Tagged: Reputation Legal

The legal side of Reputation Management

Image Link Building – Whiteboard Friday 0

Posted by BritneyMuller

Image link building is a delicate art. There are some distinct considerations from traditional link building, and doing it successfully requires a balance of creativity, curiosity, and having the right tools on hand. In today’s Whiteboard Friday, Moz’s own SEO and link building aficionado Britney Muller offers up concrete advice for successfully building links via images.

Image Link Building

Click on the whiteboard image above to open a high-resolution version in a new tab!

<span id=”selection-marker-1″ class=”redactor-selection-marker”></span>

Video Transcription

Hey, Moz fans, welcome to another edition of Whiteboard Friday. Today we’re going to go over all things image link building, which is sort of an art. I’m so excited to dig into this with you.

Know your link targets

So first and foremost, you need to know your link targets:

I. Popular industry platforms – top pages

What are those top platforms or websites that you would really like to acquire a link from? Then, from there, you can start to understand who might be influencers on those platforms, who’s writing the content, who might you contact, and also what are the top pages currently for those sites. There are a number of tools that give you a glimpse into that information. Moz’s OSE, Open Site Explorer, will show you top pages. SEMrush has a top page report. SimilarWeb has a popular page report. You can dig into all that information there, really interesting stuff.

II. Old popular images – update!

You can also start to dig into old, popular images and then update them. So what are old popular images within your space that you could have an opportunity to revamp and update? A really neat way to sort of dig into some of that is BuzzSumo’s infographics filter, and then you would insert the topic. You enter the industry or the topic you’re trying to address and then search by the infographics to see if you can come across anything.

III. Transform popular content into images

You can also just transform popular content into images, and I think there is so much opportunity in doing that for new statistics reports, new data that comes out. There are tons of great opportunities to transform those into multiple images and leverage that across different platforms for link building.

IV. Influencers

Again, just understanding who those influencers are.

Do your keyword research

So, from here, we’re going to dive into the keyword research part of this whole puzzle, and this is really understanding the intent behind people searching about the topic or the product or whatever it might be. Something you can do is evaluate keywords with link intent. This is a brilliant concept I heard about a couple weeks back from Dan Shure’s podcast. Thank you, Dan. Essentially it’s the idea that keywords with statistics or facts after the keyword have link intent baked into the search query. It’s brilliant. Those individuals are searching for something to reference, to maybe link to, to include in a presentation or an article or whatever that might be. It has this basic link intent.

Another thing you want to evaluate is just anything around images. Do any of your keywords and pictures or photos, etc. have good search volume with some opportunities? What does that search result currently look like? You have to evaluate what’s currently ranking to understand what’s working and what’s not. I used to say at my old agency I didn’t want anyone writing any piece of content until they had read all of the 10 search results for that keyword or that phrase we were targeting. Why would you do that until you have a full understanding of how that looks currently and how we can make something way better?

Rand had also mentioned this really cool tip on if you find some keywords, it’s good to evaluate whether or not the image carousel shows up for those searches, because if it does, that’s a little glimpse into the searcher intent that leads to images. That’s a good sign that you’re on the right track to really optimize for a certain image. It’s something to keep in mind.

Provide value

So, from here, we’re going to move up to providing value. Now we’re in the brainstorming stage. Hopefully, you’ve gotten some ideas, you know where you want to link from, and you need to provide value in some way. It could be a…

I. Reference/bookmark Maybe something that people would bookmark, that always works.

II. Perspective is a really interesting one. So some of the most beautiful data visualizations do this extremely well, where they can simplify a confusing concept or a lot of data. It’s a great way to leverage images and graphics.

III. Printouts still work really well. Moz has the SEO Dev Cheat Sheet that I have seen printed all over at different agencies, and that’s really neat to see it adding value directly.

IV. Curate images. We see this a lot with different articles. Maybe the top 25 to 50 images from this tradeshow or this event or whatever it might be, that’s a great way to leverage link building and kind of getting people fired up about a curated piece of content.

Gregory Ciotti — I don’t know if I’m saying that right — has an incredible article I suggest you all read called “Why a Visual Really Is Worth a Thousand Words,” and he mentions don’t be afraid to get obvious. I love that, because I think all too often we tend to overthink images and executing things in general. Why not just state the obvious and see how it goes? He’s got great examples.

