Google CEO: Search Will Change Profoundly In 2025 via @sejournal, @martinibuster

Sundar Pichai, Google’s CEO, was interviewed by Andrew Ross Sorkin at the New York Times DealBook Summit, where he discussed what to expect from Google Search in 2025 but also struggled to articulate Google’s concern for content creators.

When asked to compare where Google is today relative to the rest of the industry and whether Google should be the “default winner” Pichai reminded the interviewer that these were “the earliest stages of a profound shift” and underlined that Google is a leader in AI and not the follower. The entire AI industry is built on top of Google research discoveries that were subsequently open sourced, particularly transformers, without which the AI industry would not exist as it is today.

Pichai answered:

“Look, it’s a such a dynamic moment in the industry. When I look at what’s coming ahead, we are in the earliest stages of a profound shift. We have taken such a deep full stack approach to AI.

…we do world class research. We are the most cited, when you look at gen AI, the most cited… institution in the world, foundational research, we build AI infrastructure and when I’m saying AI infrastructure all the way from silicon, we are in our sixth generation of tensor processing units. You mentioned our product reach, we have 15 products at half a billion users, we are building foundational models, and we use it internally, we provide it to over three million developers and it’s a deep full stack investment.

We are getting ready for our next generation of models, I just think there’s so much innovation ahead, we are committed to being at the state of the art in this field and I think we are. Just coming today, we announced groundbreaking research on a text and image prompt creating a 3D scene. And so the frontier is moving pretty fast, so looking forward to 2025.”

Blue Link Economy And AI

It was pointed out by the interviewer that Google was the first mover on AI and then it wasn’t (a reference to OpenAI’s breakout in 2022 and subsequent runaway success). He asked Pichai how much of that was Google protecting the “blue link economy” so as not “to hurt or cannibalize that business” which is worth hundreds of billions of dollars.

Pichai answered that out of all the projects at Google, AI was applied the most to Search, citing BERT, MUM and multimodal search as helping close the gaps in search quality. Something that some in the search industry fail to understand is that AI has has been a part of Google since 2012 when it used Deep Neural Networks for identifying images and speech recognition and in 2014 when it introduced the world to sequence to sequence learning (PDF) for understanding strings of text. In 2015 Google introduced RankBrain, an AI system directly related to ranking search results.

Pichai  answered:

“The area where we applied AI the most aggressively, if anything in the company was in search, the gaps in search quality was all based on Transformers internally. We call it BERT and MUM and you know, we made search multimodal, the search quality improvements, we were improving the language understanding of search. That’s why we built Transformers in the company.

So and if you look at the last couple of years, we have with AI overviews, Gemini is being used by over a billion users in search alone.”

Search Will Change Profoundly In 2025

Pichai continued his answer, stating directly that Search will profoundly change not just in 2025, but in early 2025. He also said that progress is going to get harder because the easier things to innovate have been done (low hanging fruit).

He said:

“And I just feel like we are getting started. Search itself will continue to change profoundly in 2025. I think we are going to be able to tackle more complex questions than ever before. You know, I think we’ll be surprised even early in 2025, the kind of newer things search can do compared to where it is today… “

Pichai also said that progress wouldn’t be easy:

“I think the progress is going to get harder when I look at 2025, the low hanging fruit is gone.

But I think where the breakthroughs need to come from where the differentiation needs to come from is is your ability to achieve technical breakthroughs, algorithmic breakthroughs, how do you make the systems work, you know, from a planning standpoint or from a reasoning standpoint, how do you make these systems better? Those are the technical breakthroughs ahead.”

Is Search Going Away?

The interviewer asked Pichai if Google has leaned into AI enough, quoting an author who suggested that Google’s “core business is under siege” because people are increasingly getting answers from AI and other platforms outside of search, and that the value of search would be “deteriorating” because so much of the content online will be AI-generated.

He answered that it’s precisely in a scenario where the Internet is flooded with inauthentic content that search becomes even more valuable.

Pichai answered:

“In a world in which you’re flooded with like lot of content …if anything, something like search becomes more valuable. In a world in which you’re inundated with content, you’re trying to find trustworthy content, content that makes sense to you in a way reliably you can use it, I think it becomes more valuable.

To your previous part about there’s a lot of information out there, people are getting it in many different ways. Look, information is the essence of humanity. We’ve been on a curve on information… when Facebook came around, people had an entirely new way of getting information, YouTube, Facebook, Tik… I can keep going on and on.

…I think the problem with a lot of those constructs is they are zero sum in their inherent outlook. They just feel like people are consuming information in a certain limited way and people are all dividing that up. But that’s not the reality of what people are doing. “

Pichai Stumbles On Question About Impact On Creators

The interviewer next asked if content is being devalued. He used the example of someone who researches a topic for a book, reads twenty books, cites those sources in the bibliography and then gets it published. Whereas Google ingests everything and then “spits” out content all day long, defeating the human who in earlier times would write a book.

