Vetted.ai is one of the many AI chatbots that are gaining traction on Google search.
It’s not much, maybe 10,000 hits per month, but Google found the site to be the top result for ~100 keywords.
Trend: increasing.
Organic traffic and top 3 keywords for vetted.ai (Image credit: Kevin Indig)
A click on one of the Vetted AI results opens a blank chatbot page that quickly populates with content about the search query.
AI Chatbot by Vetted (Image credit: Kevin Indig)
Vetted AI is a shopping assistant that targets long-tail queries like [zep vs draino], [vornado mvh vs vh200]or [ugg ansley vs dakota].
Create content programmatically for any product comparison you can imagine.
Screenshot of the search for [zep vs drano]Google, March 2024 (Image credit: Kevin Indig)
Now contrast this with another site that gained traction in the SERPs over the past few months: Reddit (see result #3 in the screenshot above).
Reddit is the opposite of AI chatbots: human, experiential, non-optimized content.
Grow your skills with Growth Memo’s weekly expert insights. Subscribe for free!
Sundar Pich-AI
There’s more to Google’s $60 million deal with Reddit than meets the eye.
Google is pushing more forum content into the SERPs to counter AI content.
Verification is the ultimate AI watermark.
While Reddit can’t prevent humans from using AI to create posts or comments, the chances are slim because of two things Google search doesn’t have: moderation and karma.
Yes, Content goblins Reddit has already been targeted, but most of the 73 million daily active users provide helpful responses.
Content moderators punish spam with bans or even kicks.
But the most powerful quality engine on Reddit is Karma, “a user’s reputation score that reflects their community’s contributions.”
By simply upvoting or downvoting, users can gain authority and trustworthiness – two integral ingredients in Google’s quality systems.
The press release states that the deal allows Google to “facilitate more content views of Reddit information that will make our products more useful to our users and make it easier to participate in Reddit communities and conversations.” .
That might be true, but the real reason for the Google-Reddit deal is Karma.
Google already has access to Reddit content: it crawls reddit.com many times every day and can use that content to train its machine learning models. Chances are you already have.
Google also displays Reddit content prominently in search through high organic ranks (Hidden Gem update) and the SERP feature for discussions and forums.
Reddit.com has experienced the fastest growth in search of any domain and is now one of the largest on the web.
Reddit is on the same level of traffic as Amazon (Image credit: Kevin Indig)
But with access to Reddit’s API, Google can access Karma to train models on the value of human content and potentially show better Reddit answers in search. For example, filtering out posts that are spammy or that violate guidelines.
The implications go far beyond model training and go straight to the heart of what search users want.
Authentic reviews
Reddit Karma is especially valuable when it comes to authentic research experiences.
Google still has miles to go when it comes to product reviews.
A recent HouseFresh article shows how big brands aren’t testing as thoroughly as they might like.
As a team that has spent the last few years testing and reviewing air purifiers, it’s disheartening to see our independent site being overtaken by reputable publications that haven’t even bothered to check if a company is out of business before telling millions of readers to buy your products.
You can also remember a research study from Germany which found that “having a separate low-quality review section to support a site’s main content is a successful and lucrative business model.” Publishers make money from reviews, often to survive.
A recent study by Schultheiß et al. investigates the compatibility between SEO and content quality in medical websites with a user study. The study finds an inverse relationship between a page’s level of optimization and its perceived experience, indicating that SEO can harm at least the subjective quality of the page.
Reddit, on the other hand, is not optimized and human. You could say it’s in-ai-efficient.
Users go to Reddit, the main source of in-ai-efficient reviews, when they want opinions from other users instead of optimized reviews from publishers or affiliates. For some products, people want reviews from other people.
There’s also a speed advantage to using Reddit’s AI.
Similar to Indexing APIGoogle gets all of Reddit’s content through the API instead of having to get the crawl budget for the massive domain.
Wikipedia has about 60 million pages.
Reddit had over 300 million posts in 2020 alone. The site is big and getting bigger.
First, we’re excited to announce a new cloud partnership that allows Reddit to integrate new AI-based capabilities through Vertex AI. Reddit intends to use Vertex AI to improve search and other capabilities on the Reddit platform.
It’s ironic that Reddit uses Vertex AI from Google (company Twins models) to improve its search capabilities as part of the deal, since it was Reddit’s poor search function that drove so many users to search for Reddit content on Google in the first place.
The Hidden Gems update, the Google Forums and Discussions SERP feature, and the eventual Google deal would never have happened without this demand for Reddit results.
Oh-efficient
Efficiencies are incremental AI improvements. Rather than making new things, the biggest added value of AI so far is making things faster and better.
I use ChatGPT a lot to create better spreadsheet formulas and macros.
I never learned RegEx well, and my JavaScript/Python skills only prevent me from embarrassing myself in front of developers.
With LLMs, I can solve these problems quickly and independently.
This week, I ranked almost 20,000 keywords in eight basic topics for a client and paid less than $20 in one hour. AI is NOS for no code.
Websites are leveraging Aificiencies in the same way: they’re not new, they’re better.
Reddit uses Vortex AI, but examples like Zillow (I cover more second state of AI report) have already pulled the trigger.
Amazon and eBay lower the barrier to entry by allowing merchants to take a photo of a product and automatically write the title and description, including product features, with AI.
Amazon Creates Helpful AI Review Summaries (Inspiration for SGE?)
Redfin allows visitors to customize interior design with AI.
AI improves products rather than creating new ones.
However, when it comes to product advice and reviews, we want non-optimized and efficient information. Only raw and authentic comments.
We don’t want a biased list of products based on which ones get the biggest affiliate cut.
Users have grown up and can tell that most reviews on Google have financial incentives, just like the arguments “I think this resource would be useful to your audience” and “Let’s connect and look for business synergies” are really about links from entry and bids closed.
User intent: amateur experience
One way to make reviews more authentic is to use opinions from fans.
If I were in the review business, I would interview non-professionals and submit their opinions to the reviews, but only for products with a lower need for expertise.
Google’s addition of “expertise” to the QA guidelines is a direct indication that not all queries require expertise on the part of professionals.
Some inquiries require little experience and expertise. Just like we can use SERP features to infer user intent(s) for a query, a top Reddit result tells us that searchers value the opinions and experiences of non-experts.
For example, when you’re looking for the best GPU to match your computer hardware, you probably want to hear from an expert. When you’re looking for card games for couples, chances are you want to hear from other couples.
Google wants to show more results that are not optimized for embeds and backlinks but for human signals like Karma.
Helpful content updates they are steps towards more authentic content. The Google-Reddit deal is the sequel.
Featured Image: Lyna™
[ad_2]
Source link