Optimize

So, from here, we are going to move into optimization. If any of you need a brush-up on image optimization, I highly suggest you check out Rand’s Whiteboard Friday on image SEO. It covers everything. But some of the basics are your…

Title

You want to make sure that the title of the image has your keyword and explains what it is that you’re trying to convey.

Alt text

This was first and foremost designed for the visually impaired, so you need to be mindful of visually impaired screen readers that will read this to people to explain what the image actually is. So first and foremost, you just need to be helpful and provide information in a descriptive way to describe that image.

Compression

Compression is huge. Page speed is so big right now. I hear about it all the time. I know you guys do too. But one of the easiest ways to help page speed is to compress those huge images. There’s a ton of great free tools out there, like Optimizilla, where you can bulk upload a bunch of large images and then bulk download. It makes it super easy. There are also some desktop programs, if you’re doing this kind of stuff all the time, that will automatically compress images you download or save. That might be worth looking into if you do this a lot.
You want to host the image. You want it to live on your domain. You want to house that. You can leverage it on other platforms, but you want sort of that original to be on your site.

SRCSET

Source set attribute is getting a little technical. It’s super interesting, and it’s basically this really incredible image attribute that allows you to set the minimum browser size and the image you would prefer to show up for different sizes. So you can not only have different images show up for different devices in different sizes, but you can also revamp them. You can revamp the same image and serve it better for a mobile user versus a tablet, etc. John Henshaw has some of the greatest stuff on source set. Highly suggest you look at some of his articles. He’s doing really cool things with it. Check that out.

Promotion

So, from here, you want to promote your images. You obviously want to share it on popular platforms. You want to reach back out to some of these things that you might have into earlier. If you updated a piece of content, make them aware of that. Or if you transformed a really popular piece of content into some visuals, you might want to share that with the person who is sharing that piece of content. You want to start to tap into that previous research with your promotion.

Inform the influencers

Ask people to share it. There is nothing wrong with just asking your network of people to share something you’ve worked really hard on, and hopefully, vice versa, that can work in return and you’re not afraid to share something a connection of yours has that they worked really hard on.

Monitor the image SERPs

From here, you need to monitor. One of the best ways to do this is Google reverse image search. So if you go to Google and you click the images tab, there’s that little camera icon that you can click on and upload images to see where else they live on the web. This is a great way to figure out who is using your image, where it’s being held, are you getting a backlink or are you not. You want to keep an eye on all of that stuff.

Two other tools to do this, that I’ve heard about, are Image Raider and TinEye. But I have not had great experience with either of these. I would love to hear your comments below if maybe you have.

Reverse image search with Google works the best for me. This is also an awesome opportunity for someone to get on the market and create a Google alert for images. I don’t think anyone is actually doing that right now. If you know someone that is, please let me know down below in the comments. But it could be a cool business opportunity, right? I don’t know.

So for monitoring, let’s say you find your image is being used on different websites. Now you need to do some basic outreach to get that link. You want to request that link for using your image.

This is just a super basic template that I came up with. You can use it. You can change it, do whatever you want. But it’s just:

Hi, [first name].
Thank you so much for including our image in your article. Great piece. Just wondering if you could link to us.com as the source.
Thanks,
Britney

Something like that. Something short, to the point. If you can make it more personalized, please do so. I can’t stress that enough. People will take you way more seriously if you have some nugget of personal information or connection that you can make.

From there, you just sort of stay in this loop. After you go through this process, you need to continue to promote your content and continue to monitor and do outreach and push that to maximize your link building efforts.
So I hope you enjoyed this. I look forward to hearing all of your comments and thoughts down below in the comments. I look forward to seeing you all later. Thanks for joining us on this edition of Whiteboard Friday. Thanks.

Video transcription by Speechpad.com

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!


Source: Moz

Narrow Definition of “Personally Identifiable Information” Kills VPPA Case–Eichenberger v. ESPN 0

Screen Shot 2017-12-13 at 8.56.07 AMThis is a Video Privacy Protection Act case. Plaintiff downloaded the WatchESPN channel on his Roku device and used it to watch videos. When he watched a video, ESPN disclosed the plaintiff’s device serial number and video title to Adobe Analytics. Plaintiff alleged that Adobe used the information to “identify specific consumers” by connecting the data shared with existing data in Adobe’s database. Adobe obtained the information in its database (used to identify consumers) from sources other than ESPN. Adobe gives the data it derives back to ESPN in an aggregated form, and ESPN “in turn provides advertisers with aggregate information about its users’ demographics.”

