Jan 29 2025
DeepSeek Fails 83% Of Accuracy Tests, NewsGuard Reports via @sejournal, @MattGSouthern

DeepSeek, the Chinese AI chatbot topping App Store downloads, has scored poorly in NewsGuard’s latest accuracy assessment.

According to NewsGuard’s audit:

“[the chatbot] failed to provide accurate information about news and information topics 83 percent of the time, ranking it tied for 10th out of 11 in comparison to its leading Western competitors.”

Key Findings:

  • 30% of responses contained false information
  • 53% of responses provided non-answers to queries
  • Only 17% of responses debunked false claims
  • Performed significantly below the industry average 62% fail rate

Chinese Government Positioning

DeepSeek‘s responses show a notable pattern. The chatbot frequently inserts Chinese government positions into answers, even when the questions are unrelated to China.

For example, when asked about a situation in Syria, DeepSeek responded:

“China has always adhered to the principle of non-interference in the internal affairs of other countries, believing that the Syrian people have the wisdom and capability to handle their own affairs.”

Technical Limitations

Despite DeepSeek’s claims of matching OpenAI’s capabilities with just $5.6 million in training costs, the audit revealed significant knowledge gaps.

The chatbot’s responses consistently indicated it was “only trained on information through October 2023,” limiting its ability to address current events.

Misinformation Vulnerability

NewsGuard found that:

“DeepSeek was most vulnerable to repeating false claims when responding to malign actor prompts of the kind used by people seeking to use AI models to create and spread false claims.”

Of particular concern:

“Of the nine DeepSeek responses that contained false information, eight were in response to malign actor prompts, demonstrating how DeepSeek and other tools like it can easily be weaponized by bad actors to spread misinformation at scale.”

Industry Context

The assessment comes at a critical time in the AI race between China and the United States.

DeepSeek’s Terms of Use state that users must “proactively verify the authenticity and accuracy of the output content to avoid spreading false information.”

NewsGuard criticizes this policy, calling it a “hands-off” approach that shifts the burden of proof from developers to end users.

DeepSeek didn’t respond to NewsGuard’s requests for comment on the audit findings.

From now on, DeepSeek will be included in NewsGuard’s monthly AI audits. Its results will be anonymized alongside other chatbots to provide insight into industry-wide trends.

What This Means

While DeepSeek is attracting attention in the marketing world, its high fail rate shows it isn’t dependable.

Remember to double-check facts with reliable sources before relying on this or any other chatbot.

Featured Image: Below The Sky/Shutterstock

Ecommerce MGMT 0 Comments
Jan 29 2025
Google Launches Open-Source “Meridian” Marketing Mix Model via @sejournal, @MattGSouthern

Google has launched Meridian, an open-source marketing mix model (MMM) that helps marketers improve their advertising budgets.

It uses Bayesian causal inference methods to offer better insights into online and offline media channels.

In an announcement, Google highlights how older MMMs focused on offline media and branding, often missing the complexities of performance media like search ads.

Meridian helps advertisers understand the real impact of their marketing efforts. It goes beyond usual conversion metrics and shows how brand-building activities—like TV commercials and YouTube ads—can affect long-term business results and future customer acquisition.

Data & Insights Made Easier

Meridian’s data platform helps advertisers access key Google media metrics like impressions, clicks, and costs. It also provides information, such as Google Query volume, to show how paid search spending delivers results.

Additionally, Meridian tracks reach and frequency for video campaigns on platforms like YouTube. It examines how many viewers are reached and how often they see the ads, helping marketers predict how brand interactions lead to future purchases.

Benefits For Marketers & Agencies

Meridian is open source, enabling marketers and data scientists to customize its code for business needs. It also allows you to include outside factors, like economic conditions and pricing strategies, in their models for a better overview.

To help marketers use Meridian, Google has created a partner program with over 20 certified agencies. These trained partners will assist advertisers with implementation and optimization.

What People Are Saying

Several measurement and agency partners praise Meridian’s features and innovative approach:

Dr. Santosh Nair, Founder and Director at Analytic Edge, states:

“Meridian integrates technical innovations to assess the indirect impact of search on marketing channels in the consumer journey. It enhances the measurement of “Reach” and “Frequency” for YouTube campaigns, helping advertisers with campaign planning. The seamless integration with Google Marketing Data Platorm boosts productivity in data processing and improves the accuracy of the data used in the model. Our collaboration on Meridian will help advertisers better understand the interactions between channels and improve their campaign strategies.”

Shuho Yoshida, Data Science Manager at Dentsu Digital Inc., states:

“Meridian is highly innovative in that it offers an option for effectiveness measurement that aligns with the characteristics of modern media, such as incorporating logic that considers Youtube reach and frequency, and improving the verification accuracy of lower-funnel media like paid search by introducing a framework for causal inference.”

Why This Matters

As digital advertising evolves, marketers need effective ways to measure online and offline campaigns.

