Overview

The flood of artificial intelligence (AI) has had profound effects on multiple sectors of business. AI in the class action settlement and notice plan space specifically has, and will continue to have, powerful impacts. If used for nefarious intent or improperly used, it can have significant consequences—from fraudulent traffic and claims, to notice plans with inflated impression totals failing to reach the true target class. However, as with most tools, when implemented and managed properly, it can drive efficiency in cost and results. When utilizing or combating the ever-changing AI technologies across the digital marketing space, Élan ensures the proper steps are taken so that ads serve correctly and comply with applicable ad verification guidelines. These measures support the success of the notice plan’s reach, brand protection, budget, and key performance indicator (KPI) goals. Tools across all digital platforms are implemented, monitored, utilized, and optimized to fit the application of advertising in the Class Action Legal space. 

Benefits Across Digital Platforms 

Ad Platform Integrated AI

Élan’s experience with AI has been extremely positive as platforms across the digital landscape have deployed AI tools. Campaigns supplementing our manual optimizations with those made by AI are seeing increasingly better and more efficient results. Platforms such as Meta, Google Ads, Bing, Reddit, Google Analytics, TikTok, etc., have implemented several products, and this list is ever-growing.

Drawbacks and How to Combat Them

Bots & Fraud: Digital Advertising Front-End Protection Efforts

While it is difficult to prevent all bot and fraudulent claims submissions, steps can be taken from the front lines of digital marketing to identify, reduce, and monitor an influx of bot and AI traffic. This includes tags, pixels, urchin tracking modules (UTMs), and site filtering. These tools ensure ads are directed to reputable locations and the subsequent traffic to the website can be tracked to the origin of the session and claims data. 

Given the multi-media, multi-platform nature of most notice campaigns, traffic to the website is tracked using UTMs or other tracking methods to determine the source and origin of both session and claims data. Élan uses Google Analytics to capture this information and relies on it as a vital, data-driven tool to support decision-making for all media optimizations. We also utilize site filtering and specific targeting to direct ads to reputable and premium inventory. At the same time, we implement blacklists to avoid publishers and inappropriate content categories. These measures support the success of the notice plan’s reach, brand protection, and budget/KPI goals. These tools also assist our team in identifying fraudulent and AI bot traffic and claim submissions through oddities or anomalies amongst the inbound website traffic. 

Landing Page: Anti-Spam Tools & Strategies

The measures above can provide the best chance of your advertisements serving to the intended audience, but what about when your website draws eyes beyond your ad spend? The website can still be at risk of spam and bots, which are becoming ever more intelligent through the growth of AI.

Below are a couple of the many strategies that can be utilized to combat this: 

reCAPTCHA Placement

reCAPTCHA can be placed alongside your claim form to support traffic verification based on the version selected. 

• reCAPTCHA v3
  • Provides a score for each use based on the user’s behavior, helping the team identify spam traffic.
• reCAPTCHA v2 (“I’m not a robot” Checkbox)
  • Requires the user to check a box to proceed. If Google deems the online behavior to be human, the user proceeds without delay. Alternatively, if the review is failed, a user will be challenged with an additional verification task.
• reCAPTCHA v2 (Invisible reCAPTCHA Badge)
  • This version is executed when the applicable button is selected, running a verification without interaction from a user unless the verification fails.  

      Source: reCAPTCHA Guide

Honeypot

A honeypot is a hidden form field built into the claims form intending to have bot traffic fill it in while humans will not. This prevention tactic can be developed in many different ways through creative coding. 

Analytics: Monitoring Traffic & Goals

Continuous monitoring, analysis of web traffic, and notice plan goals are crucial to ensuring bad actors and spam traffic are identified. Luckily, AI-powered products are growing within analytics platforms as well. Google Analytics has developed Analytics Intelligence, which is a set of features powered by machine learning that can be used to detect anomalies, trends, and actionable data insights:

• Analytics Insights
• Anomaly Detection
• Trend Change Detection
• Contribution Analysis
• Recommendations  

These tools allow Élan analysts to effectively identify, verify, and optimize the appropriate pacing of each campaign KPI across the notice plan.