Andrew Ross Sorkin said:

“You get to spit it out a million times. A million times a day. And I just wonder what the economics of that should be for the folks that create it in the beginning.”

Sundar Pichai defended Google by saying that Google spends a lot of time thinking about the impact to the “ecosystem” of publishers and how much traffic it sends to them. The interviewer listened to Sundar’s answer without mentioning the elephant in the room, search results stuffed with Reddit and advertising that crowds out content created by actual experts, and the de-prioritization of news content which has negatively impacted traffic to news organizations around the world.

It was at this point that Pichai appeared to stumble as he tried to find the words to respond. He avoids mentioning websites, speaking in the abstract about the “ecosystem” and then when he runs out of things to say changes course and begins speaking about how Google compensates copyright holders who sign up for YouTube’s Content ID program.

He answered:

“Look I… uh… It’s a… very important question… uhm… look I… I… think… I think more than any other company… look you know… we for a long time through… you know… be it in search making sure… while it’s often debated, we spend a lot of time thinking about the traffic we send to the ecosystem.

Even through the moment through the transition over the past couple of years. It’s an important priority for us.”

At this point he started talking about Google’s content platform YouTube and how they use “Content ID” which is used to identify copyright-protected content. Content ID is a program that benefits the corporate music, film, and television industries, copyright owners who “own exclusive rights to a substantial body of original material that is frequently uploaded to YouTube.”

Pichai continued:

“In YouTube we put a lot of effort into understanding and you know identifying content and with content ID and uh creating monetization for creators.

I think… I think those are important principles, right. I think um… there’s always going to be a balance between understanding what is fair use uh… when new technology comes versus how do you… give value back proportionate to the value of the IP, the hard work people have put in.”

Insightful Interview Of Alphabet’s CEO

The interviewer did a great job at asking the hard questions but I think many in the search marketing community who are more familiar with the search results would have asked follow up questions about content creators who are not on Google’s YouTube platform or the non-expert content that pushes down content by actual experts.

Watch the New York Times Interview here:

Featured Image by Shutterstock/Shutterstock AI Generator
(no irony intended)

Google Rolls Out One-Click Event Tracking In GA4 via @sejournal, @MattGSouthern

Google simplifies analytics tracking with new one-click key event features in GA4, powered by machine learning.

  • Google released one-click event tracking in GA4 with two features: “Mark as key event” and “Create key event.”
  • Machine learning identifies important site events automatically, eliminating manual setup time.
  • These features are now available for all GA4 properties and enable better tracking and reporting.
Google: Focus On Field Data For Core Web Vitals via @sejournal, @MattGSouthern

Google stresses the importance of using actual user data to assess Core Web Vitals instead of relying only on lab data from tools like PageSpeed Insights (PSI) and Lighthouse.

This reminder comes as the company prepares to update the throttling settings in PSI. These updates are expected to increase the performance scores of websites in Lighthouse.

Field Data vs. Lab Data

Core Web Vitals measure a website’s performance in terms of loading speed, interactivity, and visual stability from the user’s perspective.

Field data shows users’ actual experiences, while lab data comes from tests done in controlled environments using tools like Lighthouse.

Barry Pollard, a Web Performance Developer Advocate at Google, recently emphasized focusing on field data.

In a LinkedIn post, he stated:

“You should concentrate on your field Core Web Vitals (the top part of PageSpeed Insights), and only use the lab Lighthouse Score as a very rough guide of whether Lighthouse has recommendations to improve performance or not…

The Lighthouse Score is best for comparing two tests made on the same Lighthouse (e.g. to test and compare fixes).

Performance IS—and hence LH Scores also ARE—highly variable. LH is particularly affected by where it is run from (PSI, DevTools, CI…), but also on the lots of other factors.

Lighthouse is a GREAT tool but it also can only test some things, under certain conditions.

So while it’s great to see people interested in improving webperf, make sure you’re doing just that (improve performance) and not just improving the score”

Upcoming Changes To PageSpeed Insights

Pollard discussed user concerns about PageSpeed Insights’s slow servers, which can cause Lighthouse tests to take longer than expected.

To fix this, Google is changing the throttling settings in PageSpeed Insights, which should lead to better performance scores when the update is released in the coming weeks.

These changes will affect both the web interface and the API but will not impact other versions of Lighthouse.

However, Pollard  reminds users that  “a score of 100 doesn’t mean perfect; it just means Lighthouse can’t help anymore.”

Goodhart’s Law & Web Performance

Pollard referenced Goodhart’s Law, which says that when a measure becomes a goal, it stops being a good measure.

In the web performance context, focusing only on improving Lighthouse scores may not improve actual user experience.

Lighthouse is a helpful tool, but it can only assess certain aspects of performance in specific situations.

Alon Kochba, Web Performance and Software Engineer at Wix, added context to the update, stating:

“Lighthouse scores may not be the most important – but this is a big deal for Lighthouse scores in PageSpeed Insights.