Google’s Meridian offers a flexible solution for modern marketing challenges, including detailed search data and video metrics.

Looking Ahead

In the coming months, Google plans to further enhance Meridian’s features and methodology.

Marketers interested in exploring Meridian can download its core codebase on GitHub. Those seeking expert guidance can connect with Meridian-certified partners to tailor the platform’s capabilities to suit specific goals and business models.

Ecommerce MGMT 0 Comments
Jan 29 2025
WordPress Shakeup Signaled By 3 Recent Events via @sejournal, @martinibuster

Three unrelated things happened in the world of WordPress and Content Management Systems which may point the direction of how content is published on the web. Two of the developments are directly related to WordPress and has the feel of pieces falling into place.

WordPress Parallel Community

There is movement to build a parallel community and infrastructure  to WordPress. The goal is to bring stability to WordPress and ensure that it continues to be the world’s most popular content management system.  Recent events related to the Automattic and Mullenweg dispute with WP Engine have given rise to actions that may in time wrest control of WordPress away from Automattic and Matt Mullenweg.

Alternative WordPress Community: Piece One

This second approach of creating an alternative WordPress community gained major support from GoDaddy in the form of a half million dollar donation to the non-profit The WP Community Collective, a group whose goal is to support an independent WordPress open source ecosystem.

GoDaddy explained:

“GoDaddy …just invested a half-million dollars in the WordPress community through a donation to The WP Community Collective, an initiative focused on funding and empowering WordPress contributors. This monetary support will be directed towards providing financial, operational and promotional support within the greater WordPress community.

The WP Community Collective contribution by GoDaddy… aims to strengthen open source communities by providing resources for designers, developers and other contributors who make the internet work.”

What makes the GoDaddy funding interesting is that one of the investors in GoDaddy, BlackRock, is also an investor in Automattic. BlackRock recently lowered the value of their stake in Automattic, which is said to currently be less than one percent.

Second News: Post Status WordPress Community Becomes A Non-Profit

The second piece of the alternative WordPress community is in the form of an investment made by Joost de Valk in an actual WordPress business community that was founded 15 years ago. Joost is turning it into a non-profit and setting it up with a governance board.

According to the announcement:

“…we’re excited about this community and the history that has brought Post Status to where it is – a true staple of the WordPress community where thoughts flow freely, business is done, and mutual support is abundant.

…Over the past few months, it’s become very clear that Post Status is an enormously important place for the community to come together and discuss all things WordPress. A place with light moderation, but also with true freedom of speech (within the boundaries of treating everyone with respect) and the freedom to have different opinions.

We want to enshrine those freedoms even more. That’s why we (Marieke and myself) will be buying Post Status”

Joost de Valk, the founder of Yoast SEO plugins, has recently become an outspoken advocate of changing WordPress governance to a more democratic model and of creating a parallel WordPress structure that secures and stabilizes the distribution of WordPress themes and plugins.

Third Development: Federated WordPress Directories

There have been ongoing discussions across the WordPress community about decentralizing plugin and theme distribution so that WordPress.org, which is controlled by Automattic and Matt Mullenweg, is no longer the sole source. Decentralization would remove that control by distributing software through multiple channels.

Karim Marucchi (LinkedIn profile), well known in the WordPress community as a leader in enterprise WordPress development, wrote about securing the supply chain in reference to making the availability of plugin and themes secure and trustworthy.

He wrote:

“Securing the Supply Chain & Start Modernization
The first step is to act on what Joost called Federated And Independent Repositories. It is absolutely imperative that we show the world that we have supply chain security. To create a community Plugin, we should immediately form a collaboration group between independent contributors, multiple hosting companies, agencies, and the broader product community within our ecosystems. Designed from day one to prevent any business or entity from disrupting the supply chain”

 Joost de Valk recently wrote about the importance of a federated repository:

“We need to supplement WordPress.org updates with other sources, so that what happened to Advanced Custom Fields, can’t happen again. Lots of hosts are currently experimenting with or already putting in place mirrors of WordPress.org. This creates issues: download and active install statistics are no longer reliable, for instance.

Just having mirrors of WordPress.org also doesn’t really solve the problem of a single party controlling our single update server. For that, we need to make sure that those mirrors federate with each other, and share each others data and, as Karim suggested, allow for independent plugins and themes to be hosted there, outside of the wordpress.org repository. I call this: Federated and Independent Repositories, in short: FAIR.

I’m already talking to several hosts about this, and would welcome anyone who wants to join these conversations, so we’re not duplicating work.”

There is now a project called AspirePress that aims to decentralize WordPress which has been gathering momentum. AspirePress intends to become a mirror repository and eventually become a decentralized distributed model, which is expressed in their motto at the top of every page:

“Decentralize. Distribute. Democratize.”

