Microsoft is making publicly available a new technology called GraphRAG, which enables chatbots and answer engines to connect the dots across an entire dataset, outperforming standard Retrieval-Augmented Generation (RAG) by large margins.

What’s The Difference Between RAG And GraphRAG?

RAG (Retrieval-Augmented Generation) is a technology that enables an LLM to reach into a database like a search index and use that as a basis for answering a question. It can be used to bridge a large language model and a conventional search engine index.

The benefit of RAG is that it can use authoritative and trustworthy data in order to answer questions. RAG also enables generative AI chatbots to use up to date information to answer questions about topics that the LLM wasn’t trained on. This is an approach that’s used by AI search engines like Perplexity.

The upside of RAG is related to its use of embeddings. Embeddings is a way of representing the semantic relationships between words, sentences, and documents. This representation enables the retrieval part of RAG to match a search query to text in a database (like a search index).

But the downside of using embeddings is that it limits the RAG to matching text at a granular level (as opposed to a global reach across the data).

Microsoft explains:

“Since naive RAG only considers the top-k most similar chunks of input text, it fails. Even worse, it will match the question against chunks of text that are superficially similar to that question, resulting in misleading answers.”

The innovation of GraphRAG is that it enables an LLM to answer questions based on the overall dataset.

What GraphRAG does is it creates a knowledge graph out of the indexed documents, also known as unstructured data. The obvious example of unstructured data are web pages. So when GraphRAG creates a knowledge graph, it’s creating a “structured” representation of the relationships between various “entities” (like people, places, concepts, and things) which is then more easily understood by machines.

GraphRAG creates what Microsoft calls “communities” of general themes (high level) and more granular topics (low level). An LLM then creates a summarization of each of these communities, a “hierarchical summary of the data” that is then used to answer questions. This is the breakthrough because it enables a chatbot to answer questions based more on knowledge (the summarizations) than depending on embeddings.

This is how Microsoft explains it:

“Using an LLM to summarize each of these communities creates a hierarchical summary of the data, providing an overview of a dataset without needing to know which questions to ask in advance. Each community serves as the basis of a community summary that describes its entities and their relationships.

…Community summaries help answer such global questions because the graph index of entity and relationship descriptions has already considered all input texts in its construction. Therefore, we can use a map-reduce approach for question answering that retains all relevant content from the global data context…”

Examples Of RAG Versus GraphRAG

The original GraphRAG research paper illustrated the superiority of the GraphRAG approach in being able to answer questions for which there is no exact match data in the indexed documents. The example uses a limited dataset of Russian and Ukrainian news from the month of June 2023 (translated to English).

Simple Text Matching Question

The first question that was used an example was “What is Novorossiya?” and both RAG and GraphRAG answered the question, with GraphRAG offering a more detailed response.

The short answer by the way is that “Novorossiya” translates to New Russia and is a reference to Ukrainian lands that were conquered by Russia in the 18th century.

The second example question required that the machine make connections between concepts within the indexed documents, what Microsoft calls a “query-focused summarization (QFS) task” which is different than a simple text-based retrieval task. It requires what Microsoft calls, “connecting the dots.”

The question asked of the RAG and GraphRAG systems:

“What has Novorossiya done?”

This is the RAG answer:

“The text does not provide specific information on what Novorossiya has done.”

GraphRAG answered the question of “What has Novorossiya done?” with a two paragraph answer that details the results of the Novorossiya political movement.

Here’s a short excerpt from the two paragraph answer:

“Novorossiya, a political movement in Ukraine, has been involved in a series of destructive activities, particularly targeting various entities in Ukraine [Entities (6494, 912)]. The movement has been linked to plans to destroy properties of several Ukrainian entities, including Rosen, the Odessa Canning Factory, the Odessa Regional Radio Television Transmission Center, and the National Television Company of Ukraine [Relationships (15207, 15208, 15209, 15210)]…

…The Office of the General Prosecutor in Ukraine has reported on the creation of Novorossiya, indicating the government’s awareness and potential concern over the activities of this movement…”

The above is just some of the answer which was extracted from the limited one-month dataset, which illustrates how GraphRAG is able to connect the dots across all of the documents.

