Everything You Need to Know About the Huge Google Search Leak

[00:00:00] It is Tuesday May 28th.

[00:00:03] Today it's the biggest leak in Google's history and the documents in it detail thousands

[00:00:09] of previously secret factors that decide how high you rank in their search engine.

[00:00:14] What are those factors and have marketers been lied to all this time?

[00:00:20] We have extensive coverage today in this special episode.

[00:00:23] I'm Tod Maffin, that's ahead today in Digital Marketing.

[00:00:29] At first glance you might think the video is of a software engineer doing a job interview.

[00:00:34] He's got the software engineer uniform, pressed jeans, white polo shirt, neatly groomed facial

[00:00:39] hair.

[00:00:40] He's sitting down looking right into the camera.

[00:00:43] Hi there.

[00:00:44] Only this is no job interview.

[00:00:47] This is the guy.

[00:00:48] I'm the guy.

[00:00:49] This past weekend leaked Google's most important trade secrets.

[00:00:55] Because he's mad.

[00:00:57] For decades we've been lied to.

[00:01:01] Literally lied.

[00:01:04] His name is Efran Azimi and today he uploaded a video revealing that he was the person

[00:01:08] behind the massive weekend leak of internal Google search documents that sparked chaos

[00:01:14] in the marketing industry.

[00:01:16] He's mad because the documents, 2500 pages worth, seem to contradict what Google has

[00:01:21] told marketers for decades about how to rank higher in the search results.

[00:01:26] The leak is a set of confidential internal documents, API documentation to be precise,

[00:01:32] detailing Google's secret sauce, the parameters and levers that run their money printing

[00:01:38] search engine.

[00:01:39] We are devoting our full episode to this leak today, what's in it and what it

[00:01:43] means for you going forward.

[00:01:46] First though, a bit of background.

[00:01:48] For decades now, Google engineers and spokespeople have given marketers advice about how to

[00:01:52] rank better.

[00:01:54] That advice has been general.

[00:01:56] For the last several years it's basically boiled down to, don't worry about the algorithm,

[00:02:00] just make good content.

[00:02:03] But often that advice seems to contradict what practitioners are seeing in the search

[00:02:07] results.

[00:02:08] For instance, Google claims that click data isn't considered a significant ranking

[00:02:12] factor.

[00:02:13] That might be true, though many outsiders who've tried to test this says it doesn't

[00:02:17] seem to be.

[00:02:18] Until now we've just had to take Google's word for it.

[00:02:22] Essentially, is what Mr. Azimi says caused him to leak these documents.

[00:02:27] As Googleers came and directly said, especially from the search relations team, that certain

[00:02:38] signals such as Chrome data and click data are not used in the ranking.

[00:02:47] This is massive.

[00:02:49] Barry Schwartz has been covering the SEO industry for most of his career.

[00:02:52] He writes at seroundtable.com.

[00:02:54] I spoke with him this morning.

[00:02:56] It looks to be a legit release of Google's search factors of some sort in a real structured

[00:03:03] fashion so people can actually search it and so forth.

[00:03:05] It seems like a lot of the specific signals that Google may be using in some level

[00:03:10] for ranking has been fully released to the public, which is Google's secret sauce

[00:03:14] really.

[00:03:15] Mike King was one of two people who were first given the leak.

[00:03:18] He spent 18 years in the SEO industry.

[00:03:20] Today he runs the agency iPollrank.com.

[00:03:24] I spoke with him earlier today as well and asked him if this means the Google reps

[00:03:27] were just lying to marketers about what factors matter.

[00:03:30] There is this dynamic between them and the SEO community where because there's

[00:03:37] this kind of shield of information, they have some wiggle room to say things that

[00:03:45] misdirect us.

[00:03:46] I think something like saying that clicks don't impact rankings is something they

[00:03:53] have to say because they don't want everyone trying to manipulate clicks.

[00:03:57] That is their core signal that you're using to reinforce what should rank.

[00:04:03] If everyone goes out there and they just make a bunch of bots to click on

[00:04:08] things, then that signal is no longer valuable.

[00:04:11] So I don't think they've lied to us necessarily out of animosity or anything

[00:04:17] like that.

[00:04:17] I think they had to, but I think the problem is the way that they've done it.

[00:04:23] There is another possibility here too, that maybe the public facing reps that

[00:04:27] Google uses to talk with us about SEO, spokespeople, the evangelists, the

[00:04:31] liaisons, maybe they were inside the black box too and didn't even know it.

[00:04:38] Tom Kapper is senior search scientist at Moz.com.

[00:04:40] I wonder to what extent the people that are employed by Google to communicate

[00:04:45] with the SEO industry, what are to what extent they're in the dark here?

[00:04:49] Did they know that they were lying or has this stuff been a secret within

[00:04:55] Google?

[00:04:56] Lying might be a strong and judgmental word.

[00:04:59] Did they know that they weren't saying the whole truth, shall we say?

[00:05:02] Or were the firewalls within Google so substantial that actually they were

[00:05:07] being given the same information that SEOs were?

[00:05:11] I could believe something like that.

[00:05:13] The debate about was Google lying to marketers aside, the documents do

[00:05:17] reveal a lot more about the inner workings of the search engine than we

[00:05:21] knew before.

[00:05:22] Here are some of those highlights.