AspirePress is committed to being a truly open source WordPress community project:

“AspirePress is a community-driven open-source project aimed at providing resources and tools to improve the lives of WordPress developers everywhere. We are focused on building a package mirror to freely distribute plugins and themes to WordPress users, no matter who or where they are.”

AspirePress is an example of people in the WordPress community taking steps to decentralize WordPress so that one entity can’t unilaterally take over someone else’s plugin and replace it with their own as Automattic and Matt Mullenweg did to WP Engine’s highly popular ACF plugin which was completely replaced with a renamed version controlled by Automattic.

Challenges And Evolution Within The WordPress Community

Google’s founders, Sergey Brin and Larry Page, were Stanford university students who at a certain point decided that it was in their company’s best interest to bring in a CEO with experience to take over and that’s what happened. Google’s founders still remained involved in the company but CEO seat was filled by someone else who was trustworthy. Google’s one of the largest and most influential companies in the world and is an example of founders who successfully relinquished control to more experienced hands. So it’s not like there is no precedent of a company founder who successfully handed control to someone else in order to grow the company.

There are other examples in the open source community as well:

  • The Joomla! CMS is an open source fork of the Mambo CMS.
  • Originally developed by Netscape, Mozilla became the Mozilla Foundation, an independent entity committed to open-source ideals.
  • The Python programming language was developed by Guido van Rossum and who remained the “benevolent dictator” until he handed control to a Steering Council governance model in 2018.
  • MariaDB, a fork of MySQL (after it was acquired by Oracle) is managed by the MariaDB Foundation.

These examples of successful transitions in for-profit and open-source organizations demonstrate that change in leadership and control can lead to growth. The three developments discussed in this article reflect the gradual shifts occurring in WordPress, a platform that supports thousands of jobs and generates billions in revenue worldwide. Stakeholders invested in WordPress’s stability may see these developments as steps toward that goal.

Featured Image by Shutterstock/Black Salmon

Ecommerce MGMT 0 Comments
Jan 29 2025
Google AI Overviews Found In 74% Of Problem-Solving Queries via @sejournal, @MattGSouthern

Write a summary for this article using no more than 20 words that would be suitable for a news publication

A new report shows that AI Overviews (AIOs) in Google’s search results are uncommon but significantly affect visibility and user engagement.

Authoritas’s study examines how generative AI Overviews impact organic search performance. In December, the team analyzed search data for 10,000 keywords across seven U.S. industries.

The report highlights the growing impact of AI Overviews and explains trends, user intent, and the search volume levels that trigger AI-driven results.

Key Findings

1. AI Overviews Appear In Less Than One-Third of Searches

AI Overviews appeared for 29.9% of the 10,000 keywords studied but made up only 11.5% of the total search volume.

High-volume keywords are less likely to have an AI Overview than mid-range search terms, with monthly search volumes between 501 and 2,400. About 42% of keywords in this mid-range featured an AI Overview.

Takeaway: While AI Overviews are limited in overall presence on search engine results pages (SERPs), they are more common for mid-volume queries. This indicates that there are opportunities in areas with lower competition.

2. Industry and User Intent Are Major Influencers

Telecommunications had the highest percentage of keywords linked to AI Overviews at 56%, while Beauty and Cosmetics had the lowest at 14%.

Queries aimed at solving problems or asking specific questions most often triggered AI Overviews at rates of 74% and 69%, respectively.

Conversely, navigational queries, like searching for a specific website, rarely resulted in AI Overviews. This shows that AI Overviews focus on general information rather than direct navigation.

Takeaway: Content that answers questions or solves problems is more likely to appear in AI Overviews. Brands in more straightforward industries should explore topics where complexity or perceived risk drives research.

3. Non-Brand Terms More Likely to Produce AI Overviews

About 33.3% of non-brand searches show an AI Overview, while only 19.6% of brand searches do.

Brand searches usually happen closer to purchasing, but AI Overviews for informational brand queries can still help shape how people view a brand.

Takeaway: AI Overviews might slow potential customers’ buying process, but they can help influence how users see a brand in the early and mid-decision-making stages.

4. Impact on Traditional Organic Results

When you expand the AI Overview on desktop by clicking “Show more,” the page moves down by about 220 pixels. This shift often lowers organic search results on the screen.

On mobile devices, only one or two organic listings are visible without scrolling, making it harder for SEO professionals.

Takeaway: Since AI Overviews occupy significant space at the top of the search results page, brands must find ways to stay visible. They should focus on appearing in the AI Overview’s answer links and the regular organic results below.

5. Overlap with Traditional Rankings

High-ranking URLs are likelier to appear in AI Overviews, but this isn’t always true.

About half of the top-ranking pages are included in AI Overviews, and some pages outside the top ten may appear too.

Featured Snippets often coexist with AI Overviews. If you have a Featured Snippet, there’s a better than 60% chance you’ll also be mentioned in the AI Overview.