4x -> 1.2x CPU throttling for Mobile device simulation, which was way off for quite a while.”

Key Takeaway: Prioritize User Experience

As the update rolls out, website owners and developers should focus on user experience using field data for Core Web Vitals.

While Lighthouse scores can help find areas for improvement, they shouldn’t be the only goal.

Google encourages creating websites that load quickly, respond well, and are visually stable.


Featured Image: GoodStudio/Shutterstock

How To Optimize TikTok Ads After You Set Up Your Campaign via @sejournal, @brookeosmundson

TikTok isn’t just a place for dance challenges or trendy lip-sync videos; it’s now one of the most powerful platforms for advertisers.

With its unique mix of authenticity and creativity, TikTok offers marketers a golden opportunity to reach engaged audiences in fresh ways.

But setting up a campaign is only the beginning. Once your ads are live, optimizing them to meet your goals is where the real magic happens.

In this guide, we’ll dive into practical, data-driven steps to improve TikTok ad performance after setup.

Whether your goal is to boost brand awareness or drive conversions, this article will show you how to refine your strategy for maximum impact.

1. Know Your Objectives And KPIs

Before diving into optimization, it’s essential to clarify your campaign’s objectives and identify key performance indicators (KPIs).

Are you aiming for brand awareness, engagement, or conversions?

Each objective requires different tactics and metrics to gauge success.

Review KPIs like impressions, clicks, click-through rates (CTR), and conversions regularly. Tracking these lets you adjust your approach based on how close you are to your goals.

For instance, if your primary goal is brand awareness, focus on engagement rates rather than immediate conversions. Keeping your objectives front and center will help shape every decision as you optimize.

2. Audience Refinement And Targeting Adjustments

TikTok’s wide-reaching audience is one of its strengths, but casting too broad a net can dilute your results.

Once you have initial data from your live ads, it’s time to refine your targeting:

  • Demographic Adjustments: Analyze the demographics of users interacting with your ads. Adjust your target audience to match the profiles that are responding best. For example, if you see a strong engagement from a particular age group or gender, double down there.
  • Interest And Behavior-Based Refinement: TikTok allows for interest-based targeting, which can help you reach specific niches. If you see that users interested in “travel” engage with your content more, consider adjusting your interests or behaviors to focus on that.
  • Custom And Lookalike Audiences: Custom audiences (based on engagement or website traffic) and lookalike audiences (based on your existing customers) are invaluable for re-targeting and prospecting. These segments let you focus your budget on high-intent users who are likely to convert, creating a win-win for efficiency and results.

Additionally, TikTok Ads has the ability to create “AND” audiences, allowing for further refinement based on behavioral performance.

For example, you can create an audience that includes a certain demographic factor, AND portrays a specific interest. This type of targeting narrows down the focus instead of using an “OR” behavior, which targets anyone in either of those categories chosen.

3. Optimize Your Ad Creative

TikTok is all about eye-catching content, so optimizing ad creative is non-negotiable.

Dive into your metrics – CTR, engagement rate, completion rate – and identify trends in what’s working.

  • A/B Testing: Split-testing different video creations can provide insights into what resonates most. Test elements like video length, style, and music choice to see what improves results.
  • Trendy, Authentic Content: TikTok users are drawn to authenticity. Swap polished brand visuals for raw, relatable content that feels natural in users’ feeds. Experiment with user-generated content (UGC) styles to tap into TikTok’s unique vibe.
  • Adjust Visuals Regularly: TikTok trends move fast. Refresh visuals, sounds, and messaging to stay relevant and avoid “ad fatigue,” where users tune out repetitive content.

If possible, try to have five to 10 creatives per ad group running at all times to reduce creative fatigue.

When first starting out, try testing creatives with big differences between the two, not small nuances, to really understand what’s going to capture your audience from the start.

These steps help you create a feedback loop where the data from each test iteration informs the next, sharpening your creative with each optimization cycle.

TikTok has a variety of creative tools in its Business Help Center, including:

  • Video template tool.
  • Smart creative.
  • Creative exchange.
  • Creative best practices.
  • And more.

4. Use Automated Bidding Strategically

TikTok’s automated bidding options, like “Maximum Delivery” and “Bid Cap,” can help streamline budget management. Knowing when to use them effectively, though, is key.

  • Start With Manual, Then Test Automated: Many advertisers find value in starting with manual bidding for tighter control, then testing TikTok’s automated options once they gather performance data.
  • Bid Adjustments Based On Engagement: If an ad shows high engagement but low conversions, consider using manual bid adjustments to focus spend on high-engagement periods or audiences.

Here’s a quick breakdown of the two automated bid strategy options:

Cost Cap Bidding

  • How It Works: Keeps your average cost similar to your bid, regardless of daily/lifetime budget.
  • Advertising Objectives available: App Installs, Conversions, or Lead Generation objectives can be chosen with this strategy.
  • Cost Metrics: Cost per Click (CPC), Cost per Thousand (CPM), Cost per View (CPV), and Optimized Cost per Thousand (oCPM).