Plaintiff alleged Adobe used the data ESPN provided to identify him as having watched specific videos. He asserted that ESPN’s disclosure violated his rights under the VPPA. The district court dismissed, and the Ninth Circuit affirmed.

Standing: The Ninth Circuit first tackles standing under Spokeo and has no trouble concluding that plaintiff satisfies Article III standing. Spokeo dealt with “bare procedural violations” and held that in some circumstances, these violations could be sufficient to satisfy standing. However, plaintiff alleged the violation of a restriction on disclosure, and unlike a procedural rule, every violation of this restriction “offends the interests that the statute protects.” The court also cites to Congressional judgment as supporting this interpretation. The statute was intended to give consumers greater control over their personal information. Similarly, privacy torts don’t require actual harm (beyond the disclosure) to be legally sufficient. Finally, the court looks to the VPPA’s history and says that the disclosure of Judge Bork’s viewing records were not themselves harmful (they were “decidedly commonplace”). Under ESPN’s interpretation–that he needed to have been “aggrieved” beyond the disclosure–the VPPA would not have even provided him with recourse.

“Personally Identifiable Information”: The court says “personally identifiable information” extends beyond actual viewing or rental records, and covers some information that “can be used” to identify an individual. But it’s unclear precisely what the definition encompasses. The key question is whether the statute embodies a purely subjective or a foreseeability-type standard. Some courts have said that if information would be reasonably likely to reveal a person’s identity, then it’s personally identifiable information (Yershov). Others have looked to the perspective of the disclosing party (Nickelodeon), and the Ninth Circuit adheres to this view:

the statute views disclosure from the perspective of the disclosing party. It looks to what information a video service provider discloses, not what the recipient of that information decides to do with it.

The court follows the Third Circuit and adopts a subjective, “ordinary person” test.

The court says this interpretation “fits most neatly” with the regime Congress must have had in mind in 1988. In that time, a video store clerk would know:  if he or she disclosed someone’s name and viewing history that they were violating the statute. But if they merely described the person in more generic terms, there would be no violation. The court says the Roku device serial number is more like the latter scenario: “It created a sizeable ‘pool’ of possible viewers . . . ” The court notes that technological changes have undoubtedly altered the ability of companies to derive identity information using data but

the advent of the internet did not change the disclosing-party focus of the statute.

__

It would be really interesting to see data on how the numerous VPPA class actions have panned out. There have been many!

The court’s test, like many other Ninth Circuit tests, leaves you with a lot of questions. The court says that some information beyond conventional identity information still constitutes personally identifiable information, but it never provides any meaningful examples. Nor does it provide any scenario in which disclosure of this other category of information that can be used to derive identity would trigger the statute.

As to the subjective test, does this mean that the disclosing party has to know that the recipient can use the information to de-identify? Would contractual assurances from the receiving party help insulate the disclosing party’s conduct? The court’s concerns about liability being imposed by something outside of the disclosing party’s control resonate, but the test it articulates seems to leave a lot of room for workarounds. And ultimately, perhaps the court’s decision is driven by the fact that there’s no human reviewing the de-identified records, and it all occurs as part of a large-scale data transaction between companies.

The court describes what looks like a mild circuit split. I wonder if there’s any chance the Supreme Court could take up a case presenting this question?

Eric’s Comments

1) We’ve occasionally blogged on privacy concerns about analytics services (e.g., Garcia v. Zimride). This case raised the issue squarely, and it worked out for the defendants this time. However, if you are running an analytics service, you’re probably increasingly nervous about the wide range of potential privacy risks you face. And if you are a publisher using an analytics service, this case highlights the potential exposure you might face from your analytics provider’s activities.

2) The court’s approach to reidentification is interesting: the disclosing publisher isn’t responsible for an analytics service’s potential or actual reidentification unless the publisher discloses information that readily permits an “ordinary person” to identify a particular individual as having watched certain videos. This is a defendant-favorable definition of reidentification, and it will be interesting to see if courts similarly do not hold publishers accountable for the reidentification capacities of third parties in non-VPPA contexts.