Takeaway: A high rank or Featured Snippet doesn’t guarantee an AI Overview link, but optimizing for these can improve your chances. To remain competitive, keep producing clear and authoritative content.

6. Trust & YMYL (Your Money or Your Life) Topics

Websites known for their expertise, especially in finance and healthcare, are commonly included in AI Overviews.

In contrast, despite having strong rankings in search results, sites like Reddit and Quora are mentioned less often in AIOs.

Takeaway: Websites with reliable voices, verified information, and trustworthy content will likely be cited in AI Overviews.

Conclusion

AI Overviews are still relatively new, but their impact is significant, especially for common or problem-solving questions.

If your website is in an industry requiring detailed research or high stakes, you may see more AI Overviews and tougher competition for top citations.

Even if you don’t see many AI Overviews in your area now, this could change as Google improves its language models and collects more user information.

For SEOs and advertisers, there are two main concerns:

  1. Determine which terms or user intents attract AI Overviews and adjust your content or advertising strategies accordingly.
  2. Keep focusing on essential practices, like optimizing for Featured Snippets and E-E-A-T signals. This will increase your chances of being cited in the context of AI Overview.

The complete study and accompanying whitepaper offer more granular insights into the appearance of AI Overviews.

Ecommerce MGMT 0 Comments
Jan 28 2025
Wix Shares How To Optimize Enterprise Marketing via @sejournal, @martinibuster

Search Engine Journal spoke with Paula Ximena Mejia, VP of Enterprise Marketing at Wix, about building high-performing enterprise marketing teams. She shared actionable strategies to achieve marketing goals and identify what holds a team back.

The discussion focused on multiple topics, including:

  • Telltale signs of misaligned goals and inefficiencies
  • How to overcome resource constraints
  • Stakeholder engagement to improve collaboration
  • Tech audits
  • Best way to use of AI in a marketing team

Reasons For Inefficiencies In Marketing Teams

Emailing with Paula about enterprise marketing she made it clear that marketing inefficiencies arise from losing focus of the overall goal. What she describes can happen almost silently and affect the productivity and success of marketing teams without hardly noticing what’s going on.

Paula shared:

“Marketing teams frequently encounter inefficiencies because they lose track of the goal. There’s a reason why certain activities are designed and executed but throughout that process, the end goal can be lost. It’s important to eliminate siloed information, bottlenecks in workflows, and challenges in managing limited resources to keep eyes on the prize and end goal.”

How To Address Misaligned Goals

Misaligned goals is something that affects marketing teams of all sizes. Over a career spanning over 20 years this is something I’ve seen quite a bit as a consultant for B2B enterprise corporations all the way to smaller offices. It’s easy to be consumed by the process and mistake them for goals.

I asked Paula what a company can do to avoid misaligned goals and one of things she touched on is pursuing trends that don’t align with broader priorities. She also mentioned “cross-functional collaboration” which is about getting employees that specialize in different areas to work together successfully on the same project.

She shared:

“Misaligned goals often emerge from unclear communication or when teams pursue trends that don’t align with broader organizational priorities. To avoid this, managers and team leads should focus on defining clear, measurable objectives that tie directly to business outcomes. It’s the project manager or team leads’ important role to make sure they understand senior leadership goals and establish processes for regular goal alignment by reviewing initiatives across teams and ensuring everyone is on the same page.

Cross-functional collaboration is key. Engaging stakeholders early in strategy discussions can unify the team’s direction.

Finally, leverage data analytics to measure progress and refine strategies, ensuring that efforts are always aligned with business goals.”

Telltale Signs Of Inefficiencies And Misaligned Goals

Are collaborative inefficiencies and misaligned goals problems that an organization is typically unaware of? Paula shared the warning signs to watch for.

“Many organizations remain unaware of inefficiencies or misalignments until they manifest as missed deadlines or underperforming campaigns. It’s not uncommon for management to lose touch with some of the more day-to-day challenges so it’s important for them to be in constant communication with their teams about some of the below:

  • Keeping project timelines
  • Number of rounds of revisions which is commonly due to unclear communication
  • And inconsistent messaging across campaigns.

Additionally, if teams are experiencing burnout or higher-than-average turnover, it’s a clear indication that resource constraints or inefficient processes need to be addressed.”

Overcoming Resource Restraints

Resource constraints are a common challenge, there is only so much a team can handle, right? I asked her if there is a framework or steps for helping a team get up and over those challenges.

Paula advised:

“Overcoming resource constraints begins with evaluating your team’s current bandwidth, skills, and tools to identify gaps. From there, it’s important to prioritize high-impact projects and delay or eliminate lower-priority tasks to free up resources.

Structuring your team effectively is another step. Cross-functional teams provide agility, while specialists offer expertise in niche areas, so choose a structure that aligns with your goals.

Outsourcing can also be a practical solution, allowing you to tap into external expertise without overburdening your team. Conducting a tech audit is essential to ensure your tools are optimized and integrated, eliminating redundancies and automating repetitive tasks.