GraphRAG Now Publicly Available

Microsoft announced that GraphRAG is publicly available for use by anybody.

“Today, we’re pleased to announce that GraphRAG is now available on GitHub, offering more structured information retrieval and comprehensive response generation than naive RAG approaches. The GraphRAG code repository is complemented by a solution accelerator, providing an easy-to-use API experience hosted on Azure that can be deployed code-free in a few clicks.”

Microsoft released GraphRAG in order to make the solutions based on it more publicly accessible and to encourage feedback for improvements.

Read the announcement:

GraphRAG: New tool for complex data discovery now on GitHub

Featured Image by Shutterstock/Deemerwha studio

WordPress announced over the weekend that they were pausing plugin updates and initiating a force reset on plugin author passwords in order to prevent additional website compromises due to the ongoing Supply Chain Attack on WordPress plugins.

Supply Chain Attack

Hackers have been attacking plugins directly at the source using password credentials exposed in previous data breaches (unrelated to WordPress itself). The hackers are looking for compromised credentials used by plugin authors who use the same passwords across multiple websites (including passwords exposed in a previous data breach).

WordPress Takes Action To Block Attacks

Some plugins have been compromised by the WordPress community has rallied to clamp down on further plugin compromises by instituting a forced password reset and encouraging plugin authors to use 2 factor authentication.

WordPress also temporarily blocked all new plugin updates at the source unless they received team approval in order to make sure that a plugin is not being updated with malicious backdoors. By Monday WordPress updated their post to confirm that plugin releases are no longer paused.

The WordPress announcement on the forced password reset:

“We have begun to force reset passwords for all plugin authors, as well as other users whose information was found by security researchers in data breaches. This will affect some users’ ability to interact with WordPress.org or perform commits until their password is reset.

You will receive an email from the Plugin Directory when it is time for you to reset your password. There is no need to take action before you’re notified.”

A discussion in the comments section between a WordPress community member and the author of the announcement revealed that WordPress did not directly contact plugin authors who were identified as using “recycled” passwords because there was evidence that the list of users found in the data breach list whose credentials were in fact safe (false positives). WordPress also discovered that some accounts that were assumed to be safe were in fact compromised (false negatives). That is what led to to the current action of forcing password resets.

Francisco Torres of WordPress answered:

“You’re right that specifically reaching out to those individuals mentioning that their data has been found in data breaches will make them even more sensitive, but unfortunately as I’ve already mentioned that might be inaccurate for some users and there will be others that are missing. What we’ve done since the beginning of this issue is to individually notify those users that we’re certain have been compromised.”

Read the official WordPress announcement:

Password Reset Required for Plugin Authors

Featured Image by Shutterstock/Aleutie

It’s been a year of considerable disruptions in digital marketing so far.

Right now, the industry is dealing with the integration of generative AI and the impact this is going to have on user behaviour and how people search. Alongside the relentless updates that Google keeps throwing at us.

SEO is changing and the industry is trying to adapt whilst accepting the uncertainty.

But, it’s not all catastrophic, there is a lot of opportunity ahead for those that can evolve to embrace the new.

To help marketers and brands thrive amidst uncertainty, I’ve outlined trends to focus on, guided by strategic insights and Yogi Berra’s timeless wisdom,

“Predictions are hard, especially about the future.” – Yogi Berra

Digital marketers can no doubt relate to Yogi’s sentiment, acknowledging the challenge of what lies ahead.

These, then, are the top 10 digital marketing trends for 2024:

1. Strategy: “If You Don’t Know Where You Are Going, You Might Wind Up Someplace Else.”

Why is “strategy” this year’s top trend instead of the latest technology?

Well, as Yogi once observed, “If you don’t know where you are going, you might wind up someplace else.”