Case citation: Eichenberger v. ESPN, Inc., 2017 WL 5762817 (9th Cir. Nov. 29, 2017)

Related news:

Netflix recently tweeted the following and predictably came under fire for it:

Related posts:

VPPA Still Doesn’t Protect App Downloaders–Perry v. CNN

Important and Troubling Video Privacy Protection Act (VPPA) Ruling From First Circuit–Yershov v. Gannett

App Users Aren’t “Subscribers” Under the VPPA–Ellis v. Cartoon Network

9th Circuit Rejects VPPA Claims Against Netflix For Intra-Household Disclosures

Court Rejects VPPA Claim Against Viacom and Google Based on Failure to Disclose Identity

Court Says Plaintiff Lacks Standing to Pursue Failure-to-Purge Claim Under the VPPA – Sterk v. Best Buy

Judge Dismisses Claims Against Pandora for Violating Michigan’s Version of the VPPA – Deacon v. Pandora Media

Disney Not Liable For Disclosing Device IDs And Viewing Habits

Ninth Circuit Rejects Video Privacy Protection Act Claims Against Sony

AARP Defeats Lawsuit for Sharing Information With Facebook and Adobe

Lawsuit Fails Over Ridesharing Service’s Disclosures To Its Analytics Service–Garcia v. Zimride

Android ID Isn’t Personally Identifiable Information Under the Video Privacy Protection Act

Minors’ Privacy Claims Against Viacom and Google Over Disclosure of Video Viewing Habits Dismissed

Hulu Unable to Shake Video Privacy Protection Act Claims


Source: Eric Goldman Legal

Moz the Monster: Anatomy of an (Averted) Brand Crisis 0

Posted by Dr-Pete

On the morning of Friday, November 10, we woke up to the news that John Lewis had launched an ad campaign called “Moz the Monster“. If you’re from the UK, John Lewis needs no introduction, but for our American audience, they’re a high-end retail chain that’s gained a reputation for a decade of amazing Christmas ads.

It’s estimated that John Lewis spent upwards of £7m on this campaign (roughly $9.4M). It quickly became clear that they had organized a multi-channel effort, including a #mozthemonster Twitter campaign.

From a consumer perspective, Moz was just a lovable blue monster. From the perspective of a company that has spent years building a brand, John Lewis was potentially going to rewrite what “Moz” meant to the broader world. From a search perspective, we were facing a rare possibility of competing for our own brand on Google results if this campaign went viral (and John Lewis has a solid history of viral campaigns).

Step #1: Don’t panic

At the speed of social media, it can be hard to stop and take a breath, but you have to remember that that speed cuts both ways. If you’re too quick to respond and make a mistake, that mistake travels at the same speed and can turn into a self-fulfilling prophecy, creating exactly the disaster you feared.

The first step is to get multiple perspectives quickly. I took to Slack in the morning (I’m two hours ahead of the Seattle team) to find out who was awake. Two of our UK team (Jo and Eli) were quick to respond, which had the added benefit of getting us the local perspective.

Collectively, we decided that, in the spirit of our TAGFEE philosophy, a friendly monster deserved a friendly response. Even if we chose to look at it purely from a pragmatic, tactical standpoint, John Lewis wasn’t a competitor, and going in metaphorical guns-blazing against a furry blue monster and the little boy he befriended could’ve been step one toward a reputation nightmare.

Step #2: Respond (carefully)

In some cases, you may choose not to respond, but in this case we felt that friendly engagement was our best approach. Since the Seattle team was finishing their first cup of coffee, I decided to test the waters with a tweet from my personal account:

I’ve got a smaller audience than the main Moz account, and a personal tweet as the west coast was getting in gear was less exposure. The initial response was positive, and we even got a little bit of feedback, such as suggestions to monitor UK Google SERPs (see “Step #3”).

Our community team (thanks, Tyler!) quickly followed up with an official tweet:

While we didn’t get direct engagement from John Lewis, the general community response was positive. Roger Mozbot and Moz the Monster could live in peace, at least for now.

Step #3: Measure

There was a longer-term fear – would engagement with the Moz the Monster campaign alter Google SERPs for Moz-related keywords? Google has become an incredibly dynamic engine, and the meaning of any given phrase can rewrite itself based on how searchers engage with that phrase. I decided to track “moz” itself across both the US and UK.

In that first day of the official campaign launch, searches for “moz” were already showing news (“Top Stories”) results in the US and UK, with the text-only version in the US:

…and the richer Top Stories carousel in the UK:

The Guardian article that announced the campaign launch was also ranking organically, near the bottom of page one. So, even on day one, we were seeing some brand encroachment and knew we had to keep track of the situation on a daily basis.