Maximum Delivery Bidding

  • How It Works: Maximizes budget usage by driving the most results possible, given a specified budget and a certain time period.
  • Advertising Objectives available: Traffic, Reach, Video Views, App Installs, Conversions, Lead Generation, and Catalog Sales objectives can be chosen with this strategy.
  • Cost Metrics: CPC, CPM, CPV, oCPM.

Automated bidding is particularly beneficial when managing multiple campaigns, but always monitor closely.

TikTok’s bidding algorithm can free up time, but a “set it and forget it” approach rarely leads to optimal results.

5. Ad Placement And Schedule Adjustments

TikTok gives you control over placement and scheduling – two levers you can use to ensure your ads reach the right audience at the right time.

By default, TikTok will set an ad group with Automatic Placements, meaning your ad will be available to be chosen in each available ad placement (depending on your ad type, bid strategy, etc.)

If you’re looking to harness more control at the beginning, you can turn that setting off. The following manual placements can be selected or de-selected:

  • TikTok: Allows for placement on TikTok based on your advertising objective.
  • Global App Bundle: An integrated traffic solution that places ads on other popular apps, like CapCut and Fizzo.
  • Pangle: The ad network of TikTok for business that allows you to reach users across top local publishers on the platform.

Below are a few tips for getting started with ad placement and schedule adjustments.

  • Review Placement Performance: Ads can show up in different TikTok placements. Start by examining how each is performing. If the “For You” feed drives better engagement than other placements, consider focusing your budget there.
  • Dayparting For Maximum Impact: Dayparting – scheduling ads to run at specific times – can improve efficiency. Look at when your target audience is most active and adjust your schedule to maximize impact during those hours.

These adjustments ensure your ads are not just “on” but active in ways that maximize visibility and relevance.

6. Retargeting And Sequencing For Engagement

As users engage with your brand on TikTok, retargeting allows you to keep them in the loop without starting from scratch.

  • Engagement-Based Retargeting: Use retargeting to reach users who have interacted with your brand but haven’t converted. Target viewers who’ve engaged with your videos or clicked but haven’t completed an action. This high-intent group is more likely to convert with the right nudge.
  • Sequencing Ads To Tell A Story: Ad sequencing is a powerful way to build a narrative across multiple videos. A sequential approach lets you nurture leads by gradually building brand familiarity and trust, guiding users down the conversion funnel naturally.

Considering 72% of TikTok users agree that a brand is memorable even three weeks after initial ad exposure, remarketing efforts on TikTok are a no-brainer.

These techniques can boost engagement and help guide audiences from mere interest to genuine investment in your brand.

7. Analyze And Optimize Landing Pages

If your landing page or app experience doesn’t align with your TikTok ad, you’ll likely see high bounce rates.

As with any ad, optimize those landing pages for a seamless transition from ad to action.

  • A/B Test Landing Pages: Experiment with different layouts, messaging, and calls to action on your landing pages. TikTok users prefer a smooth, distraction-free experience, so every element on your landing page should support your ad’s call to action.
  • Page Speed Matters: TikTok is a fast-paced platform, and users expect speed. Ensure your landing pages load quickly to reduce drop-off.

In short, your ad and landing page should feel like a cohesive journey that’s frictionless and aligned with your audience’s expectations.

8. Experiment With TikTok’s Advanced Features

As TikTok evolves, so do its advertising options.

Advanced features like Spark Ads, Dynamic Showcase Ads, and Catalog Sales provide unique ways to reach audiences.

  • Spark Ads For Organic Feel: Spark Ads allow brands to partner with influencers or use existing organic posts as ads. This format enhances trust and engagement as it appears less like a traditional ad and more like organic content.
  • Dynamic Showcase Ads And Catalog Sales: For ecommerce brands, these features let you showcase products dynamically, creating a more personalized ad experience. Optimize these ads by using high-quality visuals and up-to-date catalog information.

Taking advantage of these advanced features allows your brand to stay ahead of the curve and engage audiences in innovative ways.

Embrace Optimization With Continuous Testing

The TikTok landscape moves quickly, making continuous testing essential for ad performance. Every ad format, audience adjustment, and creative tweak contributes to your overall learning.

Setting up TikTok ads is just the first step in a larger optimization journey.

To unlock the full potential of TikTok, marketers need to embrace data-driven adjustments, experiment with new formats, and consistently align their content with TikTok’s fast-moving trends.

By refining targeting, leveraging creative testing, and optimizing placements, you can push your campaigns to perform at their best.

Let’s face it: TikTok is one of the most dynamic platforms today. With a bit of experimentation and a lot of flexibility, your ads can thrive here, bringing both visibility and results to your brand.

