1/ Here's a thread with my thoughts, comments and observations from my interview with @methode at last night's @BaySearch event at @stitchfix. 
2/ This was actually my first interview and Gary made it very easy. We had a great crowd (perhaps 120) and one person commented that the events are beginning to feel like reunions. That's definitely the type of feeling we'd like to foster.
3/ In terms of content, I think we covered a lot. I'm not sure there were any 'aha' type of moments but if you listened closely I think there were a lot of interesting mentions to mull over.
4/ The most interesting concept to me was the idea that organic search rankings are essentially an auction model, similar to paid search. Multiple signals combine to provide each candidate a bid for a position on a SERP.
5/ What I'd hoped to connect that with was the idea that each position might not be available to you in that auction. If one or more of those slots is for transactional content your bid is for a smaller subset of those spots.
6/ This is not documented anywhere but there are plenty of folks who have highlighted different intent/content in a SERP and there are other documents out there that indicate that something like this is likely happening.
7/ It was also nice to get confirmation of just how good Google has gotten at understanding language. Stories and examples of relevant content with thin or no backlink profiles winning were shared.
8/ That said, the details of how Google understands language was essentially off-limits. I wanted to talk vectors! But Gary didn't. Not because he doesn't know or want to but that his experience is people don't get it. They go glassy-eyed and slack-jawed.
9/ That might be true. But I tend to think someone needs to make it easier for laypeople to understand. Why? Telling people that your content was simply less or more relevant after an update is confusing if you don't understand the context.
10/ Adjacent to language understanding was further confirmation of Google's use of entities as a way to better understand documents. I didn't press Gary on this and I wish I had.
11/ The talk about parameters was super interesting but also frustrating. What I heard was that Google has learned how to deal with parameter bloat. They can see the patterns. That's great but it makes it harder for me to convince clients not to use them.
12/ This goes back to the idea of what is acceptable versus what is optimal. My job is to optimize. So having tons of parameters and robots'ing them out seems inefficient. And it still doesn't solve the internal link equity issue.
13/ Gary put the kibosh on using fragments (#) instead. Not because it wouldn't work but because it was a non-standard use of that feature and Google could change how they interpret those in the future and you'd be screwed.
14/ I've seen a number of sites use this technique and do quite well and it simplifies a number of issues. Maybe @g33konaut can speak to that more since Gary is not particularly enthused about JavaScript.
15/ Time got the best of me (even with an hour and fifteen minutes of interview) so we didn't get to talk as much about crawl issues as I would have liked. How crawl rates are determined and how to trigger a reassessment were high on my list.
16/ But not getting to those things simply means asking @methode back for a future event. So I again thank Gary for his time and expertise. I hope those in attendance enjoyed the event and took something from it that they can use over the course of the next year.