Just two days later (November 12), Moz the Monster had captured four page-one organic results for “moz” in the UK (at the bottom of the page):

While it still wasn’t time to panic, John Lewis’ campaign was clearly having an impact on Google SERPs.

Step #4: Surprises

On November 13, it looked like the SERPs might be returning to normal. The Moz Blog had regained the Top Stories block in both US and UK results:

We weren’t in the clear yet, though. A couple of days later, a plagiarism scandal broke, and it was dominating the UK news for “moz” by November 18:

This story also migrated into organic SERPs after The Guardian published an op-ed piece. Fortunately for John Lewis, the follow-up story didn’t last very long. It’s an important reminder, though, that you can’t take your eyes off of the ball just because it seems to be rolling in the right direction.

Step #5: Results

It’s one thing to see changes in the SERPs, but how was all of this impacting search trends and our actual traffic? Here’s the data from Google Trends for a 4-week period around the Moz the Monster launch (2 weeks on either side):

The top graph is US trends data, and the bottom graph is UK. The large spike in the middle of the UK graph is November 10, where you can see that interest in the search “moz” increased dramatically. However, this spike fell off fairly quickly and US interest was relatively unaffected.

Let’s look at the same time period for Google Search Console impression and click data. First, the US data (isolated to just the keyword “moz”):

There was almost no change in impressions or clicks in the US market. Now, the UK data:

Here, the launch spike in impressions is very clear, and closely mirrors the Google Trends data. However, clicks to Moz.com were, like the US market, unaffected. Hindsight is 20/20, and we were trying to make decisions on the fly, but the short-term shift in Google SERPs had very little impact on clicks to our site. People looking for Moz the Monster and people looking for Moz the search marketing tool are, not shockingly, two very different groups.

Ultimately, the impact of this campaign was short-lived, but it is interesting to see how quickly a SERP can rewrite itself based on the changing world, especially with an injection of ad dollars. At one point (in UK results), Moz the Monster had replaced Moz.com in over half (5 of 8) page-one organic spots and Top Stories – an impressive and somewhat alarming feat.

By December 2, Moz the Monster had completely disappeared from US and UK SERPs for the phrase “moz”. New, short-term signals can rewrite search results, but when those signals fade, results often return to normal. So, remember not to panic and track real, bottom-line results.

Your crisis plan

So, how can we generalize this to other brand crises? What happens when someone else’s campaign treads on your brand’s hard-fought territory? Let’s restate our 5-step process:

(1) Remember not to panic

The very word “crisis” almost demands panic, but remember that you can make any problem worse. I realize that’s not very comforting, but unless your office is actually on fire, there’s time to stop and assess the situation. Get multiple perspectives and make sure you’re not overreacting.

(2) Be cautiously proactive

Unless there’s a very good reason not to (such as a legal reason), it’s almost always best to be proactive and respond to the situation on your own terms. At least acknowledge the situation, preferably with a touch of humor. These brand intrusions are, by their nature, high profile, and if you pretend it’s not happening, you’ll just look clueless.

(3) Track the impact

As soon as possible, start collecting data. These situations move quickly, and search rankings can change overnight in 2017. Find out what impact the event is really having as quickly as possible, even if you have to track some of it by hand. Don’t wait for the perfect metrics or tracking tools.

(4) Don’t get complacent

Search results are volatile and social media is fickle – don’t assume that a lull or short-term change means you can stop and rest. Keep tracking, at least for a few days and preferably for a couple of weeks (depending on the severity of the crisis).

(5) Measure bottom-line results

As the days go by, you’ll be able to more clearly see the impact. Track as deeply as you can – long-term rankings, traffic, even sales/conversions where necessary. This is the data that tells you if the short-term impact in (3) is really doing damage or is just superficial.

The real John Lewis

Finally, I’d like to give a shout-out to someone who has felt a much longer-term impact of John Lewis’ succesful holiday campaigns. Twitter user and computer science teacher @johnlewis has weathered his own brand crisis year after year with grace and humor:

So, a hat-tip to John Lewis, and, on behalf of Moz, a very happy holidays to Moz the Monster!

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!


Source: Moz

Facebook Defeats Lawsuit Over Tracking Logged-Out Users–In re Facebook Internet Tracking 0

Screen Shot 2017-12-12 at 7.18.12 AMFacebook allegedly improperly tracked the activity of logged-out Facebook users on third party websites. Plaintiffs asserted claims based on common law rights and based on federal and state statutes, but the court previously rejected those. In the latest ruling, the court dismisses Plaintiffs’ claims based on breach of contract and breach of the duty of good faith.