More resources:


Featured Image: Roman Samborskyi/Shutterstock

The US Department of Defense is investing in deepfake detection

The US Department of Defense has invested $2.4 million over two years in deepfake detection technology from a startup called Hive AI. It’s the first contract of its kind for the DOD’s Defense Innovation Unit, which accelerates the adoption of new technologies for the US defense sector. Hive AI’s models are capable of detecting AI-generated video, image, and audio content. 

Although deepfakes have been around for the better part of a decade, generative AI has made them easier to create and more realistic-looking than ever before, which makes them ripe for abuse in disinformation campaigns or fraud. Defending against these sorts of threats is now crucial for national security, says Captain Anthony Bustamante, a project manager and cyberwarfare operator for the Defense Innovation Unit.

“This work represents a significant step forward in strengthening our information advantage as we combat sophisticated disinformation campaigns and synthetic-media threats,” says Bustamante. Hive was chosen out of a pool of 36 companies to test its deepfake detection and attribution technology with the DOD. The contract could enable the department to detect and counter AI deception at scale.

Defending against deepfakes is “existential,” says Kevin Guo, Hive AI’s CEO. “This is the evolution of cyberwarfare.”

Hive’s technology has been trained on a large amount of content, some AI-generated and some not. It picks up on signals and patterns in AI-generated content that are invisible to the human eye but can be detected by an AI model. 

“Turns out that every image generated by one of these generators has that sort of pattern in there if you know where to look for it,” says Guo. The Hive team constantly keeps track of new models and updates its technology accordingly. 

The tools and methodologies developed through this initiative have the potential to be adapted for broader use, not only addressing defense-specific challenges but also safeguarding civilian institutions against disinformation, fraud, and deception, the DOD said in a statement.

Hive’s technology provides state-of-the-art performance in detecting AI-generated content, says Siwei Lyu, a professor of computer science and engineering at the University at Buffalo. He was not involved in Hive’s work but has tested its detection tools. 

Ben Zhao, a professor at the University of Chicago, who has also independently evaluated Hive AI’s deepfake technology, agrees but points out that it is far from foolproof. 

“Hive is certainly better than most of the commercial entities and some of the research techniques that we tried, but we also showed that it is not at all hard to circumvent,” Zhao says. The team found that adversaries could tamper with images in a way that bypassed Hive’s detection.

And given the rapid development of generative AI technologies, it is not yet certain how it will fare in real-world scenarios that the defense sector might face, Lyu adds.  

Guo says Hive is making its models available to the DOD so that the department can use the tools offline and on their own devices. This keeps sensitive information from leaking.

But when it comes to protecting national security against sophisticated state actors, off-the-shelf products are not enough, says Zhao: “There’s very little that they can do to make themselves completely robust to unforeseen nation-state-level attacks.” 

New Ecommerce Tools: December 5, 2024

Every week we curate and publish a list of new products from companies offering services to ecommerce merchants. This installment includes updates on virtual stores, tax management, cross-border commerce, circular packaging, sponsored product display, product images, and quick commerce.

Got an ecommerce product release? Email releases@practicalecommerce.com.

New Tools for Merchants

Walmart updates virtual store with seasonal shopping features. Expanding its Realm experience with virtual-world developer Emperia, Walmart has added four new virtual shops themed as mountain chalet, holiday party, the white elephant game, and the sweetest gifts. Influencers curate gift ideas within the virtual spaces. Walmart also debuted a feature to create and share wish lists. Earlier this year, Walmart launched Realm with three environments.

Web page for Walmart Realm

Walmart Realm

Shopify supports multi-currency payouts. Shopify now supports payouts in multiple currencies for European merchants. The feature is available for users of Shopify Advanced and Plus in six European countries: the United Kingdom, Germany, Spain, Italy, The Netherlands, and Austria. Payout currencies include the U.S. dollar, Australian dollar, Canadian dollar, Danish krone, euro, Hungarian forint, Hong Kong dollar, Japanese yen, New Zealand dollar, Norwegian krone, Polish złoty, Romanian leu, Singapore dollar, South African rand, Swedish krona, Swiss franc, and Czech koruna.

Solidgate launches cross-regional tax management platform. Solidgate, a payment processing platform, has launched its tax compliance portal, a self-service, pay-per-use solution to simplify VAT and sales tax management for multinational businesses. The new portal automatically synchronizes data from hundreds of payment processors — including Worldpay, Adyen, Checkout.com, Nuvei, and Stripe — and calculates applicable taxes based on predefined parameters, such as customer location and transaction type.

Movopack raises funds to launch circular packaging in U.K. Movopack has secured $2.5 million in seed funding to launch its circular ecommerce packaging in the U.K. The company produces its reusable packaging from recycled plastic bottles and recycled polypropylene (plastic), designed to withstand postal handling and be reused up to 20 times. Movopack’s packaging can be returned via Royal Mail. The E.U. mandates that at least 10% of ecommerce packaging must be reusable by 2030, increasing to 50% by 2040.

