Connect with us


Build Your Own SEO AnswerBox With GPT-3 Codex & Streamlit



Build Your Own SEO AnswerBox With GPT-3 Codex & Streamlit

Google has integrated a new method of querying data in GA4 whereby you simply type a question or a phrase, and it creates the dashboard.

Imagine if you could do the same with your own SEO data – what would that do for your productivity?

In this article, you’ll learn how to configure your own dashboards using phrases and questions, creating an AnswerBox of your own with GPT-3 Codex and Streamlit.

Google’s Answer Box

In Google Analytics v4, you’ve probably noticed a clever search bar that allows you to get:

  • Instant answers.
  • Reports.
  • Answers on configuration.
  • Help.

Instant answers are particularly useful. By asking questions about your data, you get answers and – most importantly – ready-to-use reports.

There’s nothing magical about it. This technology relies on natural language processing (NLP), so you have to be precise about metrics, dimensions, and timing when asking for an answer.

For example, you can search for [conversions last week from the United States] and see the results in the search panel that appears on the right.

Screenshot from Google Analytics v4, February 2022

This new way of using a data visualization tool is incredibly powerful and will surely be integrated into all solutions of this type.

The time savings for the user are impressive, as users don’t have to search through all the views of the tool and no longer have to configure the view.

Everything is done automatically based on the instructions provided in the search box.

Can we easily do the same thing? And which data should we use?

Smart SEO Dashboard

Before requesting a report, you need to think about the important data to be taken into account.

I suggest you look into the concept of the Smart SEO Dashboard.

  • The first requirement is to keep the graphs simple and specific. Less is always more.
  • Next, the abscissa or ordinates must refer to measurable data. Otherwise, it is impossible to see the evolution.
  • In addition, graphs must focus on meaningful parameters. It is useless to monitor parameters that will have no influence on your activity. Weather is an excellent example: it plays a crucial role on some sites and none on others.
  • Dashboards should always include relevant summaries in order to be quickly read and understood. Generally speaking, if it takes more than three seconds to understand a dashboard, it can definitely be improved.
  • Finally, the most important data is time. It is imperative to track time data by comparing each day, month, year, etc.

Now, you need to identify the best technology to generate this type of dashboard.

GPT-3 Codex

GPT-3 Codex is a computer code generator that was created in August 2021.

Access to GPT-3 Codex was given much faster than access to GPT-3.

Not surprisingly, GPT-3 Codex has been fed millions of quality source codes available on GitHub – that is, more than 54 million GitHub repositories.

Like GPT-3, it is a sophisticated neural network that is capable of self-learning.

GPT-3 Codex does not only work in Python. You can also generate code in Go, Javascript, Perl, and PHP.

On the other hand, GPT-3 Codex has only 12 billion parameters, unlike its big brother GPT-3 Da Vinci which has 175 billion.

Let’s take a closer look at this size-versus-cost ratio.

OpenAI’s experiments show that the size-versus-performance ratio of Codex follows a logarithmic scale.

This means that the performance gains gradually decrease as the size of the model increases.

Therefore, the additional costs of collecting data, training, and running a larger model are not at all worth the slight increase in performance.

All of these reasons explain why the model has only 12 billion parameters for its first version.

We’ve found an AI to generate the code.

Now let’s look for the best framework available at the moment to execute it all in a friendly interface with clicks and drag-and-drop.


Streamlit is an open source technology that allows you to quickly build very advanced user interfaces.

Streamlit also includes many very useful components to have even more interactions like:

  • Session management.
  • Password management.
  • User management.

The community is particularly active and shares many handy custom modules for SEO.

To start, we’ll use GPT-3 Codex to generate graphs with Streamlit, and then attempt to produce a Streamlit app that generates the code and runs it automatically.

Two Excellent Examples For Querying The Data

First, we need to generate an app for Streamlit and run it.

1. With OpenAI (Semi-Automatic)

The first thing to generate is a Streamlit web app that retrieves all the logs from the month of May 1995 from NASA and displays the number of URLs crawled per day.

First of all, we need to retrieve the CSV file by specifying the name of the columns and the format if necessary.

For our example, it is important that the date is in UTC format.

Then you can ask OpenAI to display the graph of your choice, once it has understood your data.

how to configure using OpenAIScreenshot from OpenAI, February 2022

From these instructions, you will have a working code.

Remember that we don’t want to copy and paste code, but drive everything through English instructions with a no-code approach.

2. With Streamlit (Full Automatic)

Here is an open source example based on one of Streamlit’s applications.

It is an app directly connected to GPT-3 Codex that generates computer code and allows you to execute it.

