My List ♥ ()

AI / LLM Content Analysis Fields

Source Type: AI / LLM

LLMs (Large Language Models, like ChatGPT) can be used in a variety of ways in content analysis, but one way they can be useful is by categorizing or summarizing content. Using LLMs at scale can be expensive, so the sophistication, cost, and energy efficiency of the model should be right-sized.

In Content Chimera

Content Chimera deeply embeds LLMs, from semantic search to chat to adding fields to the analysis. Chimera supports LLM Fields by:

Defining sets of fields, so that you can far more efficiently get multiple field values from one overall prompt.
Allowing users to define sets of potential field values, and ensuring that the values from the prompts are within that set. Chimera also has a retry mechanism in place if the first LLM response was outside that set of acceptable values.
Storing the values in a consistent manner for charting, rules, and other analysis.
Supporting either single-level or multi-level taxonomies
Leveraging existing Chimera site cache for faster execution
Ability to select from a wide range of potential models from Anthropic (Claude), Mistral, and OpenAI (ChatGPT) in order to match cost, speed, and energy efficiency to the needs

See AI / LLM fields below. Or show fields for all field types.

Answer some questions to help select fields

If we take the traditional view of a content inventory or audit, we have rows representing each page (so each row has a unique URL) and then we have columns for things like the meta description or crawl depth. These columns are the different fields we have available to us in our content analysis.

①. Define what you are trying to accomplish.

Your content analysis needs to be grounded on your analyze goal.

Examples: Plan Digital Transformation, Test Content Hypothesis, Provide Better Bid.

②. Define your analysis approach.

Size and complexity of your digital presence

Size and complexity of your digital presence should drive your content analysis approach.

My digital presence is:

Use the calculator

Your approach

Content analysis does not necessarily mean opening up a spreadsheet. Before diving in, you should define your basic approach to the analysis.

Examples: Brute Force; Sample, Rules, Repeat; Quick Take.

③. Select fields toward your goal, grounded in your prioritized list of questions you want answered.

Although you can use this database however you like, in general we recommend that you build up a list of fields that will be useful for your analysis. To do so, just click on the heart next to any field name. After you have hearted some fields, you can see an analysis of your list at My List ♥ (at which point you can move to ④. Start iterating on your analysis, starting with the basics).

Audience ♡

Audience the content is actually written to target (whether or not it is supposed to).

General Usefulness:

Ease of Automation:

Compare with other User fields.

Author ♡

The person(s) who wrote the content. This may be different than who published or crafted the page.

General Usefulness:

Ease of Automation:

Compare with other Org fields.

Content Type ♡

Content Type (semantic type of content, such as Product Page or Event) is usually an extremely effective way to group and look for patterns across a digital presence.

General Usefulness:

Ease of Automation:

Compare with other Category fields.

Tone ♡

How the content is communicated with language. Tone may reasonably vary across a digital presence.

General Usefulness:

Ease of Automation:

Compare with other Brand fields.

Topic ♡

The topic/subject of the content.

General Usefulness:

Ease of Automation:

Compare with other Category fields.

Voice ♡

Brand voice

General Usefulness:

Ease of Automation:

Compare with other Brand fields.

Legend