Home page of Movopack

Movopack

Flipkart to expand quick commerce to more cities. Flipkart, an India-based ecommerce company backed by Walmart, is planning to expand Minutes, its fast delivery service, amid rising demand for 10-minute deliveries. Flipkart Minutes is currently live in the pilot stage in Bengaluru, Mumbai, and Delhi. Flipkart began piloting Minutes in Bengaluru in August and expanded to the Delhi region in September. Flipkart’s move came in response to the rise of Zomato-owned Blinkit, Swiggy Instamart, Zepto, and Tata-owned BigBasket.

Walmart and DoorDash partner across Canada. DoorDash Canada and Walmart Canada have announced a collaboration spanning 300 Walmart Supercenters across 10 Canadian provinces. Customers shopping on DoorDash can now browse thousands of unique items from Walmart, ranging from fresh produce and kitchenware to pet food and electronics, including private-label brands. All Walmart Supercenters on DoorDash will be available on DashPass (a subscription program), providing $0 delivery fees and reduced service fees on eligible orders.

GroupBy and Topsort partner to optimize sponsored product display. GroupBy, a provider of ecommerce search and product discovery, has announced an integration with Topsort, an AI and auction-based retail media infrastructure company. By combining GroupBy’s AI-driven platform with Topsort’s retail media capabilities, merchants can increase visibility for sponsored products, ensuring only relevant products will be sent for validation in Topsort’s retail media platform.

Home page of GroupBy

GroupBy

Aiarty Image Matting released for ecommerce photos. Aiarty, part of Digiarty Software, has launched Image Matting, an AI-powered tool to streamline image processing for ecommerce businesses. Following the release of Aiarty’s Image Enhancer, Image Matting utilizes AI models for background removal and bulk background replacement. Together, the tools enable ecommerce sellers to enhance product visuals across channels and remove and replace the backgrounds on up to 3,000 images at once.

HSBC launches a platform for domestic and international payments. HSBC, a London-based bank and financial services provider, has launched Smart Transact, a one-stop domestic and cross-border payments platform. Users can access additional services, including corporate cards and integrated HSBC tools. Smart Transact is available in India, the U.K., the U.S., Hong Kong, Singapore, Australia, Ireland, The Netherlands, and France.

ImageKit unveils API to simplify video post-production and delivery. ImageKit, a developer of media solutions, has launched its video API to automate the time-consuming and repetitive tasks required to publish videos across the web. ImageKit allows users to resize and transform videos in real-time by modifying URL parameters and optimizing for any device or platform without manual effort. The video API features real-time tools such as resizing and cropping, adaptive streaming capabilities, overlays, styling options for captions, and more.

Mollie and PayPal partner on payments for marketplace platforms. Mollie, a payment processing company, has partnered with PayPal to deliver payment solutions for marketplace platforms across Europe. The partnership enhances Mollie’s range of Europe-based payment options, including credit and debit cards, digital wallets, bank transfers, buy-now-pay-later options, and local payment methods.

Home page of Mollie

Mollie

Keywordless Search Ads Are Coming

I’ve run pay-per-click campaigns for nearly 20 years and witnessed significant change. Tactics I once considered cutting-edge are now outdated. Exciting new strategies have emerged, which I enjoy learning and testing.

In this post, I’ll share a momentous change I believe we’ll see in 2025.

Keywordless Search Campaigns

The “death of the keyword” prediction isn’t new. It suggests that keywords will go away as an optimization tactic, replaced by audiences and other factors.

Yet Search campaigns without keywords already exist. Dynamic Search Ads and Performance Max campaigns match user queries to a site’s content. Advertisers can see the queries, but there are no keywords to bid on. It’s a holistic shift to keywordless campaigns and hints at the future.

Google has redefined how broad match works. Queries that do not relate to keywords can trigger ads based on other factors, such as previous searches and the user’s location. For example, an advertiser bidding in broad match on the keyword “office desk accessories” could trigger an ad for “coffee mugs” if the user’s previous searches indicated an interest in desk items.

Based on that logic and how Google triggers ads for phrase and exact match, true match types no longer exist. Google decides which queries are relevant. Keywords are a component, not the sole focus.

Google says broad match uses the most audience signals and is thus the most effective for advertisers. I disagree, at least partially, because I still see better performance with phrase and exact match. But the writing is on the wall: Match type isn’t critical.

Interface in Google Ads admin for broad match keywords

Per Google, broad match uses the most audience signals and performs best for advertisers. Click image to enlarge.

A campaign beta in Performance Max called “search themes” confirms the direction. Instead of submitting keywords, advertisers enter themes related to their product or service. Advertisers selling recliners, for example, could enter “recliners,” “recliner chairs,” and “sofa recliners.”

Keywordless search allows Google to cast a wider net with more data and rely on machine learning to show the right ad.

Same Structure

Despite the absence of keywords, I don’t see the structure of Search campaigns changing. Campaigns organized by topics with ad groups in sub-topics will continue. Ad copy will reflect the ad group. Targeting will be a mix of audience signals, including search themes, and enhanced by first-party data.

