Welcome to the first Surfaced newsletter, outlining important insights relevant to search.
What to expect in the 5-minute read:
Signal this week: Important market shifts or news in search
Surfacing social: What’s top of mind in the community
Research & insights: Recent data trends worth looking at
The move of the week: Something to act on, today
Office hours: Answering common questions I get or see
Excited to share these insights with you. Let’s get into it.
Google I/O happened on May 19th, and SEO’s have been arguing about it on the internet for the last 2 weeks. I wrote about it a bit more on LinkedIn, but the condensed version is here:
The winners are consumers. So many new features, including agents, a more intuitive “intelligent” search box, and commerce options.
Zero-click will continue to consume top-of-funnel information queries as they integrate AI Mode into more queries and serve more individualistic SERPs.
“Ask YouTube” is coming this summer, leveraging its powerful video library to provide users with the ability to use search to scan transcripts and answer questions with video. Hugely important channel that a lot are ignoring.
The shift that’s happening in platforms like Google is a tell for the future: Plan like “organic search” could go away/look much different in channel reports in the next few years. Act accordingly.
Also, need validation that AI search matters?
ChatGPT changed its hyperlink formula, which is now showing more referral traffic coming from the platform. If you’ve seen ChatGPT historically coming at fractions of a percentage of share, look again recently. I can’t help but feel they are doing this to win over marketers and measurement.
The measurement crisis is the #1 pain point people are asking in social. Practitioners can't show AI search ROI on legacy dashboards & leadership is asking for the same metrics they always have.
The basics are: traffic down, CTR down, but brand awareness up - with no dashboard that shows all three.
Reddit is asking, "What do I track?" while white-knight product warriors swoop in and recommend buying something. This shows nascent adoption in subreddits, lagging behind LinkedIn.
LinkedIn is asking, "How do I defend my budget?"
Well, Semrush says AI search converts 4.4x better than organic & I’ve seen this at lots of companies in the last couple of years — but almost nobody can show that to leadership yet. Beyond referral, visibility is so much more important than attribution. This is the pain point we’re all feeling now.
Also, at Google’s I/O conference, there was a slip of LLM.txt being shown on slides, and generally they have said it doesn’t matter. As of the last week in May, they added llms.txt to their “agentic browsing audits,” which is now a Lighthouse feature.
McKinsey says $750B will run through AI-powered search by 2028. Cool. The issue you and I will have is attributing revenue to this number, and measurement will become increasingly difficult. Currently, last-click referral is the only way to do this, and with the agentic web coming, this will be all the more important to see how agents are being reported.
G2 released a study that says that 51% of B2B buyers use AI chatbots to make software purchasing decisions. While this is self-serving, I’ve seen this in the wild. I’ve researched this first-hand and worked with leaders who do due diligence. The steps people are taking are:
Social proof: asking people in their networks, Slack groups, colleagues, and hard-to-attribute sources
Discovery on search or comparison sites they are familiar with, or conversational search
The actual hand-raising moment happens & you probably see it in your sales and web dashboards
Seer Interactive has reported an 85% rebound in organic CTR on AIO-present queries from December to February among their 53 brands investigated across 2 billion impressions. This is a win in recent months, but time will tell if it’ll hold or continue to slow or be dragged down further. One thing is for sure, as zero click is our new world, demand is still there, so people clicking means business.
The move this week is coming from data provided by Cyrus Sherpard/Zyppy, coming from an excellent correlation analysis he performed:
Here are 2 things you can do for the top factor, URL accessibility, which is used for training and grounding (very important to be higher in consideration set):
Make sure you are interlinking to your most important pages. Think of your site like a champagne tower, your homepage being the top glass (and very heavily weighted in AI search), with your most important pages a step down, like pricing, case studies, and other bottom-of-funnel content. Glasses at the top get the most juice.
Check your log files to see how often trained vs live robots are visiting different areas of your site. The way that the internet is understood is through user agentic crawling, rendering, or parsing. Seeing how many “agentic robot window shoppers” there are. If user agents are not visiting your most important pages very often, you might have an eligibility problem.
“Our brand is being mentioned negatively on Reddit, and it's showing up in AI answers…what do we do?”
Practitioners are realizing that Reddit threads they can't edit are being pulled into LLM responses verbatim. Most answers are either "hire a PR firm" or "nothing you can do," both of which are wrong.
You can't delete Reddit threads, but you can out-publish them.
Upvote positive sentiment threads and bring actual value to users
Create more community-native content in adjacent threads that shifts the citation pool (this works, I’ve done it).
This is no different than reputation management on any other platform, but my take is that managing individual threads on Reddit might not be the best use of time. Pick off the loudest and most problematic, and move on.
The real answer, though? Fix what they are complaining about and improve your product or service; it’s free primary customer research.
POLL TIME! What is your biggest challenge?
Welp, the first one is a long one. There is a lot to keep up with that’s happening every week, and even in this issue there was more I wanted to stuff in - but I do want to keep it curated for now. I also wrote a piece on enterprise challenges that brands have, if that interests you.
If something in here sparked a question or there’s something that moved you, let me know. Happy for the feedback to improve future sends.
Appreciate you all. Until the next time,
Andrew