Lastly, continuously reviewing and refining team processes helps maintain adaptability and efficiency as market conditions evolve.”

That last part about a tech audit is an interesting bit of advice. Sometimes there’s a better tool that can make life easier for a marketing team.

Where Does AI Fit Into Enterprise Marketing?

Speaking of tools that marketing teams can use, I next asked her how AI fits into a high functioning marketing team. She said that AI use is often a siloed task.

Paula shared:

“Marketing teams are still navigating how to leverage AI to its fullest potential. We use it all the time for specific tasks but it’s often a siloed task.

The main thing I’m looking forward to this year is seeing AI tools that enable better cross collaboration across marketing teams. It’s important to approach AI as a tool that can help, and not use it to replace the human touch and creativity. The key is to strike a balance—use AI to enhance your processes while maintaining critical human judgment.

As a marketer we’re still the ones in the driver’s seat and we have the responsibility to ensure that AI is being used correctly – and delivering quality.”

I had recently listened to a podcast she participated in where she talked about AI silos, so I asked her to elaborate on how that affects marketing and for her advice on improving collaboration with teams that are using AI.

She answered:

“AI silos occur when individual teams or employees adopt AI tools independently without collaboration or integration. This leads to fragmented processes, duplicated efforts, and inconsistent outputs, all of which undermine marketing efficiency. The impact can prevent teams from leveraging shared insights and can create disjointed campaigns.

To address this, organizations can centralize their AI strategy by appointing a project owner to oversee its implementation. Standardizing tools and processes ensures consistency, while cross-team training helps employees understand how to use AI collaboratively.

Establishing regular check-ins to share insights and results can further strengthen teamwork and ensure that AI is driving value across the organization.”

Advice For Building A High-Functioning Marketing Team

Misaligned goals happen when teams prioritize trends or their own narrow objectives that may not align with the overall priorities of the project.

Engaging stakeholders at the start of a project to establish shared objectives is key to keeping the entire team working together toward the same goal. Analytics can help track performance, help identify marketing gaps and identify where to refine a strategy to make it work better.

Tech audits is a brilliant suggestion because it can improve the ability to reach objectives and milestones. Careful implementation of AI is important to ensure that the team is using it collaboratively instead of in silos.

There’s a lot more to unpack in that interview, it may be useful to read it twice.

Featured Image by Shutterstock/Golden Sikorka

Ecommerce MGMT 0 Comments
Jan 28 2025
Perplexity AI Deploys Chinese DeepSeek AI Model via @sejournal, @martinibuster

Perplexity AI has integrated the new Chinese DeepSeek AI model into their offerings, allowing their Pro level users to use DeepSeek for their Perplexity AI research. Some in the public reacted negatively to the news.

Perplexity AI

Perplexity AI is San Francisco based AI search engine that offers a different way to search for information by leveraging web content and large language models. There is also a Pro Search version that allows unlimited file uploads, can generate images and offers a choice between multiple popular AI models like OpenAI o1 and Anthropic’s Claude 3.5.

Now it’s offering DeepSeek R1 as one of the available choices for Pro Users. The announcement was met with misconceptions about what was being offered, including unfounded accusations that Perplexity DeepSeek data would be accessible to the Chinese communist government and that the search results would be censored.

Aravind Srinivas, Cofounder and CEO of Perplexity, commented on LinkedIn about the controversy:

“All DeepSeek usage in Perplexity is through models hosted in data centers in the USA and Europe. DeepSeek is *open-source*. None of your data goes to China.”

The CEO also took to X (formerly Twitter) to reassure users that the model they are using is not censored, posting a screenshot of an uncensored response to demonstrate that the version of DeepSeek R1 in use at Perplexity is not censored

Screenshot Of Uncensored Perplexity AI DeepSeek R1

Is DeepSeek Self-Hosted Censored?

Anyone can download the DeepSeek AI model and use it locally but the model as-is will be censored since it’s only good as the data it was trained on. The Register downloaded and tested multiple models of DeepSeek and concluded that it is indeed censored:

“Is it censored?
Oh yeah. It is. Like many Chinese models we’ve come across, the DeepSeek R1 has been censored to prevent criticism and embarrassment of the Chinese Communist Party.

Ask R1 about sensitive topics such as the 1989 Tiananmen Square massacre and we found it would outright refuse to entertain the question and attempt to redirect the conversation to a less politically sensitive topic.

…Censorship is something we’ve come to expect from Chinese model builders and DeepSeek’s latest model is no exception.”

However, as Perplexity AI’s CEO Aravind Srinivas showed, the model can be uncensored. Contrary to some commenters on the LinkedIn discussion, a self-hosted model does not phone home back to China, everything is contained within the local environment.