With Charly Wargnier, we did the same thing but for SEO use cases in an app called “Codex for SEO”.

In one click, you can import your data.

Then, you can describe the content of the imported file: What are the columns? What are the data types?

data analysis script in codex for seoScreenshot from Codex For SEO, February 2022

Then you specify your instructions.

In our example, we ask it to group the queries together and to sum the clicks and impressions.

We’ll tell it to keep only the columns we’re interested in (the Queries and Clicks columns), and then click the Execute button.

No line of code is needed to get the results, and everything is generated by the OpenAI Codex and executed by Streamlit.

Our proof of concept is therefore validated with many different use cases.

Moreover, if you need assistance, everything you need for this is accessible via a training program with 150 minutes of video.

results from codex for seo scriptScreenshot from Codex For SEO, February 2022

For educational and transparency reasons, we have provided the generated code as well as the results.

And with that, the SEO AnswerBox is now available for everyone to create!

More resources:

Featured Image: NicoElNino/Shutterstock

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


Google’s AI Overviews Shake Up Ecommerce Search Visibility




Google's AI Overviews Shake Up Ecommerce Search Visibility

An analysis of 25,000 ecommerce queries by Bartosz Góralewicz, founder of Onely, reveals the impact of Google’s AI overviews on search visibility for online retailers.

The study found that 16% of eCommerce queries now return an AI overview in search results, accounting for 13% of total search volume in this sector.

Notably, 80% of the sources listed in these AI overviews do not rank organically for the original query.

“Ranking #1-3 gives you only an 8% chance of being a source in AI overviews,” Góralewicz stated.

Shift Toward “Accelerated” Product Experiences

International SEO consultant Aleyda Solis analyzed the disconnect between traditional organic ranking and inclusion in AI overviews.

According to Solis, for product-related queries, Google is prioritizing an “accelerated” approach over summarizing currently ranking pages.

She commented Góralewicz’ findings, stating:

“… rather than providing high level summaries of what’s already ranked organically below, what Google does with e-commerce is “accelerate” the experience by already showcasing what the user would get next.”

Solis explains that for queries where Google previously ranked category pages, reviews, and buying guides, it’s now bypassing this level of results with AI overviews.

Assessing AI Overview Traffic Impact

To help retailers evaluate their exposure, Solis has shared a spreadsheet that analyzes the potential traffic impact of AI overviews.

As Góralewicz notes, this could be an initial rollout, speculating that “Google will expand AI overviews for high-cost queries when enabling ads” based on data showing they are currently excluded for high cost-per-click keywords.

An in-depth report across ecommerce and publishing is expected soon from Góralewicz and Onely, with additional insights into this search trend.

Why SEJ Cares

AI overviews represent a shift in how search visibility is achieved for ecommerce websites.

With most overviews currently pulling product data from non-ranking sources, the traditional connection between organic rankings and search traffic is being disrupted.

Retailers may need to adapt their SEO strategies for this new search environment.

How This Can Benefit You

While unsettling for established brands, AI overviews create new opportunities for retailers to gain visibility without competing for the most commercially valuable keywords.

Ecommerce sites can potentially circumvent traditional ranking barriers by optimizing product data and detail pages for Google’s “accelerated” product displays.

The detailed assessment framework provided by Solis enables merchants to audit their exposure and prioritize optimization needs accordingly.


What are the key findings from the analysis of AI overviews & ecommerce queries?

Góralewicz’s analysis of 25,000 ecommerce queries found:

  • 16% of ecommerce queries now return an AI overview in the search results.
  • 80% of the sources listed in these AI overviews do not rank organically for the original query.
  • Ranking positions #1-3 only provides an 8% chance of being a source in AI overviews.

These insights reveal significant shifts in how ecommerce sites need to approach search visibility.

Why are AI overviews pulling product data from non-ranking sources, and what does this mean for retailers?

Google’s AI overviews prioritize “accelerated” experiences over summarizing currently ranked pages for product-related queries.

This shift focuses on showcasing directly what users seek instead of traditional organic results.

For retailers, this means:

  • A need to optimize product pages beyond traditional SEO practices, catering to the data requirements of AI overviews.
  • Opportunities to gain visibility without necessarily holding top organic rankings.
  • Potential to bypass traditional ranking barriers by focusing on enhanced product data integration.

Retailers must adapt quickly to remain competitive in this evolving search environment.

What practical steps can retailers take to evaluate and improve their search visibility in light of AI overview disruptions?