Plaintiffs argued that Facebook promised not to track logged out users, but these promises were contained in documents other than Facebook’s terms of service (its “Statement of Rights and Responsibilities”).

Plaintiffs cited to language from Facebook’s “data use policy” stating that Facebook receives data from the websites that users visit, including “if [a user is] logged into Facebook” the user ID. The key question was whether this language was incorporated by reference into the terms of use. The court says no, because the version of the privacy policy plaintiffs point to was not in existence at the time of the relevant versions of the terms of use plaintiffs allegedly relied on. In fact, Facebook changed the title of its privacy policy to a “data use policy,” and the terms of service plaintiffs relied on did not reference a “data use policy” at all.

Plaintiffs also pointed to certain “help center pages” but the court says none of those are referenced in the privacy policy. Even to the extent the terms incorporate the privacy policy, this doesn’t help plaintiffs. Plaintiffs in response argued that all of the various help center pages are part of a “single broader document” but the court says this argument has no factual basis. The various pages have different URLs and were attached by plaintiffs as different exhibits. Moreover:

[n]o evidence suggests that a Facebook user who reads one Help Center page has also read, or is even aware of, any of the others.

Plaintiffs also relied on a breach of the duty of good faith but the court says that this duty has to be anchored to a specific contractual provision and plaintiffs cannot point to one here.

__

Ouch. As the court notes, this lawsuit has been through several rounds of motion practice. It was whittled down until it was extinguished by the most recent order. Plaintiffs have yet to file a notice of appeal, although their appeal deadline has not yet run.

That websites have not given consumers meaningful choice on being tracked while logged out is one of the great failures of modern U.S. privacy laws. This seems like a basic privacy practice that a website should make obvious to a user and allow the user to opt-out from. Yet, as this case illustrates, no rule exists to force websites to do this.

Facebook’s terms of service and related documents are voluminous, and the court appears to employ a fairly technical reading in determining whether certain help page documents form a part of the contract. You should not need to be a lawyer to figure this stuff out.

Facebook is under a 20 year consent decree with the FTC. Perhaps this activity does not come under the scope of the consent decree, because it doesn’t involve any affirmative misrepresentations or over-riding of expressed consumer preferences? Either way, this seems like the type of thing a regulator may be well suited (or perhaps better suited than the plaintiffs’ bar) to address, and which may be falling through the cracks.

Case citation: In re Facebook Internet Tracking Litig., 2017 U.S. Dist. LEXIS 190819 (N.D. Cal. Nov. 17, 2017) [pdf]

Related posts:

Facebook Scores Initial Win Against Privacy Plaintiffs Over Data Leakage Claims — In re Facebook Privacy Litigation

Facebook Privacy Class Action Filed by Lanier Firm Voluntarily Dismissed — Melkonian v. Facebook

Facebook Gets Bad Ruling In Face-Scanning Privacy Case–In re Facebook Biometric Information Privacy Litigation

Facebook Beats Privacy Lawsuit Alleging Persistent Tracking

The “I Didn’t Understand Facebook’s Privacy Settings” Argument Isn’t Persuasive to Judges–Sumien v. CareFlite

Facebook and Zynga Privacy Litigation Dismissed With Prejudice [Catch up Post]


Source: Eric Goldman Legal

Keyword Research Beats Nate Silver’s 2016 Presidential Election Prediction 0

Posted by BritneyMuller

100% of statisticians would say this is a terrible method for predicting elections. However, in the case of 2016’s presidential election, analyzing the geographic search volume of a few telling keywords “predicted” the outcome more accurately than Nate Silver himself.

The 2016 US Presidential Election was a nail-biter, and many of us followed along with the famed statistician’s predictions in real time on FiveThirtyEight.com. Silver’s predictions, though more accurate than many, were still disrupted by the election results.

In an effort to better understand our country (and current political chaos), I dove into keyword research state-by-state searching for insights. Keywords can be powerful indicators of intent, thought, and behavior. What keyword searches might indicate a personal political opinion? Might there be a common denominator search among people with the same political beliefs?

It’s generally agreed that Fox News leans to the right and CNN leans to the left. And if we’ve learned anything this past year, it’s that the news you consume can have a strong impact on what you believe, in addition to the confirmation bias already present in seeking out particular sources of information.