According to Spencer Stuart’s 2024 CMO Tenure Study, the average tenure of chief marketing officers (CMOs) at Fortune 500 companies in 2023 was 4.2 years.

The study also found the average tenure of CMOs at B2B companies was 4.5 years. It was 4.0 years for CMOs at B2C companies. And it was just 3.1 years at the consumer-heavy top 100 advertisers.

So, developing a digital marketing strategy that will generate above-average results within a couple of years is the key to keeping your job as the CMO of a big brand.

And if you don’t master the art and science of creating a digital marketing strategy that generates business outcomes, then you won’t land one of the CMO jobs that turn over each year.

In other words, learning to use the latest technology is necessary, but it won’t get digital marketing leaders and executives where they want to go.

2. Generative AI: “Predictions Are Hard, Especially About The Future.”

Yogi also said, “Predictions are hard, especially about the future.” So, it’s tempting to ask generative AI tools to predict their own future.

OpenAI’s ChatGPT 3.5 says:

“AI and machine learning have been transforming digital marketing for years, but 2024 is poised to see these technologies become even more integral. From chatbots and predictive analytics to personalized content and ad targeting, AI will enhance customer experiences and drive efficiency. Brands leveraging AI can expect to see improved ROI and deeper customer insights.”

Google’s Gemini (formerly Bard) says:

“It ain’t science fiction anymore. Artificial intelligence (AI) is taking center stage, from crafting personalized chatbots that convert like crazy to optimizing campaigns with laser precision. Don’t fear the robot takeover, embrace it!”

And Anthropic’s Claude 3 says:

“Artificial intelligence writing assistants like Claude have been making waves, and in 2024 we’ll see these tools become ubiquitous in content marketing. They’ll help scale content creation while maintaining quality.”

But AI can’t see the big picture for your organization. It can’t empathize with people. And it can’t be creative like you. So, AI needs you in the driver’s seat to make it work effectively.

3. SEO: “It Ain’t Over Till It’s Over.”

Some pundits think SEO is dead. But as Yogi declared, “It ain’t over till it’s over.”

That’s because SEO pros have the remarkable ability to adapt to constant change or new information. Often, this means adjusting to the latest Google algorithm updates. But this also includes rethinking strategies based on the recent Google API “leak.”

Now, Rand Fishkin and Mike King were the first to report on the leaked documents. Although Google has officially acknowledged that these internal documents are authentic, it has also cautioned against jumping to conclusions based on the leaked files alone.

What should savvy SEO pros do?

Well, I’ve known Fishkin for more than 20 years. And he has the experience, expertise, authoritativeness, and trustworthiness (E-E-A-T) you’ve heard about.

So, I’m going to follow Fishkin’s recommendations, including:

• Hire writers with established reputational authority that Google already associates with quality content.
• Supplement link-building with public relations to increase branded search demand. (I’ll say more on this below.)
• “Think about SEO as being more geographically specific than you think it is even for web search results.”
• Move beyond parsing Google’s public statements and embrace experimentation and testing to uncover what produces results.

4. Link Building: “Always Go To Other People’s Funerals; Otherwise, They Won’t Go To Yours.”

I spotted this trend a long time ago, and I spoke about it at SES London 2009 in a session titled, “Beyond Linkbait: Getting Authoritative Mentions Online.”

Back then, I said link bait tactics can be effective “if you focus on the underlying quality as well as ingenuity needed to get other websites to link to you.”

I also provided a couple of case studies that showed British SEO professionals how to “approach journalists, bloggers, and other authoritative sources to enhance your company’s online reputation, whether or not you get links.”

But getting authoritative mentions without links didn’t translate. People on the other side of the pond thought I was saying something unintentionally funny like, “Always go to other people’s funerals; otherwise, they won’t go to yours.”

Hopefully, Fishkin’s recommendation will enable a lot more SEO pros to finally understand the underlying wisdom of supplementing link building with public relations.

As he clearly explained at MozCon, “If you get a whole bunch of links in one day and nothing else, guess what? You manipulated the link graph. If you’re really a big brand, people should be talking about you.”