Retailers can take several practical steps to evaluate and improve their search visibility:

  • Utilize the spreadsheet provided by Aleyda Solis to assess the potential traffic impact of AI overviews.
  • Optimize product and detail pages to align with the data and presentation style preferred by AI overviews.
  • Continuously monitor changes and updates to AI overviews, adapting strategies based on new data and trends.

These steps can help retailers navigate the impact of AI overviews and maintain or improve their search visibility.

Featured Image: Marco Lazzarini/Shutterstock

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


Google’s AI Overviews Go Viral, Draw Mainstream Media Scrutiny




Google's AI Overviews Go Viral, Draw Mainstream Media Scrutiny

Google’s rollout of AI-generated overviews in US search results is taking a disastrous turn, with mainstream media outlets like The New York Times, BBC, and CNBC reporting on numerous inaccuracies and bizarre responses.

On social media, users are sharing endless examples of the feature’s nonsensical and sometimes dangerous output.

From recommending non-toxic glue on pizza to suggesting that eating rocks provides nutritional benefits, the blunders would be amusing if they weren’t so alarming.

Mainstream Media Coverage

As reported by The New York Times, Google’s AI overviews struggle with basic facts, claiming that Barack Obama was the first Muslim president of the United States and stating that Andrew Jackson graduated from college in 2005.

These errors undermine trust in Google’s search engine, which more than two billion people rely on for authoritative information worldwide.

Manual Removal & System Refinements

As reported by The Verge, Google is now scrambling to remove the bizarre AI-generated responses and improve its systems manually.

A Google spokesperson confirmed that the company is taking “swift action” to remove problematic responses and using the examples to refine its AI overview feature.

Google’s Rush To AI Integration

The flawed rollout of AI overviews isn’t an isolated incident for Google.

As CNBC notes in its report, Google made several missteps in a rush to integrate AI into its products.

In February, Google was forced to pause its Gemini chatbot after it generated inaccurate images of historical figures and refused to depict white people in most instances.

Before that, the company’s Bard chatbot faced ridicule for sharing incorrect information about outer space, leading to a $100 billion drop in Google’s market value.

Despite these setbacks, industry experts cited by The New York Times suggest that Google has little choice but to continue advancing AI integration to remain competitive.

However, the challenges of taming large language models, which ingest false information and satirical posts, are now more apparent.

The Debate Over AI In Search

The controversy surrounding AI overviews adds fuel to the debate over the risks and limitations of AI.

While the technology holds potential, these missteps remind everyone that more testing is needed before unleashing it on the public.

The BBC notes that Google’s rivals face similar backlash over their attempts to cram more AI tools into their consumer-facing products.

The UK’s data watchdog is investigating Microsoft after it announced a feature that would take continuous screenshots of users’ online activity.

At the same time, actress Scarlett Johansson criticized OpenAI for using a voice likened to her own without permission.

What This Means For Websites & SEO Professionals

Mainstream media coverage of Google’s erroneous AI overviews brings the issue of declining search quality to public attention.

As the company works to address inaccuracies, the incident serves as a cautionary tale for the entire industry.

Important takeaway: Prioritize responsible use of AI technology to ensure the benefits outweigh its risks.

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


New Google Search Ads Resemble AI Assistant App




New Google Search Ads Resemble AI Assistant App

A keynote at Google’s Marketing Live event showed a new AI-powered visual search results that feature advertisements that engage users within the context of an AI-Assisted search, blurring the line between AI-generated search results and advertisements.

Google Lens is a truly helpful app but it becomes unconventional where it blurs the line between an assistant helping users and being led to a shopping cart. This new way of engaging potential customers with AI is so far out there that the presenter doesn’t even call it advertising, he doesn’t even use the word.

Visual Search Traffic Opportunity?

Google’s Group Product Manager Sylvanus Bent, begins the presentation with an overview of the next version of Google Lens visual search that will be useful for surfacing information and for help finding where to buy them.

Sylvanus explained how it will be an opportunity for websites to receive traffic from this new way to search.

“…whether you’re snapping a photo with lens or circling to search something on your social feed, visual search unlocks new ways to explore whatever catches your eye, and we recently announced a newly redesigned results page for Visual search.

Soon, instead of just visual matches, you’ll see a wide range of results, from images to video, web links, and facts about the knowledge graph. It gets people the helpful information they need and creates new opportunities for sites to be discovered.”

It’s hard to say whether or not this will bring search traffic to websites and what the quality of that traffic will be. Will they stick around to read an article? Will they engage with a product review?

Visual Search Results

Sylvanus shares a hypothetical example of someone at an airport baggage claim who falls in like with someone else’s bag. He explains that all the person needs to do is snap a photo of the luggage bag and Google Lens will take them directly to shopping options.