[00:05:24] First, there are about 14,000 ranking elements listed.

[00:05:29] The document did not detail the weighting of each element, just that

[00:05:33] they exist.

[00:05:34] There is a metric called site authority, which hadn't been previously

[00:05:38] disclosed and based on name alone seems to run counter to Google's claim

[00:05:42] that broad site trustworthiness was not a strong ranking metric.

[00:05:47] Third, content can be demoted in search ranking position if it's a

[00:05:51] product page with bad reviews.

[00:05:53] We were able to infer this from one of Google's recent updates, but

[00:05:56] nothing's been this clear.

[00:05:58] Fourth, Google apparently keeps a copy of every version of every page.

[00:06:03] It's ever indexed so it can go back and see what changes were made, though

[00:06:08] it only uses the last 20 content changes when analyzing links.

[00:06:13] Fifth, yes, links to your content still matter despite the direction

[00:06:17] away from them.

[00:06:18] Google spokespeople have tried to steer marketers.

[00:06:20] And six, there are penalties for domain names that exactly match

[00:06:24] unbranded search queries.

[00:06:26] For example, cheap fashion shoes.com or women's luxury watches.com.

[00:06:33] Use those domain names and your ranking will drop.

[00:06:37] It'll be hard for Google to claim this is an out of date document.

[00:06:40] It seems to have been updated as recently as March.

[00:06:44] The leaker says the documentation came from GitHub where software engineers

[00:06:47] store their code and documentation.

[00:06:50] These API documents had been protected and only accessible by

[00:06:53] specific Google employees, but between March and this month, those

[00:06:58] documents had accidentally been left open to the public.

[00:07:02] Think of it like a library, but as SEO veteran Rand Fishkin, who was the

[00:07:06] other person initially given this leak said in his blog post about this all

[00:07:09] quote, whereas libraries are public.

[00:07:12] Google search is one of the most secretive, closely guarded

[00:07:15] black boxes in the world.

[00:07:17] In the last quarter century, no leak of this magnitude or detail has ever

[00:07:22] been reported from Google's search division unquote.

[00:07:27] As for how much of the code is actively being used by Google's search

[00:07:30] engine, that's hard to know for sure.

[00:07:33] It's possible Google has since deprecated some of these factors.

[00:07:36] Some may have been internal only or were test API calls that

[00:07:40] were never implemented, but all that seems unlikely.

[00:07:43] Partially because the code clearly notes when features have been sunsetted.

[00:07:48] So what now given that some of the advice Google's given us in the

[00:07:51] past seems to have been misdirection at best should marketers

[00:07:55] change our SEO practices because of this?

[00:07:59] I don't see why.

[00:08:00] Again, Barry Schwartz.

[00:08:01] I don't see there's no reason to change your strategies.

[00:08:03] Ultimately you want to build content that people like, that people

[00:08:06] want to click on, that people want to consume, that people want to

[00:08:07] share, that people just want to go back to.

[00:08:10] So ultimately it's all in line with building out great content because

[00:08:14] that's the type of stuff that generates clicks, generates visits.

[00:08:17] People have long divorced SEO from user experience.

[00:08:22] This is Mike King again.

[00:08:24] I think the two need to work more closely together because ultimately

[00:08:27] you need someone to like come to your page and stay there.

[00:08:29] Like that is a signal that Google is using.

[00:08:33] Historically, we've all just been like, cool, let's just get them

[00:08:35] to the site, get them to the site and then, you know, leave

[00:08:37] the conversion to someone else.

[00:08:39] I think that we need to be thinking more about UX principles

[00:08:44] and how do we ensure that people get to the thing that they're looking

[00:08:47] for so that we can drive more successful sessions.

[00:08:51] Barry Schwartz is at seroundtable.com.

[00:08:54] Mike King is at ipolerank.com and Tom Kapper works at moz.com.

[00:09:00] For the record, we did reach out to Google for comment, but

[00:09:03] did not hear back by deadline.

[00:09:04] We have even more detail about these documents in today's email

[00:09:08] newsletter, including links to Barry's coverage, Rand Fishkin's

[00:09:11] full blog post, and Mike King's breakdown of what's in the docs.

[00:09:14] You can get all that by signing up to the newsletter for free by

[00:09:17] tapping the link at the top of today's show notes or going to

[00:09:19] todayindigital.com slash newsletter.

[00:09:22] Tomorrow, we will update this story with any reaction from

[00:09:25] Google on all this if they comment.

[00:09:27] Plus back to other stories making news this week, including PayPal's

[00:09:31] aggressive new push into the ad market.

[00:09:34] I'm Todd Maffin.

[00:09:35] Thanks for listening.

[00:09:36] See you tomorrow.

[00:09:40] Ah, the sizzle of McDonald's sausage.

[00:09:43] It's enough to make you crave your favorite breakfasts,

[00:09:47] enough to head over to McDonald's, enough to make you really wish

[00:09:52] this commercial were scratch and sniff.

[00:09:55] And if you're a sausage person now get two satisfyingly savory

[00:09:58] sausage, McGriddle sausage, biscuits, or sausage, burritos for

[00:10:01] just three 33 or mix and match price and participation may vary.

[00:10:05] Cannot be combined with any other offer.

[00:10:06] A combo meal, single item at regular price.