5. Paid Media: “It’s Déjà Vu All Over Again.”

Everyone knows that Google, Meta, and other paid media are adding AI to their advertising platforms faster than the speed of sound. So, this might be mistaken as background noise.

But I’ve spotted the signal in the noise. Today’s frenzy to provide AI solutions is remarkably like the frenzy to provide programmatic solutions a decade ago. As Yogi said, “It’s déjà vu all over again.”

This means that digital marketers – and their agencies – can quickly refresh their “programmatic” workflow and turn it into “AI” best practices.

For example, Google touted a five-step programmatic workflow five years ago.

It consisted of:

• Organize audience insights.
• Design compelling creative.
• Execute with integrated technology.
• Reach audiences across screens.
• Measure the impact.

Why is today’s process of buying and selling digital media in an automated fashion so similar? Because AI is just fulfilling the early promise of programmatic to engage with consumers in the moments that matter most.

But there’s one significant difference between then and now.

As you’ll read below, it’s the improved ability to integrate your advertising platforms with your analytics platform to measure the impact of campaigns on brand awareness and lead generation.

6. Analytics: “You Can Observe A Lot By Watching.”

Performance marketers integrated their advertising platforms with their analytics platform more than a decade ago to measure the impact of their campaigns on “conversions.”

But brand marketers rarely focused on their analytics data because “brand awareness” was something they measured when consumers initially saw their display ads or watched their video ads.

A funny thing happened after Google Analytics 4 rolled out last summer. A “Business objectives” collection replaced the “Life cycle” collection of reports and one business objective you can now track is “Raise brand awareness.”

For example, brand marketers can now use traffic acquisition, demographic details, user acquisition, as well as which pages and screens users visit to measure brand awareness in places that are less vulnerable to ad fraud.

Another business objective you can now track is “Generate leads.”

So, digital marketers can measure any user action that’s valuable to their organization, including:

• Scrolling to 90% or more of their blog post.
• Downloading a whitepaper.
• Subscribing to their newsletter.
• Playing at least 50% of a product video.
• Completing a tutorial.
• Submitting a registration form.

And as Yogi noted, “You can observe a lot by watching.”

7. Content Marketing: “When You Come To A Fork In The Road, Take It.”

In the summer of 2020, the Content Marketing Institute and MarketingProfs fielded their annual survey and found that “Content marketers are resilient. Most have met the challenges of the pandemic head-on.”

In response to the pandemic, B2B and B2C marketers:

• Increased time spent talking with customers.
• Revisited their customer/buyer personas.
• Re-examined the customer journey.
• Changed their targeting/messaging strategy.
• Changed their distribution strategy.
• Adjusted their editorial calendar.
• Put more resources toward social media/online communities.
• Changed their website.
• Changed their products/services.
• Adjusted their key performance indicators (KPIs).
• Changed their content marketing metrics (e.g., set up new analytics/dashboards).

In other words, many content marketers totally overhauled their process for creating a content marketing plan from stem to stern.

For some, 2020 was the year of quickly adapting their content marketing strategy. For others, it was the year to finally develop one.

According to BrightEdge, content marketers are now “preparing for a Searchquake,” a tectonic shift in the content marketing landscape triggered by Google’s Search Generative Experiences (SGE).

But content marketers now know exactly what to do. As Yogi directed, “When you come to a fork in the road, take it.”

8. Video Creation: “If You Can’t Imitate Him, Don’t Copy Him.”

I teach an online class at the New Media Academy in Dubai on “Influencer Marketing and AI.” This may seem like an odd combination of topics, but they’re related to another class I teach on “Engaging Audiences through Content.”

I tell my students that creating great content is hard. That’s why marketers start using influencers or AI to create video content that their audience will find valuable and engaging. Then, they learn that there’s more to learn.

For example, AI can create realistic and imaginative scenes from text instructions. But AI can’t be creative like humans. So, the heart of every great video is still innovative, surprising, human-led creativity.

