Connect with us

MARKETING

How bias in AI can damage marketing data and what you can do about it

Published

on

How bias in AI can damage marketing data and what you can do about it

Algorithms are at the heart of marketing and martech. They are used for data analysis, data collection, audience segmentation and much, much more. That’s because they are at the heart of the artificial intelligence which is built on them. Marketers rely on AI systems to provide neutral, reliable data. If it doesn’t, it can misdirect your marketing efforts..

We like to think of algorithms as sets of rules without bias or intent. In themselves, that’s exactly what they are. They don’t have opinions.. But those rules are built on the suppositions and values of their creator. That’s one way bias gets into AI. The other and perhaps more important way is through the data it is trained on. 

Dig deeper: Bard and ChatGPT will ultimately make the search experience better

For example, facial recognition systems are trained on sets of images of mostly lighter-skinned people. As a result they are notoriously bad at recognizing darker-skinned people. In one instance, 28 members of Congress, disproportionately people of color, were incorrectly matched with mugshot images. The failure of attempts to correct this has led some companies, most notably Microsoft, to stop selling these systems to police departments. 

ChatGPT, Google’s Bard and other AI-powered chatbots are autoregressive language models using deep learning to produce text. That learning is trained on a huge data set, possibly encompassing everything posted on the internet during a given time period — a data set riddled with error, disinformation and, of course, bias.

Only as good as the data it gets

“If you give it access to the internet it, inherently has whatever bias exists,” says Paul Roetzer, founder and CEO of The Marketing AI Institute. “It’s just a mirror on humanity in many ways.”

The builders of these systems are aware of this.

In [ChatGPT creator] OpenAI’s disclosures and disclaimers they say negative sentiment is more closely associated with African American female names than any other name set within there,” says Christopher Penn, co-founder and chief data scientist at TrustInsights.ai. “So if you have any kind of fully automated black box sentiment modeling and you’re judging people’s first names, if Letitia gets a lower score than Laura, you have a problem. You are reinforcing these biases.”

OpenAI’s best practices documents also says, “From hallucinating inaccurate information, to offensive outputs, to bias, and much more, language models may not be suitable for every use case without significant modifications.”

What’s a marketer to do?

Mitigating bias is essential for marketers who want to work with the best possible data. Eliminating it will forever be a moving target, a goal to pursue but not necessarily achieve. 

“What marketers and martech companies should be thinking is, ‘How do we apply this on the training data that goes in so that the model has fewer biases to start with that we have to mitigate later?’” says Christopher Penn. “Don’t put garbage in, you don’t have to filter garbage out.”

There are tools which can help you do this. Here are the five best known ones:

  • What-If from Google is an open source tool to help detect the existence of bias in a model by manipulating data points, generating plots and specifying criteria to test if changes impact the end result.
  • AI Fairness 360 from IBM is an open-source toolkit to detect and eliminate bias in machine learning models.
  • Fairlearn from Microsoft designed to help with navigating trade-offs between fairness and model performance.
  • Local Interpretable Model-Agnostic Explanations (LIME) created by researcher Marco Tulio Ribeiro lets users manipulate different components of a model to better understand and be able to point out the source of bias if one exists.
  • FairML from MIT’s Julius Adebayo is an end-to-end toolbox for auditing predictive models by quantifying the relative significance of the model’s inputs. 

“They are good when you know what you’re looking for,” says Penn. “They are less good when you’re not sure what’s in the box.”

Judging inputs is the easy part

For example, he says, with AI Fairness 360, you can give it a series of loan decisions and a list of protected classes — age, gender, race, etcetera. It can then identify any biases in the training data or in the model and sound an alarm when the model starts to drift in a direction that’s biased. 

“When you’re doing generation it’s a lot harder to do that, particularly if you’re doing copy or imagery,” Penn says. “The tools that exist right now are mainly meant for tabular rectangular data with clear outcomes that you’re trying to mitigate against.”

The systems that generate content, like ChatGPT and Bard, are incredibly computing-intensive. Adding additional safeguards against bias will have a significant impact on their performance. This adds to the already difficult task of building them, so don’t expect any resolution soon. 

Can’t afford to wait

Because of brand risk, marketers can’t afford to sit around and wait for the models to fix themselves. The mitigation they need to be doing for AI-generated content is constantly asking what could go wrong. The best people to be asking that are from the diversity, equity and inclusion efforts.

“Organizations give a lot of lip service to DEI initiatives,” says Penn, “but this is where DEI actually can shine. [Have the] diversity team … inspect the outputs of the models and say, ‘This is not OK or this is OK.’ And then have that be built into processes, like DEI has given this its stamp of approval.”