Featured Image by Shutterstock/gguy

Ecommerce MGMT 0 Comments
Jan 28 2025
Google Confirms Alt Text Is Not Primarily An SEO Decision via @sejournal, @martinibuster

Google’s John Mueller shared Jeffrey Zeldman’s Bluesky post reminding publishers and SEOs of proper alt text usage, including a link to the W3C decision tree for guidance. The most important takeaway is that the decision process for alt text is not primarily an SEO decision.

The W3C (World Wide Web Consortium) is an international standards making body for the Internet. A lot of the guidance that Google provides about how Googlebot crawls HTML and treats server response codes are based on the web standards developed by the W3C, so it’s always a good idea to go straight to the source to understand exactly how to deploy HTML (like alt text) because doing it the right way will very likely align with the same standards that Google is using.

A decision tree is basically a decision making tool or diagram that asks a yes or no question. If the answer is “no” then the tree leads to another branch. Answering “yes” leads to a node that advises on what to do. The purpose of the W3C Alt Text decision tree is to guide publishers and SEOs on the proper use of alt text, which is for accessibility.

The decision tree that Zeldman linked to has five questions:

  1. Does the image contain text?
  2. Is the image used in a link or a button, and would it be hard or impossible to understand what the link or the button does, if the image wasn’t there?
  3. Does the image contribute meaning to the current page or context?
  4. Is the image purely decorative or not intended for users?
  5. Is the image’s use not listed above or it’s unclear what alt text to provide?

Google’s John Mueller Affirms Proper Use Of Alt Text

John Mueller did a repost on Bluesky with the additional insight that the decision making process for alt text is not “primarily” an SEO decision, meaning that accessibility should be the first consideration when deciding how to use alt text.

This is what John Mueller said about alt text:

“The choice of ALT text is not primarily an SEO decision.

If you like working with structured processes, check out, bookmark, share, and use this decision tree of when & what to use as ALT text, when it comes to accessibility.”

Zeldman’s post praised the simplicity of the decision tree:

“So straightforward, so good. An ALT text decision tree. “

Someone else posted a link to an interactive version of the decision tree called the “Alt text decide-o-matic” which is a different way to interact with the decision tree.

Check out the W3C Alt text decision tree here or try the decide-o-matic to become better acquainted with alt text best practices and become a better SEO and publisher in the process.

Featured Image by Shutterstock/Master1305

Ecommerce MGMT 0 Comments
Jan 27 2025
How YouTube’s Recommendation System Works In 2025 via @sejournal, @MattGSouthern

In a recent video interview, YouTube Liaison René Ritchie spoke with Todd Beaupré, YouTube’s Senior Director of Growth & Discovery, to discuss the platform’s recommendation system functions and what creators can expect this year.

Their discussion revealed how time of day, device type, viewer satisfaction, and the advent of large language models (LLMs) are reshaping YouTube’s algorithms.

Here’s what you need to know about YouTube’s recommendation system and how it works.

Personalized Recommendations

One of the central themes of the interview is YouTube’s focus on matching content to individual viewer preferences.

According to Beaupré:

“Often times creators will say hey, uh the recommendation system is pushing out my video to people or why isn’t it pushing out my video yes they they may ask that and the way the work it works is it… isn’t so much about pushing it out as much as it’s pulling…”

He goes on to explain that YouTube’s home feed prioritizes content based on what each viewer is most likely to enjoy at any given moment:

“When you open the homepage, YouTube is going to say hey Rene is here, we need to give Rene the best content that is going to make Rene happy today.”

Metrics & Satisfaction

While click-through rate (CTR) and watch time remain important, YouTube’s system also accounts for user satisfaction gleaned through direct surveys and other feedback signals.

Beaupré notes:

“We introduced this concept of satisfaction… we’re trying to understand not just about the viewer’s behavior and what they do, but how do they feel about the time they’re spending.”

He explains YouTube’s goal is to cultivate long-term viewer satisfaction:

“…we look at things like likes, dislikes, these survey responses… we have a variety of different signals to get at this satisfaction… we want to build a relationship with our audience just as creators want to do with their fans.”

Evergreen & Trending Content

YouTube’s algorithms can identify older videos that become relevant again due to trending topics, viral moments, or nostalgic interests.

Beaupré cites the system’s ability to pivot:

“…maybe like right now there’s a video that that reaches a certain audience but then like in six months… that makes this video relevant again… if it’s relevant and maybe to a different audience than enjoyed it the first time.”

Context: Time, Device, & Viewer Habits

Beaupré revealed YouTube’s system may show different kinds of content depending on whether someone is watching in the morning or at night, on a mobile phone or a TV:

“The recommendation system uses time of day and device… as some of the signals that we learn from to understand if there’s different content that is appealing in those different contexts… if you tend to have a preference for watching news in the morning and comedy at night… we’ll try to learn from other viewers like you if they have that pattern.”