He explains:

“No words, no problem. Just open Lens, take a quick picture and immediately you’ll see options to purchase.

And for the first time, shopping ads will appear at the very top of the results on linked searches, where a business can offer what a consumer is looking for.

This will help them easily purchase something that catches their eye.”

These are image-heavy shopping ads at the top of the search results and as annoying as that may be it’s nowhere near the “next level” advertising that is coming to Google’s search ads where Google presents a paid promotion within the context of an AI Assistant.

Interactive Search Shopping

Sylvanus next describes an AI-powered form advertising that happens directly within search. But he doesn’t call it advertising. He doesn’t even use the word advertising. He suggests this new form of AI search experience is more than offer, saying that, “it’s an experience.”

He’s right to not use the word advertisement because what he describes goes far beyond advertising and blurs the boundaries between search and advertising within the context of AI-powered suggestions, paid suggestions.

Sylvanus explains how this new form of shopping experience works:

“And next, imagine a world where every search ad is more than an offer. It’s an experience. It’s a new way for you to engage more directly with your customers. And we’re exploring search ads with AI powered recommendations across different verticals. So I want to show you an example that’s going live soon and you’ll see even more when we get to shopping.”

He uses the example of someone who needs to store their furniture for a few months and who turns to Google to find short term storage. What he describes is a query for local short term storage that turns into a “dynamic ad experience” that leads the searcher into throwing packing supplies into their shopping cart.

He narrated how it works:

“You search for short term storage and you see an ad for extra space storage. Now you can click into a new dynamic ad experience.

You can select and upload photos of the different rooms in your house, showing how much furniture you have, and then extra space storage with help from Google, AI generates a description of all your belongings for you to verify. You get a recommendation for the right size and type of storage unit and even how much packing supplies you need to get the job done. Then you just go to the website to complete the transaction.

And this is taking the definition of a helpful ad to the next level. It does everything but physically pick up your stuff and move it, and that is cool.”

Step 1: Search For Short Term Storage

1716722762 15 New Google Search Ads Resemble AI Assistant App

The above screenshot shows an advertisement that when clicked takes the user to what looks like an AI-assisted search but is really an interactive advertisement.

Step 2: Upload Photos For “AI Assistance”

1716722762 242 New Google Search Ads Resemble AI Assistant App

The above image is a screenshot of an advertisement that is presented in the context of AI-assisted search.  Masking an advertisement within a different context is the same principal behind an advertorial where an advertisement is hidden in the form of an article. The phrases “Let AI do the heavy lifting” and “AI-powered recommendations” create the context of AI-search that masks the true context of an advertisement.

Step 3: Images Chosen For Uploading

1716722762 187 New Google Search Ads Resemble AI Assistant App

The above screenshot shows how a user uploads an image to the AI-powered advertisement within the context of an AI-powered search app.

The Word “App” Masks That This Is An Ad

Screenshot of interactive advertisement for that identifies itself as an app with the words

Above is a screenshot of how a user uploads a photo to the AI-powered interactive advertisement within the context of a visual search engine, using the word “app” to further the illusion that the user is interacting with an app and not an advertisement.

Upload Process Masks The Advertising Context

Screenshot of interactive advertisement that uses the context of an AI Assistant to mask that this is an advertisement

The phrase “Generative AI is experimental” contributes to the illusion that this is an AI-assisted search.

Step 4: Upload Confirmation

1716722762 395 New Google Search Ads Resemble AI Assistant App

In step 4 the “app” advertisement is for confirming that the AI correctly identified the furniture that needs to be put into storage.

Step 5: AI “Recommendations”

1716722762 588 New Google Search Ads Resemble AI Assistant App

The above screenshot shows “AI recommendations” that look like search results.

The Recommendations Are Ad Units

1716722762 751 New Google Search Ads Resemble AI Assistant App

Those recommendations are actually ad units that when clicked takes the user to the “Extra Space Storage” shopping website.

Step 6: Searcher Visits Advertiser Website

1716722762 929 New Google Search Ads Resemble AI Assistant App

Blurring The Boundaries

What the Google keynote speaker describes is the integration of paid product suggestions into an AI assisted search. This kind of advertising is so far out there that the Googler doesn’t even call it advertising and rightfully so because what this does is blur the line between AI assisted search and advertising. At what point does a helpful AI search become just a platform for using AI to offer paid suggestions?

Watch The Keynote At The 32 Minute Mark

Featured Image by Shutterstock/Ljupco Smokovski

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