How companies define and mitigate against bias in all these systems will be significant markers of its culture.

“Each organization is going to have to develop their own principles about how they develop and use this technology,” says Paul Roetzer. “And I don’t know how else it’s solved other than at that subjective level of ‘this is what we deem bias to be and we will, or will not, use tools that allow this to happen.”


Get MarTech! Daily. Free. In your inbox.


Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address

MARKETING

Take back your ROI by owning your data

Published

on

Treasure Data 800x450

Treasure Data 800x450

Other brands can copy your style, tone and strategy — but they can’t copy your data.

Your data is your competitive advantage in an environment where enterprises are working to grab market share by designing can’t-miss, always-on customer experiences. Your marketing tech stack enables those experiences. 

Join ActionIQ and Snowplow to learn the value of composing your stack – decoupling the data collection and activation layers to drive more intelligent targeting.

Register and attend “Maximizing Marketing ROI With a Composable Stack: Separating Reality from Fallacy,” presented by Snowplow and ActionIQ.


Click here to view more MarTech webinars.


About the author

Cynthia RamsaranCynthia Ramsaran

Cynthia Ramsaran is director of custom content at Third Door Media, publishers of Search Engine Land and MarTech. A multi-channel storyteller with over two decades of editorial/content marketing experience, Cynthia’s expertise spans the marketing, technology, finance, manufacturing and gaming industries. She was a writer/producer for CNBC.com and produced thought leadership for KPMG. Cynthia hails from Queens, NY and earned her Bachelor’s and MBA from St. John’s University.

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading

MARKETING

Revolutionizing Auto Retail: The Game-Changing Partnership Between Amazon and Hyundai

Published

on

Revolutionizing Auto Retail: The Game-Changing Partnership Between Amazon and Hyundai

Revolutionizing Auto Retail The Game Changing Partnership Between Amazon and Hyundai

In a groundbreaking alliance, Amazon and Hyundai have joined forces to reshape the automotive landscape, promising a revolutionary shift in how we buy, drive, and experience cars.

Imagine browsing for your dream car on Amazon, with the option to seamlessly purchase, pick up, or have it delivered—all within the familiar confines of the world’s largest online marketplace. Buckle up as we explore the potential impact of this monumental partnership and the transformation it heralds for the future of auto retail.

Driving Change Through Amazon’s Auto Revolution

Consider “Josh”, a tech-savvy professional with an affinity for efficiency. Faced with the tedious process of purchasing a new car, he stumbled upon Amazon’s automotive section. Intrigued by the prospect of a one-stop shopping experience, Josh decided to explore the Amazon-Hyundai collaboration.

The result?

A hassle-free online car purchase, personalized to his preferences, and delivered to his doorstep. Josh’s story is just a glimpse into the real-world impact of this game-changing partnership.

Bridging the Gap Between Convenience and Complexity

Traditional car buying is often marred by complexities, from navigating dealership lots to negotiating prices. The disconnect between the convenience consumers seek and the cumbersome process they endure has long been a pain point in the automotive industry. The need for a streamlined, customer-centric solution has never been more pressing.

1701235578 44 Revolutionizing Auto Retail The Game Changing Partnership Between Amazon and Hyundai1701235578 44 Revolutionizing Auto Retail The Game Changing Partnership Between Amazon and Hyundai

Ecommerce Partnership Reshaping Auto Retail Dynamics

Enter Amazon and Hyundai’s new strategic partnership coming in 2024—an innovative solution poised to redefine the car-buying experience. The trio of key developments—Amazon becoming a virtual showroom, Hyundai embracing AWS for a digital makeover, and the integration of Alexa into next-gen vehicles—addresses the pain points with a holistic approach.

In 2024, auto dealers for the first time will be able to sell vehicles in Amazon’s U.S. store, and Hyundai will be the first brand available for customers to purchase.

Amazon and Hyundai launch a broad, strategic partnership—including vehicle sales on Amazon.com in 2024 – Amazon Staff

This collaboration promises not just a transaction but a transformation in the way customers interact with, purchase, and engage with their vehicles.

Pedal to the Metal

Seamless Online Purchase:

  • Complete the entire transaction within the trusted Amazon platform.
  • Utilize familiar payment and financing options.
  • Opt for convenient pick-up or doorstep delivery.
Ecommerce CertificationEcommerce Certification

Become A Certified E-Commerce Marketing Master

The Industry’s Most Comprehensive E-Commerce Marketing Certification For The Modern Marketer. Turn Products Into Profit, Browsers Into Buyers, & Past Purchasers Into Life-Long Customers

Click here

Hyundai’s Cloud-First Transformation:

  • Experience a data-driven organization powered by AWS.
  • Benefit from enhanced production optimization, cost reduction, and improved security.