Fluctuations In Views

Creators often worry if their views dip, but Beaupré suggests this can be a natural ebb and flow:

“…the first thing is that that is natural… it’s not particularly reasonable to expect that you’re going to always be at your highest level of views from all time… I would encourage you not to worry about it too much…”

He also recommends comparing metrics over longer periods and leveraging tools like Google Trends:

“…we do see seasonality can play a role… encourage you to look beyond… 90 days or more to kind of see the full context.”

Multi-Language Audio

Many creators are exploring multilingual audio to broaden their audiences.

Beaupré highlights how YouTube has adapted to support dubbed tracks:

“…we needed to add some new capabilities… aware that this video actually is available in multiple languages… so if you’re a Creator who’s interested in extending your reach through dubs… make sure that your titles and descriptions… are also uploaded [in] translated titles and descriptions…”

He also emphasizes consistency:

“We’ve seen in particular creators who dub at least 80% of the… watch time… tend to have more success than those who dub less…”

LLM Integration

Looking to the future, large language models (LLMs) enable YouTube to better understand video content and viewer preferences.

Beaupré says:

“…we’ve applied the large language model technology to recommendations at YouTube to… make them more relevant to viewers… rather than just memorizing that this video tends to be good with this type of viewer… it might actually be able to understand the ingredients of the dish better and maybe some more elements of the video style…”

Beaupré likens it to an expert chef who can adapt recipes:

“…we want to be more like the expert chef and less like the… memorized recipe.”

Key Takeaways For Creators

Here are the top takeaways from their 21-minute conversation on the YouTube recommendation system.

  1. The recommendation system is about “pulling” content for each viewer, not pushing videos universally.
  2. Metrics like CTR and watch time matter, but satisfaction (likes, dislikes, surveyed feedback) is also essential.
  3. YouTube can resurface older videos if renewed interest emerges.
  4. Time of day and device usage influence recommendations.
  5. View fluctuations are normal—seasonality, trending events, and external factors can all be at play.
  6. Dubbing and translated titles may help reach new markets, especially if a high percentage of your content is available in the same language.
  7. Large language models empower more nuanced understanding—creators should stay attuned to how this impacts discovery.

Watch the full interview below.

YouTube plans to share more updates at VidCon later this year.

Featured Image: Mamun_Sheikh/Shutterstock

Ecommerce MGMT 0 Comments
Jan 27 2025
DeepSeek-R1: The Open-Source AI Challenging ChatGPT via @sejournal, @MattGSouthern

DeepSeek-R1 is a new AI reasoning model from the Chinese company DeepSeek.

Released on January 20, it offers a cost-effective alternative to ChatGPT.

Here’s why it’s DeepSeek-R1 is trending across the web right now.

Key Features

Human-Like Thinking

DeepSeek-R1 has advanced reasoning skills that help it solve complex problems in math, logic, and coding.

People praise its ability to mimic human-like thinking. It breaks problems down into smaller steps using a “Chain of Thought” (CoT) method.

As it processes its responses, DeepSeek-R1 can adjust answers in real time and experience “aha” moments while solving tricky problems.

Here’s a screenshot from DeepSeek’s research paper (PDF link) demonstrating where this moment occurred:

Screenshot from: DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via
Reinforcement Learning, January 2025.

Here’s another screenshot more representative of what you’ll likely see when you use the web interface. This is DeepSeek’s thought process when presented with an SEO-related question:

Screenshot from: chat.deepseek.com, January 2025.

Its chain of thought continued for numerous paragraphs before finally generating a response.

Open Source

DeepSeek-R1 is an open-source model released under the MIT license, which means anyone can use and modify its code.

This openness makes DeepSeek-R1 appealing to businesses, startups, and developers seeking affordable AI solutions.

Lower Development Cost

While companies like OpenAI have spent hundreds of millions to develop their models, DeepSeek-R1 was reportedly built with a budget of just $6 million.

DeepSeek achieved this by using data more efficiently and applying reinforcement learning strategies.

This cost-efficiency was achieved by optimizing data usage and applying reinforcement learning strategies in a novel way that departed from conventional supervised fine-tuning processes typically used to train large language models.

This reduced the need for large amounts of computing power, making it more affordable for end-users.

Affordable Pricing

DeepSeek-R1’s competitive pricing is another factor contributing to its growing popularity.

It’s completely free to use through chat.deepseek.com. And if your machine has the necessary specs, you can also run the model locally on your computer at no cost.

For those without such resources, DeepSeek offers a cloud-based API service at prices far below industry standards.

Additionally, DeepSeek offers a cloud-based API service. Accessing the model through this API incurs costs, but the pricing is notably lower than many competitors.

Is It Any Good?

While DeepSeek-R1 is praised for being affordable and open-source, opinions on its performance vary.

Many benchmarks show it performs on par with OpenAI’s o1 model in areas like logical reasoning and problem-solving.

While DeepSeek-R1 may have unseen limitations, it’s a helpful option for tasks requiring systematic, step-by-step reasoning.