Still, an advertiser might question the removal of keywords from Search campaigns when Performance Max targets the Search network. Why bother with a Search campaign if the targeting options are the same as Performance Max?

I see a couple of reasons.

First, advertisers cannot choose the networks in Performance Max campaigns. Ads show across all of Google’s properties. The setting is fixed, not changeable. Hence advertisers interested only in the Search network should not choose Performance Max — Search campaigns are a better option.

Second, search has been the core of Google for nearly 25 years. Removing the Search campaign type — even without keywords — would be too disruptive to advertisers who view Performance Max as not viable.

Therefore Google will likely not force advertisers to transition from Search. Google will encourage keywordless campaigns but will not make them mandatory.

OpenAI Releases ChatGPT o1, ‘World’s Smartest Language Model” via @sejournal, @martinibuster

Today OpenAI rolled out what Sam Altman says is the world’s smartest language model in the world plus a brand new Pro tier that comes with unlimited usage limits and with a higher level of computing resources.

OpenAI ChatGPT o1 Model

Sam Altman announced on X (formerly Twitter) that their new AI model is now live and available in ChatGPT right now and will be arriving to the API soon.

He tweeted:

“o1, the smartest model in the world. smarter, faster, and more features (eg multimodality) than o1-preview. live in chatgpt now, coming to api soon.

chatgpt pro. $200/month. unlimited usage and even-smarter mode for using o1. more benefits to come!”

Screenshot Of ChatGPT 01 Model Availability

ChatGPT Pro Mode $200/Month

ChatGPT Pro Mode is a new tier that has more “thinking power” than the standard version of o1, which increases it’s reliability. Answers in Pro mode take longer to generate, displaying a progress bar and triggering an in-app notification if the user navigates to a different conversation.

OpenAI describes the new ChatGPT Pro Mode:

“ChatGPT Pro provides access to a version of our most intelligent model that thinks longer for the most reliable responses. In evaluations from external expert testers, o1 pro mode produces more reliably accurate and comprehensive responses, especially in areas like data science, programming, and case law analysis.

Compared to both o1 and o1-preview, o1 pro mode performs better on challenging ML benchmarks across math, science, and coding.”

The new tier is not a price increase from the regular plan, which is called Plus. It’s an entirely new plan called Pro.

OpenAI’s new o1 Pro plan provides unlimited access to its new o1 model, along with o1-mini, GPT-4o, and Advanced Voice. It also includes o1 Pro Mode, which has access to increased computational power to generate more refined and insightful responses to complex queries.

Read more about OpenAI’s new pro plan and O1 model:

Introducing ChatGPT Pro

Featured Image by Shutterstock/One Artist

Google Uses About 40 Signals To Determine Canonical URLs via @sejournal, @MattGSouthern

In a recent episode of Google’s Search Off the Record podcast, Allan Scott from the “Dups” team explained how Google decides which URL to consider as the main one when there are duplicate pages.

He revealed that Google looks at about 40 different signals to pick the main URL from a group of similar pages.

Around 40 Signals For Canonical URL Selection

Duplicate content is a common problem for search engines because many websites have multiple pages with the same or similar content.

To solve this, Google uses a process called canonicalization. This process allows Google to pick one URL as the main version to index and show in search results.

Google has discussed the importance of using signals like rel=”canonical” tags, sitemaps, and 301 redirects for canonicalization. However, the number of signals involved in this process is more than you may expect.

Scott revealed during the podcast:

“I’m not sure what the exact number is right now because it goes up and down, but I suspect it’s somewhere in the neighborhood of 40.”

Some of the known signals mentioned include:

  1. rel=”canonical” tags
  2. 301 redirects
  3. HTTPS vs. HTTP
  4. Sitemaps
  5. Internal linking
  6. URL length

The weight and importance of each signal may vary, and some signals, like rel=”canonical” tags, can influence both the clustering and canonicalization process.

Balancing Signals

With so many signals at play, Allan acknowledged the challenges in determining the canonical URL when signals conflict.

He stated:

“If your signals conflict with each other, what’s going to happen is the system will start falling back on lesser signals.”

This means that while strong signals like rel=”canonical” tags and 301 redirects are crucial, other factors can come into play when these signals are unclear or contradictory.

As a result, Google’s canonicalization process involves a delicate balancing act to determine the most appropriate canonical URL.

Best Practices For Canonicalization

Clear signals help Google identify the preferred canonical URL.

Best practices include:

  1. Use rel=”canonical” tags correctly.
  2. Implement 301 redirects for permanently moved content.
  3. Ensure HTTPS versions of pages are accessible and linked.
  4. Submit sitemaps with preferred canonical URLs.
  5. Keep internal linking consistent.

These signals help Google find the correct canonical URLs, improving your site’s crawling, indexing, and search visibility.

Mistakes To Avoid

Here are a few common mistakes to watch out for.