Alexa Integration in Next-Gen Vehicles:

  • Enjoy a hands-free, voice-controlled experience in Hyundai vehicles.
  • Access music, podcasts, reminders, and smart home controls effortlessly.
  • Stay connected with up-to-date traffic and weather information.

Driving into the Future

The Amazon-Hyundai collaboration is not just a partnership; it’s a revolution in motion. As we witness the fusion of e-commerce giant Amazon with automotive prowess of Hyundai, the potential impact on customer behavior is staggering.

The age-old challenges of car buying are met with a forward-thinking, customer-centric solution, paving the way for a new era in auto retail. From the comfort of your home to the driver’s seat, this partnership is set to redefine every step of the journey, promising a future where buying a car is as easy as ordering a package online.

Embrace the change, and witness the evolution of auto retail unfold before your eyes.


Revolutionizing Auto Retail The Game Changing Partnership Between Amazon and Hyundai

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading

MARKETING

How to Schedule Ad Customizers for Google RSAs [2024]

Published

on

How to Schedule Ad Customizers for Google RSAs [2024]

It’s no wonder that responsive search ads have steadily grown in popularity in recent years. Through Google’s machine learning capabilities, RSAs provide a powerful way to automate the testing of multiple headlines and descriptions to ensure a closer match to user intent. The benefits are clear: RSAs mean broader reach, better engagement, and improved performance metrics.

However, all these benefits come at a significant (but reasonable) cost – they can be extremely difficult to manage, especially when it comes to updating ad copy to promote limited time offers.

I know this firsthand – I work with several ecommerce clients with promotions that constantly change. Not too long ago, I found myself going through the consistently tedious process of updating a client’s RSA headlines and copy. As I was making the changes, I thought to myself: “There must be a better way to update this ad copy. I shouldn’t have to use find and replace so many times while pausing and enabling my ad campaigns.”

After expressing this to my colleague, Jordan Stambaugh, the two of us agreed there must be a better way. But we’d have to make it happen. A few weeks later, we put that idea into action and created a more efficient process for updating RSA ad copy on a scheduled basis. If you want to try this process for yourself, just keep reading.

Responsive Search Ad Customizers 101: Basic Options & Execution

Before diving into the process of scheduling automatic updates for your RSA customizers, it’s essential to understand some key Responsive Search Ad fundamentals.

First, you can customize three main options within RSAs: the Attribute Name, the Data Type, and the Account Value. Each of these plays a vital role in personalizing your ads:

  • Attribute Name: This is essentially the identifier for the customizer. It is how you’ll reference the specific piece of information you’re customizing within the ad. For instance, if you’re running a promotion, you might name an attribute “Promotion.”
  • Data Type: This indicates the kind of data the attribute represents and it determines how the information can be formatted and used within the ad. Common data types include Text (for plain, non-numeric text), Percent (to represent percentage discounts), Price (to denote monetary values), and Number (for any numerical value).
  • Account Value: This is the default value for the attribute that you set at the account level. It acts as a fallback if more specific values aren’t provided at the campaign or ad group level.

For example, if you wanted to promote a 10% off discount using RSAs, you’d use the “Discount” attribute, a data type of “Percent,” and an account value of “10% off.” Then, when someone is searching for products, Google would test automatically inserting a copy regarding a 10% off promotion into your ad.

Once you’ve set up the right customization options, you can start to format your RSAs with customizers.

Here’s how:

  • Start by typing in {
  • Click on Ad Customizer then select your attribute
  • Google will populate your attributes that are already uploaded
  • For a simple offer, use the “Default text” attribute as a catch-all. This will ensure your ads run smoothly if Google can’t pull the right messaging from your RSA feed

 

 

How to Schedule Your Ad Customizers with a Feed

Now that we’ve covered the basics, let’s cover how to schedule your ad customizers.

Just follow this three step process:

1. Create the feed

Start by creating two sheets: The Parent sheet, and the Child sheet. The “Parent” sheet will act as the primary data source, while the child sheet will pull data from the parent sheet.

We’ll start by building the parent sheet. After opening the sheet, start by renaming the active tab to “Promotions.” Don’t skip this step, it’s crucial for referencing this range in formulas later on.

In your “Promotions” tab, head to the top row and label columns A, B, and C with the headers of your ad customizer attributes. For example, you might have “BrandSaleHeadline” as your attribute in column A, “text” as the Data Type in column B, and “Shop the Collection” as the Account Value in column C.

Once your headers are in place, move to cell C2. Here, you’ll input the expression =lookup(today(),F:G,E:E). This formula will play a key role in dynamically updating your RSA customizer based on the current date.

