Will Critchlow

GSC has AI visibility data! We have been going through a very painstaking and manual process of gathering and analysing that data tests where there was a strong AI visibility hypothesis. Punchline: we're detecting AI visibility impacts with statistical confidence for the first time.

Look at the purple line in the chart - growing both before and after the website change (dotted line) - it's only by running a controlled test where a control group is unchanged (producing the pink forecast line) that we can see that there is causality. The change we made helped!

GSC has AI visibility data! We have been going through a very painstaking and manual process of gathering and analysing that data tests where there was a strong AI visibility hypothesis. Punchline: we're detecting AI visibility impacts with statistical confidence for the first time.

Ever since we rolled out experiments that focused on AI *clicks* the top question we've been asked is "what about visibility?". Well...

I'm not exaggerating when I say that this conversation with @retailgeek.bsky.social had more moments of shock and surprise for me than any other interview I've recorded. Not least when you see my jaw drop as Jason tells the story of being stuck in Minneapolis with Jeff Bezos on 9/11.

Kelly Wortham very kindly invited me to speak to the Test & Learn Community (TLC) about AI search and discovery. Her crew are extremely experienced experimentation experts when it comes to other channels - and particularly around conversion rate optimisation, UX, and product testing.

I don't quite know how to capture how excellent this conversation was. I recorded with @retailgeek.bsky.social recently and was blown away by his stories about Bezos, Blockbuster (and how he still gets trolled by his team), Target, Best Buy and more.

I think you can tell from my shocked expression that I did NOT already know the story @retailgeek.bsky.social told me about him getting stuck with Jeff Bezos on 9/11 when planes were grounded while they were in a conference room negotiating for Amazon to take over running the Target website.

When you start doing public speaking, recording videos etc, it's pretty universal to hate the sound of your own voice.

I have worked hard to get over that, and the other day, while driving to a SearchPilot offsite, I decided to listen to this episode of my podcast, where I was chatting to @paddymoogan.bsky.social - and I thought to myself "this sounds like a real, grown-up podcast". shows.acast.com/business-cla...

I've been in the lab. I wanted to see if I could model the impact of combining randomness in LLM outputs with the branching chaos of conversational paths. I wondered how stable the recommended product and brand would be. The answer? Extremely unstable.

I have more questions than answers, but I'm excited to do more testing.

I've been in the lab. I wanted to see if I could model the impact of combining randomness in LLM outputs with the branching chaos of conversational paths. I wondered how stable the recommended product and brand would be. The answer? Extremely unstable.

You know who is the GOAT at introducing other people and their work? @zyppy.com - Cyrus Shepard.

Who else do you know who's actually worked as a Google quality rater to improve their game?

You know who is the GOAT at introducing other people and their work? @zyppy.com - Cyrus Shepard.

I have become very interested in the developing AI capabilities for forecasting. My favourite tool at the moment is @FUTURESEARCHAI which takes a simple forecasting question, builds a panel of researchers and emulates the best human forecasters' approaches to quantitative forecasting.

The other is obviously the growth of GEO: - Change in revenue attributed to combined organic search + AI discovery over the same time period (+20% central estimate)

I have become very interested in the developing AI capabilities for forecasting. My favourite tool at the moment is @FUTURESEARCHAI which takes a simple forecasting question, builds a panel of researchers and emulates the best human forecasters' approaches to quantitative forecasting.

I recently set it the task of forecasting: - Change in Google search *traffic* to large retailers over the next 3 years (-15% central estimate) - Change in total online revenue for the same retailers over the same time period (+31% central estimate)

I am pleased that @zyppy.com is continuing the work of doing ranking factor surveys, and I was pleased to be asked to contribute. I cross-referenced my answers against the published data and found something interesting that is important to know when you're reading the output.

In general, I DON'T think they are especially actionable, but they ARE interesting for understanding the consensus and where you differ.

I am pleased that @zyppy.com is continuing the work of doing ranking factor surveys, and I was pleased to be asked to contribute. I cross-referenced my answers against the published data and found something interesting that is important to know when you're reading the output.

Agree? Disagree? One of the experts who answered this differently to me? I'd love to discuss in the thread.

I am pleased that @zyppy.com is continuing the work of doing ranking factor surveys, and I was pleased to be asked to contribute. I cross-referenced my answers against the published data and found something interesting that is important to know when you're reading the output.

For the record, my view is that they are correlated, but that this is a spurious correlation based on a set of confounding variables.

I am pleased that @zyppy.com is continuing the work of doing ranking factor surveys, and I was pleased to be asked to contribute. I cross-referenced my answers against the published data and found something interesting that is important to know when you're reading the output.

One of the questions was about how much brand & reputation signals INFLUENCE Google rankings. "Influence" is the key word here - not predict, not correlate with, but influence - we are looking for causality.

We're trying something new for this webinar - the registration landing page has a LIVE poll on it. @kevin-indig.bsky.social and I placed a bet last year about what would happen in the search space by the end of 2030...