Hello again mCommunity,

I am excited to add to your regularly scheduled programing, a 3-part post for using Meltwater | Media Relations and Monitor to identify and follow people of interest - those journalists or other media contacts writing about topics germane to you and your organization.

If you missed the first post in this series, you can view it here!

Discovering and following relevant journalists and other media contacts is easy with Meltwater, but can be done in a variety of ways. Our three areas of focus this week are the following:

1. Using Media Relations Search to perform keyword and filter searches on contacts ‘writing about’ our topic of interest
2. Using PR Assistant’s AI-powered ‘Suggested journalists’ feature as a complimentary search method
3. Powering Monitor with Media Lists to keep abreast of people of interest

Creating a media list for authors of interest who are covering a specific beat or topic within that beat is easy. Following them to stay up to date on their activity is even easier. Let’s get started!

Using PR Assistant’s AI-powered ‘Suggested journalists’ feature as a complimentary search method

We just discussed how to use a combination of filters and keywords in Media Relations Search to find authors. However, we can use PR Assistant to do this in a way that is technically different from Media Relations Search. Indeed, while Media Relations Search is inherently using keywords and filters on contact attributes, PR Assistant works by applying a machine learning approach to data association called vector embeddings. 

Vector embeddings are numerical representations of words or entities in a multi-dimensional space, enabling language models to understand and manipulate language based on contextual relationships between these embeddings. Long story short, if you’re used to a xy-plane (2D) and plotting an (x, y) coordinate pair - imagine an even higher-dimensional mathematical space with hundreds of thousands or millions of variables! That is what we’re working with (just less ‘discrete’ for you math nerds).

By training our models on the millions of articles we ingest by hundreds of thousands of journalists, we’re able to take a larger collection of information, like your draft pitches and press releases, and determine which articles and authors are closely related to your content’s unique digital fingerprint.

How do we use it?

Simply navigate to Media Relations > PR Assistant, throw in your key messages or content draft (up to 4,000 characters), and click ‘generate.’ Once the content appears, click on the Suggested journalists tab to see those authors who are writing content contextually similar to your provided text.

Here’s an example using the following key messages:

After clicking ‘generate’ and clicking on the Suggested journalists tab:

And when reviewing the ‘Relevant articles:’

So, after reviewing the relevant articles for the suggested journalists, you may choose to add their profiles to media lists for performing highly-targeted outreach.

This approach is a great way to compliment keyword and filter searches, because of the technical differences in returning the authors. This simply does not compare to a simple keyword or filter search!

So our recommendation would be to combine these approaches to flesh out your media lists with well-researched prospects.

Tomorrow, we will continue our discussion by detailing how Monitor enables you to follow people of interest, drop in where you left off, and in the future - follow sources of interest too  that’s an allusion to a future feature release, indeed!] 

.

.

Until then, stay smart and sharp

Blake Corman

Product Team

Have problems, ideas or feedback? Schedule time on my calendar:

www.calendly.com/blake-corman