SEO

Liveblogging Pubcon: Optimizing Personalized Search.

Welcome back to Pubcon!  Lunch is (almost) over and everybody’s heading back to the sessions.  First up is the state of personalized search with Brian Combs, Greg Boser, and Christine Churchill. It’s scheduled for 1:30 but we’re starting late due to lunch stragglers.

Brian Combs is up first.  6 months ago he said we were at a crossroad, but it’s more like a maze.  Things are changing.

In fact, Brian doesn’t even include Google screenshots in his powerpoints anymore because they’ll probably change before he presents the deck.

Some things however, stay the same.  There’s still odd search results in local. Here’s a good example.

Review spam is still prevalent. I know this all too well. Google places support is still weak.

Is Google implementing Bugs or Features? They’re updating data on their own, grabbing images from all over, and allowing malicious reports to close down Google place pages. If Google is relying on 3rd party data, what happens when those 3rd parties stop updating?

Use this to your advantage. Find where Google is pulling that info from and clean it up.

He’s giving examples of Google places pulling images of breasts instead of a doctor’s office. you should monitor that, and figure out why it’s pulling those instead of your preferred images. I’m blogging that bit simply so I can get the word breasts onto search engine journal. Sorry Melissa.

Local SEO is still the wild west of SEO. Brian says it reminds him of SEO several years ago. You need to vette your vendor. Are they using fake reviews? Do they claim a special relationship with Google? Do they speak in half truths? If so, be wary.

Another great tip when selecting a local SEO vendor: ask them who owns the Google account. Make sure you own it and have access. Not owning that Google account could cause problems when you part ways in the future.

Interesting comscore numbers. Google search is increasing while vertical search numbers are declining. Brian theorizes that those searches are moving over to Google. (sorry Yelp!)

There’s a great heat map up now showing that for certain terms (in this case Pizza) that local listings are getting significant attention.

Closing thoughts for Brian: Google is creating a great deal of resentment – and not all of us can run to congress or the New York Times when we have issues.

Next up is Greg Boser ( @gregboser ) and the room is shocked that he not only has a powerpoint deck, but he also didn’t steal it! Way to go Greg!

So what is universal results?

  • Video/images
  • QDF content (news,blogs)
  • Google products (the ultimate evil)
  • Places localizations
  • Organic Localization

Be careful what you wish for. Generating universal results for some queries can kill your ROI. Do you really want people to click on your Youtube video instead of your website where they can actually buy something?

Figure out what types of queries are a good match and what types aren’t. (Hint: informational queries are good ones for universal results)

What’s the overall universal distribution for your keyword space? Knowing what types of phrases generate universal results can help you identify road blocks and opportunities.

What is your main query type? (see the leaked Google review document – or email Greg for a copy)

  • named entities
  • action queries (do)
  • Information queries (know)
  • navigation queries (go)

Google treats these types of queries differently – however several queries can have multiple intent. In that case, Google divides up the real estate for the various intents.

Do your phrases have geo identifiers? (Does it have a city name?) If they do, does the word order impact universal display? Try it with city name first or city name last and see how the SERPs differ.

Greg suggests changing the location on the side of Google and see how suggested search changes. They can vary greatly from Geo to Geo. I see a lot of that in my own work on Ford as well – there’s a heavy bias toward the brand in the city where they’re located.

And that’s it. It’s time for the Q&A.

We’re down a speaker for this session (Not sure what happened) so it’s Q&A time. I’ll try to cover the good questions – if I don’t get too sidetracked.

The first question is about how Greg finds the time to do all that research he just suggested. He advocates using interns which is a great idea that I’m going to steal. So to my intern, if you’re reading this, you’ve got a new job coming up!

Interesting note from somebody in the audience: Google may infer local intent and show different results, even if they don’t show Google places results. The room agrees that sometimes we all want to slap a Google engineer. In fact, I think a couple Google engineers just snuck out of the room after that comment.

Instead of slapping an engineer though, one thing you can do is look at your analytics and see what cities people are coming from and then set your location to see what they see.

Back in the day exact-match domains used to be effective (my opinion they still are) – city domains are the new exact match.

Another good point: You can’t predict where your video/image will show up in the results.

And that’s a wrap for this session. There doesn’t appear to be any more questions so I’m going to try and grab a seat with a power cord for the next session on enterprise local search with my friend @AlanBleiweiss

2bf27e1a5632db8aba77510c78aaa9a2 64 Liveblogging Pubcon: Optimizing Personalized Search.
By day Ryan Jones works at SapientNitro where he's a manager of search strategy & analytics and works mostly on fortune 500 clients. By night he’s either playing hockey or attempting to take over the world – which he would have already succeeded in doing had it not been for those meddling kids and their dog. Follow Ryan on Twitter at: @RyanJones or visit his personal website:
2bf27e1a5632db8aba77510c78aaa9a2 64 Liveblogging Pubcon: Optimizing Personalized Search.

Latest posts by Ryan Jones (see all)

You Might Also Like