Next, go to columns E, F, and G, which will be used to manage your scheduling. In these columns, you’ll list out the different values your chosen attribute might take, alongside their corresponding start and end dates. For example, under the “BrandSaleHeadline” attribute, you might schedule various promotional headlines to appear during different sale periods throughout the year.

Here’s how your sheet might look:

Now look back at the first 3 columns on your sheet. They should look like this:

Now create a second sheet. We’ll call this sheet the Child sheet. It’s going to automatically pull in data from the parent sheet you just created, and will be the one you link to Google Ads later on.

Columns A, B and C will be almost identical to the child sheet, but we will be using a special formula later so we can automatically populate this. So, start by labeling Row 1 Column A “Attribute,” then the next column as “Data type,” then column C as “Account value.” 

Then go to C2 and use this expression to populate the right account value from the parent document: =importrange(“[PARENT DOCUMENT URL HERE]”,”Promotions!C2″)

Your sheet should now look like this:

We recommend adding a date range with default text for any days you’re  not running a promotion. In the example above, we have “Shop Our Collection” appearing as default text.

2. Input attributes

Once you have your feed created, the next step involves inputting your attributes into the Google Ads platform. This can be done either manually or through a bulk upload.

For the manual approach, navigate to “Tools & Settings” in your Google Ads interface, then go to ‘Setup’ followed by “Business Data.” Here, you’ll find an option for “Ad Customizer Attributes.” Click the plus sign to add your attributes. It’s crucial to use the same attribute names that you’ve established in your Parent Google Sheet template to ensure consistency and proper data synchronization.

 

 

Alternatively, if you prefer the bulk upload method, again head to “Tools & Settings.” This time, select “Bulk Actions” and then “Uploads.” For this process, you only need to upload columns A to C from your template. 

Be aware that it might take some time for your uploaded attributes to be reflected in the business data section of Google Ads.

3. Set up an automatic schedule

At this point, you’ve almost finished scheduling your ad customizers. Navigate to Tools & Settings, then Bulk Actions, then Uploads, then click the Schedules tab at the top. Select your Child Google Sheet as the data source, and share your Google Sheet with the appropriate email.

 

 

And there you have it – Google will automatically pull in the data you populated in the sheets into your RSAs.

Common Challenges When Scheduling RSA Ad Customizers

When we test these sheets with our clients in the wild, we’ve uncovered five common challenges. Be on the lookout for these issues – solving them before they happen can save you a lot of trouble down the line.

Not scheduling your upload when the site changes 

The first and most significant hurdle is the mismatch between the scheduled data upload and website content updates. For instance, if the Google Sheet is set to upload at 11 am, but the website changes occur at 3 pm, there’s going to be a discrepancy where the wrong message could be displayed for several hours, or new messaging could appear prematurely. Conversely, if the website updates happen before the scheduled sheet upload, outdated promotions might linger until the new data is imported. Synchronizing these schedules is crucial; it’s best to align them so updates occur simultaneously.

Skipping QA during a message change

Another pitfall is neglecting quality assurance (QA) during message updates. It’s vital to regularly check the business data section to verify that the correct values are in place post-update.

Issues with the IMPORTRANGE function

Then there’s the technical aspect of setting up the IMPORTRANGE function correctly in the Google Sheets template. The ‘child’ template must reliably pull data from the ‘parent’ sheet. If this function isn’t configured correctly, data won’t be imported as needed.

Not sharing access of the Google template for automatic uploads

Pay attention to your access permissions for the Google Sheets template. Google will prompt you with the email address that needs permission to access the ‘child’ sheet for automatic uploads. Overlooking the sharing of your sheet with this address will prevent the system from working.

Having date range gaps in your parent sheet

Lastly, a common oversight is leaving date range gaps in the ‘parent’ sheet. Every single date must be accounted for without overlaps. A practical tip is to have an ‘evergreen’ backup message ready, scheduled to run continuously, ideally through the end of the year, to cover any potential gaps.

Conclusion

Leveraging Google Sheets in conjunction with Google Ads to schedule RSA ad customizers is a game-changer for managing dynamic promotional content. This process not only streamlines your workflows but also ensures that your ads remain relevant and up-to-date, reflecting current promotions without the need for constant manual intervention. 

By adopting this method, you’ll save significant time and effort, allowing you to focus more on strategy and less on the minutiae of ad copy updates. Give it a try and experience a more efficient way to manage your RSAs, keeping your campaigns fresh and engaging with minimal hassle.

Source link

Keep an eye on what we are doing
Be the first to get latest updates and exclusive content straight to your email inbox.
We promise not to spam you. You can unsubscribe at any time.
Invalid email address
Continue Reading

Trending