My crazy idea: What if Republican states showed more “fox news” searches than “cnn”? What if those searches revealed a bias and an intent that exit polling seemed to obscure?

The limitations to this research were pretty obvious. Watching Fox News or CNN doesn’t necessarily correlate with voter behavior, but could it be a better indicator than the polls? My research says yes. I researched other media outlets as well, but the top two ideologically opposed news sources — in any of the 50 states — were consistently Fox News and CNN.

Using Google Keyword Planner (connected to a high-paying Adwords account to view the most accurate/non-bucketed data), I evaluated each state’s search volume for “fox news” and “cnn.”

Eight states showed the exact same search volumes for both. Excluding those from my initial test, my results accurately predicted 42/42 of the 2016 presidential state outcomes including North Carolina and Wisconsin (which Silver mis-predicted). Interestingly, “cnn” even mirrored Hillary Clinton, similarly winning the popular vote (25,633,333 vs. 23,675,000 average monthly search volume for the United States).

In contrast, Nate Silver accurately predicted 45/50 states using a statistical methodology based on polling results.

Click for a larger image

This gets even more interesting:

The eight states showing the same average monthly search volume for both “cnn” and “fox news” are Arizona, Florida, Michigan, Nevada, New Mexico, Ohio, Pennsylvania, and Texas.

However, I was able to dive deeper via GrepWords API (a keyword research tool that actually powers Keyword Explorer’s data), to discover that Arizona, Nevada, New Mexico, Pennsylvania, and Ohio each have slightly different “cnn” vs “fox news” search averages over the previous 12-month period. Those new search volume averages are:

“fox news” avg monthly search volume

“cnn” avg monthly search volume

KWR Prediction

2016 Vote

Arizona

566333

518583

Trump

Trump

Nevada

213833

214583

Hillary

Hillary

New Mexico

138833

142916

Hillary

Hillary

Ohio

845833

781083

Trump

Trump

Pennsylvania

1030500

1063583

Hillary

Trump

Four out of five isn’t bad! This brought my new prediction up to 46/47.

Silver and I each got Pennsylvania wrong. The GrepWords API shows the average monthly search volume for “cnn” was ~33,083 searches higher than “fox news” (to put that in perspective, that’s ~0.26% of the state’s population). This tight-knit keyword research theory is perfectly reflected in Trump’s 48.2% win against Clinton’s 47.5%.

Nate Silver and I have very different day jobs, and he wouldn’t make many of these hasty generalizations. Any prediction method can be right a couple times. However, it got me thinking about the power of keyword research: how it can reveal searcher intent, predict behavior, and sometimes even defy the logic of things like statistics.

It’s also easy to predict the past. What happens when we apply this model to today’s Senate race?

Can we apply this theory to Alabama’s special election in the US Senate?

After completing the above research on a whim, I realized that we’re on the cusp of yet another hotly contested, extremely close election: the upcoming Alabama senate race, between controversy-laden Republican Roy Moore and Democratic challenger Doug Jones, fighting for a Senate seat that hasn’t been held by a Democrat since 1992.

I researched each Alabama county — 67 in total — for good measure. There are obviously a ton of variables at play. However, 52 out of the 67 counties (77.6%) 2016 presidential county votes are correctly “predicted” by my theory.

Even when giving the Democratic nominee more weight to the very low search volume counties (19 counties showed a search volume difference of less than 500), my numbers lean pretty far to the right (48/67 Republican counties):

It should be noted that my theory incorrectly guessed two of the five largest Alabama counties, Montgomery and Jefferson, which both voted Democrat in 2016.

Greene and Macon Counties should both vote Democrat; their very slight “cnn” over “fox news” search volume is confirmed by their previous presidential election results.

I realize state elections are not won by county, they’re won by popular vote, and the state of Alabama searches for “fox news” 204,000 more times a month than “cnn” (to put that in perspective, that’s around ~4.27% of Alabama’s population).

All things aside and regardless of outcome, this was an interesting exploration into how keyword research can offer us a glimpse into popular opinion, future behavior, and search intent. What do you think? Any other predictions we could make to test this theory? What other keywords or factors would you look at? Let us know in the comments.

Also, if you’ve enjoyed this post, check out Sam Wang’s Google-Wide Association Studies! –Fascinating read.

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don’t have time to hunt down but want to read!


Source: Moz