I show them “OpenAI Sora’s first short film – ‘Air Head,’ created by shy kids,” a Toronto-based production company.

Then, I ask them to apply what they have learned by using Synthesia, Runway, or invideo AI to generate a short video for their capstone project.

Invariably, they report that AI video generators can create realistic and imaginative scenes from text instructions but aren’t creative like shy kids.

Or, as Yogi put it, “If you can’t imitate him, don’t copy him.”

9. Influencer Marketing: “Nobody Goes There Anymore. It’s Too Crowded.”

The Influencer Marketing Hub says, “Most marketers believe that finding and selecting the best, most relevant influencers to be the most difficult part of influencer marketing.”

That’s ironic because HypeAuditor offers an influencer discovery platform that enables marketers to search through a database of 137.5 million influencers on Instagram, YouTube, TikTok, X (formerly Twitter), and Twitch.

It also enables marketers to apply filters to discover the perfect partners for their brand.

This apparent contradiction reminds me of Yogi’s comment, “Nobody goes there anymore. It’s too crowded.”

But it also indicates that most marketers are looking at influencer identification through the wrong end of the telescope. What should they do instead?

Well, I show the students in my “Influencer Marketing and AI” class how to use SparkToro to get a free report on the audience that searches for “Dubai.”

 

SparkToro estimates that 446,000 to 654,000 people search for “Dubai” monthly. And it uncovers the websites they visit, the keywords they search for, and their gender demographics.

 

SparkToro also identifies the sources of influence for this audience, including high-affinity accounts and hidden gems, so marketers can invest in the right ones.

10. Social Media: “The Future Ain’t What It Used To Be.”

I’m a big believer in “the rule of three.”

So, I wasn’t startled when I received an email from Jennifer Radke inviting me to attend “an exciting webinar focused on a high-level look into using ChatGPT for social media!”

But I was shocked when Katie Delahaye Paine shared a link to new research by Asana’s Work Innovation Lab and Meltwater, which found that “only 28% of marketing professionals have received training on how to use AI tools effectively.”

I was also horrified when I read a column by Mark Ritson in MarketingWeek that argued, “AI’s strength is automating high-volume, short-term marketing activity, which means social media could become a cesspool of synthetic content.”

Hey, I was having lunch with Chris Shipley in 2004 when she coined the term “social media.” So, I remember when social media still had a promising future.

But, as Yogi once declared, “The future ain’t what it used to be.”

So, social media marketing has three options:

• They can get upskilled to use AI tools more effectively.
• They can get reskilled to identify the right influencers.
• They can update their resumes and look for new jobs.

Picking Digital Marketing Trends Is Like Playing Moneyball

Some skeptics may question this counter-intuitive lineup of the top 10 digital marketing trends for 2024. Some of my selections seem to throw out conventional wisdom.

I recently watched the movie Moneyball (2011) for a second time. I was reminded that the Oakland Athletics baseball team’s general manager, Billy Beane (Brad Pitt), and assistant general manager, Peter Brand (Jonah Hill), used sabermetrics to analyze players.

This produced an epiphany: Picking digital marketing trends is like playing Moneyball. If you want to win against competitors with bigger budgets, then you need to find strategic insights, critical data, tactical advice, and digital marketing trends that conventional wisdom has overlooked.

And where did I come up with the whimsical idea of matching each trend with one of Yogi’s memorable quotes? Was it inspiration or hallucination?

I recently watched the documentary It Ain’t Over (2022) for the first time. It’s about New York Yankee Hall of Fame catcher Yogi Berra. And it supported Yogi’s claim, “I really didn’t say everything I said.”

But sportswriters kept attributing these Yogi-isms to the catcher because these “distilled bits of wisdom … like good country songs … get to the truth in a hurry,” as Allan Barra, the author of a book on Yogi, has explained.

And that strategic insight produced this year’s update – by a human – as opposed to last year’s top 10 digital marketing trends by ChatGPT.

More resources:

---

Featured Image: SuPatMaN/Shutterstock