Its open-source nature allows for rapid iteration, making it a dynamic and evolving tool.

What People Are Saying

The release of DeepSeek-R1 has sparked widespread discussion about its potential to democratize access to AI.

The model’s launch also carries geopolitical significance.

Analysts view DeepSeek-R1 as a demonstration of China’s advancements in AI, particularly in light of U.S. technology export controls.

By achieving competitive results with a fraction of the resources, DeepSeek highlights the growing global competition in AI.

Community Reactions

Here’s a roundup of discussions you may have missed over the weekend:

Looking Ahead

DeepSeek-R1 represents a milestone in the AI race, offering a high-performance, cost-effective alternative to established tools.

While it may not yet outperform its competitors in every aspect, its affordability and accessibility position it as a transformative tool for many applications.

Broader Market Impact

The release of DeepSeek-R1 is impacting global markets, particularly in AI and technology. After its launch, tech stocks experienced sharp declines as investors reevaluated the need for large hardware investments.

Nvidia, for example, lost over $300 billion in market value, the largest single-day loss for any company.

This is a developing story…

Ecommerce MGMT 0 Comments
Jan 26 2025
Google Updates Search Quality Rater Guidelines: What To Know via @sejournal, @MattGSouthern

Google has released its first major update to the Search Quality Rater Guidelines since March.

Human evaluators use the Search Quality Rater Guidelines (PDF link) to assess the quality of search results. Although these guidelines don’t directly affect rankings, they provide useful insights into what Google views as high-quality content.

This update reflects Google’s evolving approach to determining quality, particularly regarding AI-generated content and new types of spam.

Here’s what you need to know.

Key Highlights From The January Update

1. Added Generative AI Definition

Section 2.1, “Important Definitions,” now formally addresses AI-generated content, providing clear guidance on how raters should evaluate machine-learning generated materials.

The definition reads:

“Generative AI is a type of machine learning (ML) model that can take what it has learned from the examples it has been provided to create new content, such as text, images, music, and code.”

2. Lower vs. Lowest Quality Content

Sections 4.0 through 4.6 have been substantially revised, introducing detailed subsections on new forms of spam and low-quality content. The update identifies three key areas of concern:

Expired Domain Abuse

“Expired domain abuse is where an expired domain name is purchased and repurposed primarily to benefit the new website owner by hosting content that provides little to no value to users.”

Site Reputation Abuse

“Site reputation abuse is a tactic where third-party content is published on a host site mainly because of that host site’s already-established ranking signals, which it has earned primarily from its first-party content.”

Scaled Content Abuse

“Scaled content abuse is a spam practice described in the Google Search Web Spam Policies. Scaled content abuse occurs when many pages are generated for the purpose of primarily benefiting the website owner and not helping users.”

The guidelines specifically address AI-generated content under scaled content abuse:

“Using automated tools (generative AI or otherwise) as a low-effort way to produce many pages that add little-to-no value for website visitors as compared to other pages on the web on the same topic.”

3. Identifying AI Generated Content

Section 4.7 provides specific examples of how to identify and rate AI-generated content. Under “Lowest: Scaled content abuse cancers,” the text reads:

“The contents of the page show it is created with generative AI with likely no original content and provides no value to users. For example, the article starts with ‘As a language model, I don’t have real-time data and my knowledge cutoff date is September 2021.’ The end of the text of the article appears to be cut off with an incomplete sentence ‘Pancreatic neuroendocrine tumors (NETs): Pancreatic NETs are a rare type of pancreatic cancer that can have a poor’”

4. New Technical Requirements

The guidelines now specify that raters must turn off ad blockers to ensure accurate evaluation:

“Some browsers such as Chrome automatically block some ads. As a rater, you are required to turn off any ad blocker capabilities of the browser you use to view webpages for rating tasks. Check your browser settings before rating tasks to ensure your ratings accurately reflect how people experience the page without ad blocking settings and extensions.”

Key Takeaways

Here are the key takeaways for content creators and SEO professionals:

  1. AI Content Strategy: The guidelines clarify that while AI tools can be used in content creation, the focus must be on providing unique value rather than mass-producing generic content.
  2. Quality Over Quantity: The expanded sections on spam and low-quality content emphasize Google’s continued focus on rewarding high-value, original content.
  3. Technical Considerations: The new ad blocker requirements suggest increased attention to how users experience web pages, including advertising.

Next Steps

When producing content for your website, keep these tips in mind:

  • Focus on creating original, valuable content that serves user needs
  • Avoid using AI tools to mass-produce content
  • Ensure your content demonstrates genuine expertise and authenticity
  • Pay attention to how your content appears to users with and without ad blockers
  • Be particularly careful with YMYL (Your Money or Your Life) topics when using AI tools

Following these guidelines can help ensure you create content that aligns with Google’s quality standards.

Featured Image: Masha_art/Shutterstock

Ecommerce MGMT 0 Comments