Need Help Navigating AI in Your Next Class Action Notice Plan?

Élan has the media and analytics professionals to ensure your plan can utilize all of the benefits of artificial intelligence while mitigating the pitfalls. Contact us today for a free class action notice plan proposal. 

Overview

R is a powerful programming language that is widely used for statistical computing and data analysis. As a marketing analyst, you can use R to perform tasks such as data manipulation, data visualization, text mining and more. Its vast library of packages allows for effective handling of large datasets and elaborate analyses. 

Data Manipulation 

Data manipulation packages are vital when it comes to organizing your data so that you can perform your analysis and gain valuable insights. This package is one I load every time I create a new R script because I know chances are I’ll probably be using it for whatever task I am trying to accomplish.

dplyr

dplyr is my most used package and what I predominantly use for my data manipulation, as its capabilities allow me to perform all the tasks I need. The functions I use the most within the package are:

• summarize(): This function calculates summary statistics for groups of data based on the chosen grouping variables. It’s used to calculate totals and averages, but can also calculate any relevant KPI. When summarized, the grouped-by columns and the column created will only be shown.
• mutate(): Very similar to summarize(), mutate() creates new columns or updates existing ones. The only difference is that mutate() will add this column to the data frame, letting you continue to use it.
• filter(): Is self explanatory, as it filters out specific rows which you do not want to include. Being able to isolate certain audiences or events is very useful when trying to identify best performing groups (like ad groups or campaigns).
• group_by(): group_by() is a must use when creating a new column with mutate() because it allows you to group data into specific categories.
• arrange(): Once you have transformed your data to how you want it, you can use arrange() to sort the data to gain insights. A common use for it could be to sort by date, or some value like engagements. 

Data Visualization 

Data visualizations are extremely important in the analytics field, as they are the best way to see patterns and simplify complex information. Using these visualizations, analysts can present data effectively to others who may not have as much context as they do.

ggplot2

The good thing about creating visuals in R is that the ggplot2 package has almost every visual needed, and is what analysts use 90% of the time. What is also nice is that ggplot2 uses a consistent setup across different types of graphs, so switching visuals is very simple. ggplot2 also allows for a lot of customization in terms of fonts, coloring, positioning of labels, etc., so you can customize your visualization to your liking.

• geom_bar(): This creates a bar chart, which is typically used for comparing totals from multiple groups.
• geom_tile(): Not as common, but this creates a heatmap which can be a useful visual.

Example of using geom_tile() to create a heatmap with Google Ads Search data.

treemap

As with most processes in R, there are many ways to do the same thing. ggplot2 can be used to create a tree map, but I prefer using the treemap package because it is what I learned first.

• treemap(): Treemaps are a very good alternative to pie charts. They are best used when comparing values relative to the total of those values.

Example of using treemap() to make a treemap with the same Google Ads Search data.

Text Mining

There are a lot of scenarios where you may want to analyze text data, whether that be customer reviews or survey responses. In these scenarios, having a tool that helps you extract patterns and key phrases will allow you to track trends and discover actionable insights.

tidytext

One package I like is tidytext, which allows you to split large text strings into words or phrases. From there, you can work with the functions from the dplyr package to gain insights, like the most used phrases or words.

• unnest_tokens(): This is a function that allows you to extract single words or grouped words from text strings. You can choose how many words you want to be grouped together to find common phrases.

Bonus

If you or your company uses Google Analytics 4 (GA4), then this package can be very useful to you, as it allows you to pull data directly from GA4.

GoogleAnalyticsR

GoogleAnalyticsR lets you interact with the Google Analytics APIs. This can be beneficial for many reasons and allows you to do your own analysis. GA4 has a lot of very good metrics that it creates for you, but looking at this data in R can allow you to make more customized metrics and run different analyses.

Conclusion

Data manipulation, visualization, and text mining are crucial components for any marketing analyst using R. These capabilities help organize data, reveal patterns, and extract key insights from textual data. By integrating R into our analytical processes at Élan, we have been able to enhance our data analysis capabilities. R’s statistical functions and data visualization tools have improved our ability to make data driven insights from complex datasets, and streamline these processes to better our efficiency.