brands and businesses which have seen increased visitation in the most recent month relative to the prior year. Brands are filtered to those which this segment over-indexes on (in general), and those that have had the highest increase in visitation in recent months are shown.

This charts calculation has some nuance. Indexes are based on percentage (or "share") of visitation. In the screenshot, you can see that Rita's Italian Ice has had an increase of 55.7%. This indicates that share of visitation has increased 55.7%. In a hypothetical, Rita's visits made up 1.0% of #RoaringRetirees visitation in the last year, then the last month visitation to Rita's was 1.557%, a 55.7% increase. We used 1% to keep it simple, but that's more in the Walmart territory of visitation share.

### Live Web Visitation¶

View changes in the web visitation habits of this segment. Similar to live in-store visitation, this section compares the most recent month to the longer time period and highlights the web visitation behaviors that have increased the most.

## Geography¶

The geography section provides the ability to look at this segment on an interactive map. It highlights the census block groups which are dominantly this segment and provides a list of the markets which have the highest concentration of this particular segment. In the screenshot, we've zoomed in on Hartford (CT), which is the market with the 3rd highest concentration of #RoaringRetirees of all markets in the USA.

## In-Store Visitation¶

This section shows the brands, businesses, and points of interest which this segment over-indexes on visiting relative to the national average. These indexes are calculated using visitation percentage (or "share"). If 2% of #RoaringRetirees' visits go to Stein Mart but nationally only 1% of visits go to Stein Mart, then this segment has an index of 200 for Stein Mart. We do take into account whether a Stein Mart is available when calculating these values.

We've broken down the businesses into several different categories (such as Retail-Apparel, Retail-Home Improvement, Restaurant-Fast Casual, etc.) for enhanced usability and understandability.

## Brand Influence¶

Use this section to explore the brands which this segment over-indexes on. You can see that we have results at an overall level but you can also view these results for specific categories, such as Fashion, Home, Beauty, etc. A score of 200 would indicate that #RoaringRetirees engages with a given brand at twice the national rate.

## Channels¶

The Channels section uses the same approach as the Brand Influence section (index of social media engagement compared to the national average), but focuses on "marketing channels", or the places where you might be able to reach a customer. As you can see in the screenshot, that includes websites, news sources, TV Channels, TV shows, podcasts, and more.

By using this tab, you can better understand exactly who these people are (and know the exact right spot to place your ad). For this example, we decided to switch it up and show you segment A06, #TechTitans. There are 2 sections on this page that are calculated slightly differently, so we're going to go into brief details on those.

### Out of Home Exposure¶

Understanding how individuals can best be reached via out of home advertising is not necessarily answered well by using social media (as the other things on the channels page are). So, we used a combination of datasets to generate values for the above chart, while still using the same indexing approach.

Value Source
Airports, Malls, Stadiums Calculated using same approach as the brand visitation indexes, using mobile phone location data.
Billboards Calculated using Census data, specifically mode of transport to work. Calculated as percentage of people who commute via car divided by the national average.
Transit Calculated using Census data, specifically mode of transport to work. Calculated as percentage of people who commute via public transit divided by the national average.

### Social Media Platform Usage¶

Ironically, though we collect social media data, it's not necessarily straightforward to use this data for estimating platform usage. This section was estimated using data from a 2019 Pew Research study.

## Social Topics¶

Welcome to one of the most unique parts of the metadata. This tab has 3 main sections.

#### Social Media Frequency¶

This section tells you how many times per week users in this segment tend to post. Note that this is a simple average for the social media users in a given segment.

#### Social Media Topics¶

The topics section provides a list of hashtags that this segment uses at a higher rate in general than the average social media user. We also provide a companion panel which shows recent tweets using the listed hashtags. For example, the top hashtag for #TechTitans is #Cloud. The screenshot shows tweets that contain this hashtag.

#### Proximity (Geosocial) Segment¶

This section shows the top Proximity segments from Spatial.ai's Proximity dataset in areas where the given PersonaLive segment is present. We understand this may be confusing... our Proximity Dataset is a separate dataset from PersonaLive that focuses on geolocated social media alone. In short, the Proximity dataset scores areas based on the text of geolocated social media being posted in the area.

For example, #TechTitans's top Proximity segment is LGBTQ Culture. This means that areas in which #TechTitans live tend to score highly for LGBTQ Culture. This makes sense-- Tech Titans are heavily located in San Francisco and other places known for flourishing LGBTQ culture.

## Influencers¶

Who are the celebrities and influencers that a given segment follows on social media? Find out in the Influencer section. These values are calculated using the same index approach as is used in the Channels and Brand Influence Sections.

You can switch between categories of influencers and even see a live twitter feed of the posts being made by these influencers. As you can see, #TechTitans tend to follow tech leaders and VCs, such as Sundar Pichai and Ben Horowitz.

## Web Visitation¶

In the web visitation section, you will find information on the types of websites which this segment over-indexes on visiting. Each value is presented as an index.

## Demographics¶

The demographics section shows basic demographic variables for each segment. These demographics describe the makeup of areas where this segment is dominant. This data comes straight from the census and is shown in both percentage (in the screenshot on the right) and index (on the left) form. We've included variables related to:

• Household Composition
• Ethnicity
• Income
• Education
• Occupation
• Population Density
• Home Ownership
• And more...

Demographics will always be important. PersonaLive innovates by not only using demographics but also the many other datasets we've already gone over in this metadata section.