1. Incorrect or conflicting canonical tags:

  • Pointing to non-existent or 404 pages
  • Multiple canonical tags with different URLs on one page
  • Pointing to a different domain entirely

Fix: Double-check canonical tags, use only one per page, and use absolute URLs.

2. Canonical chains or loops

When Page A points to Page B as canonical, but Page B points back to A or another page, creating a loop.

Fix: Ensure canonical URLs always point to the final, preferred version of the page.

3. Using noindex and canonical tags together

Sending mixed signals to search engines. Noindex means don’t index the page at all, making canonicals irrelevant.

Fix: Use canonical tags for consolidation and noindex for exclusion.

4. Canonicalizing to redirect or noindex pages

Pointing canonicals to redirected or noindex pages confuses search engines.

Fix: Canonical URLs should be 200 status and indexable.

5. Ignoring case sensitivity

Inconsistent URL casing can cause duplicate content issues.

Fix: Keep URL and canonical tag casing consistent.

6. Overlooking pagination and parameters

Paginated content and parameter-heavy URLs can cause duplication if mishandled.

Fix: Use canonical tags pointing to the first page or “View All” for pagination, and keep parameters consistent.

Key Takeaways

It’s unlikely the complete list of 40+ signals used to determine canonical URLs will be made publicly available.

However, this was still an insightful discussion worth highlighting.

Here are the key takeaways:

  • Google uses approximately 40 different signals to determine canonical URLs, with rel=”canonical” tags and 301 redirects being among the strongest indicators
  • When signals conflict, Google falls back on secondary signals to make its determination
  • Clear, consistent implementation of canonicalization signals (tags, redirects, sitemaps, internal linking) is crucial
  • Common mistakes like canonical chains, mixed signals, or incorrect implementations can confuse search engines

Hear the full discussion in the video below:


Featured Image: chatiyanon/Shutterstock

Google Warns Of Duplicate Content “Black Holes” Caused By Error Pages via @sejournal, @MattGSouthern

Google’s “Search Off the Record” podcast recently highlighted an SEO issue that can make web pages disappear from search results.

In the latest episode, Google Search team member Allan Scott discussed “marauding black holes” formed by grouping similar-looking error pages.

Google’s system can accidentally cluster error pages that look alike, causing regular pages to get included in these groups.

This means Google may not crawl these pages again, which can lead to them being de-indexed, even after fixing the errors.

The podcast explained how this happens, its effects on search traffic, and how website owners can keep their pages from getting lost.

How Google Handles Duplicate Content

To understand content black holes, you must first know how Google handles duplicate content.

Scott explains this happens in two steps:

  1. Clustering: Google groups pages that have the same or very similar content.
  2. Canonicalization: Google then chooses the best URL from each group.

After clustering, Google stops re-crawling these pages. This saves resources and avoids unnecessary indexing of duplicate content.

How Error Pages Create Black Holes

The black hole problem happens when error pages group together because they have similar content, such as generic “Page Not Found” messages. Regular pages with occasional errors or temporary outages can get stuck in these error clusters.

The duplication system prevents the re-crawling of pages within a cluster. This makes it hard for mistakenly grouped pages to escape the “black hole,” even after fixing the initial errors. As a result, these pages can get de-indexed, leading to a loss of organic search traffic.

Scott explained:

“Only the things that are very towards the top of the cluster are likely to get back out. Where this really worries me is sites with transient errors… If those fail to fetch, they might break your render, in which case we’ll look at your page, and we’ll think it’s broken.”

How To Avoid Black Holes

To avoid problems with duplicate content black holes, Scott shared the following advice:

  1. Use the Right HTTP Status Codes: For error pages, use proper status codes (like 404, 403, and 503) instead of a 200 OK status. Only pages marked as 200 OK may be grouped together.
  2. Create Unique Content for Custom Error Pages: If you have custom error pages that use a 200 OK status (common in single-page apps), make sure these pages contain specific content to prevent grouping. For example, include the error code and name in the text.
  3. Caution with Noindex Tags: Do not use noindex tags on error pages unless you want them permanently removed from search results. This tag strongly indicates that you want the pages removed, more so than using error status codes.

Following these tips can help ensure regular pages aren’t accidentally mixed with error pages, keeping them in Google’s index.

Regularly checking your site’s crawl coverage and indexation can help catch duplication issues early.

In Summary

Google’s “Search Off the Record” podcast highlighted a potential SEO issue where error pages can be seen as duplicate content. This can cause regular pages to be grouped with errors and removed from Google’s index, even if the errors are fixed.

To prevent duplicate content issues, website owners should:

  1. Use the correct HTTP status codes for error pages.
  2. Ensure custom error pages have unique content.
  3. Monitor their site’s crawl coverage and indexation.

Following technical SEO best practices is essential for maintaining strong search performance, as emphasized by Google’s Search team.

Hear the full discussion in the video below:


Featured Image: Nazarii_Neshcherenskyi/Shutterstock