All the sessions were awesome today and the big bash at Garage was epic! I won the special edition Roger bobblehead today! *Pats myself on the back*
For all those who could not make it to the MozCon party, here are the key takeaways.
Marketing Innovations: Creative PR, Content, and SEO Strategies
Lexi Mills, @leximills
What’s the common thread between any of these marketing positions? The game is getting harder. You have to compete with huge brands to get on the first page of Google.
With the latest mobile update some of the largest publishers lost over 20% of their traffic. How are these companies responding?
30% reduction in outbound links
2o% drop in referencing and citations
Everyone is trying to keep people on their site so they link out less often.
It’s also harder to get free links, because why would an agency or publication give you space on an article, when someone else is willing to pay for it?
Publications are having less of an impact than they used to.
Tools are no longer a luxury. You need to get the best tools you can get and test the toolbox that works best for you.
Gorkana - details, phone numbers, emails for UK media
MyMediaInfo - great for getting U.S. data
Yesware: great tool for tracking email opens. Employees and Sales people keep their pace up when they’re using a tool like this, because they know when their customers are looking at their message.
If you’re trying to compete in a really competitive landscape, doesn't it make sense to push a campaign from every corner of the globe? Make content that everyone on your team can push.
To help, you need to make journalist friends.
News agencies create the content and then sell it to larger publications. Make friends at these news agencies.
Storytelling - it’s so important
“No story lives unless someone wants to listen” - JK Rowling
Be at the center of a story.
Think about what stories were big in the past. Which ones are still being talked about.
Selfies were a possible topic in the media, so we created a piece of content about the top 10 selfie locations - it got picked up really well.
Finding topics and stories that are popular are tough - learn from your research companies. They know what works.
We created a Yelp calendar with all the important dates are audience would want to be aware of.
Look at Netflix, it shows you all the trending shows and when the new ones are going to launch. Following these trends are going to help you create content ideas.
The tactic that trumps all tactics: Speed
This story was a huge success. We saw it and took advantage of it with one of our fashion clients.
We sent an employee to get that exact dress and brought in media to do a photo shoot within the day.
Is it all about a Good Idea? What’s harder than getting a good idea? Selling a good idea.
Images can make your campaign a huge success.
One important thing to do when running a great campaign - Sell it back. Report on what worked. Show your client the key highlights and sound bites that people said about you in the media and social. Don’t pack your report with social metrics that nobody reads.
Order your report based on where your budget is coming from. If you’re doing PR but the budget is coming from SEO, put your SEO stats at the front. The SEO director is going to be the one to sign off on your next project.
Building on Success
If we’ve got success with one campaign, let’s build off it.
How do you make it really difficult for people to write about you without linking to you. Answer: Product campaigns
We made sure there was always fresh links and coverage coming in → this proves to Google that the story is still relevant.
Upside Down and Inside Out
Mig Reyes, @migreyes
- We don’t see ourselves as artists like we did when we were kids
- Think in a different way about how we approach our work
Burn Your Business Cards
- Why do we have to lead with our title?
- We are not defined by what we do or what’s on our business card
- We have lots of other skill sets that we’re proud of that we don’t necessarily put on that little card
- There are no titles on the Basecamp staff business cards
- Employees are encouraged to do cool shit
- Emily wanted to develop writing skills and personalized the automatic basecamp emails
- Everyone on the basecamp support team learned how to use terminal and markdown and tried something new
- Don’t let the dev team hold you back either
- Mig, himself, learned how to code on his own in order to use an API for campfire to share fun links with the rest of his community.
- All the funny side work you do, and the business cards you toss to the side can be applied to your business.
- Use this to win the hearts and minds of people
- Mig used a custom message when people squished their desktop sites to test responsive to let them know that Basecamp works great on mobile
- That turned into commenters on YouTube asking him to do a project which ended up with him being able to move to the city and furnish his whole apartment for a year.
- All because he did something different.
- Example: Mig created an entry for a college film festival as a graphic designer - not a film student - because he knew he could.
- Nobody tells adults anymore that they’re craftspeople and artists.
- So if you have a business card - don’t worry about it.
- Let’s break things - because this is how we learn!
Go to humblepied.com to see a bunch of great advice from graphic designers
- Mig sourced this advice from people he admired and made a site out of it
If you don’t break a few things then you aren’t trying hard enough. - Mig’s boss at 37 signals
Make ugly things too!
- This resonates with people because when you’re a designer you learn the very traditional ways of doing things.
- Everything on the web looks the same, the call to actions all sound the same - Do. Something. Different.
You aren’t going to be a badass at anything if you don’t try.
- You are going to make things that suck - but who cares?
Ever thought that maybe ugly was intentional?
- IE, Craigslist
Trying new things doesn’t have to be a career switch.
Spend Less Time On Things
- When Mig worked at an agency they spent 15 minutes once every month to just create an art piece based on one word
- Doesn’t need to take a long time to do something cool
- Just practice your craft
This mentality leads to his work at Threadless.
Spending less time on things then lead to Layer Battle (check that out in Mig’s slide deck - link at the bottom of this set of notes).
Takeaway: Spend less time. Do things that scare you. Don’t be afraid of ugly work.
Mig has lots of examples in his slide deck of art he made that he’s proud of and art that he’s not. Download the deck from the MozCon site here: https://moz.com/mozcon#schedule
Rocking Your CRO Efforts with Radical Redesigns
Radical redesigns can be scary. Conversion rates could go down as easily as they could go up.
If you’re not radically redesigning key pages on your site throughout the year, you could be missing out on huge opportunities.
Testing the little things is like sharpening one chain link on your chainsaw at a time. What you should do is switch things up entirely to see what big wins you can get.
When you test a radical redesign, you reset the tables with your A/B testing process so you can level up and start your A/B tests again at a higher level.
At the start of a new test cycle
Research to identify your opportunities and strategy. Start testing, do your analysis and find a winner.
From there you move on to the iterative testing phase. You’re then going to test everything on that winning design to see how you can improve it further.
Once you’re done that, you start the whole process over again. This whole process should take more than 6 months. That’s because user behavior on the Web changes about every 6 months. There’s new trends, new best practices etc.
Once something has become a design best practice, it stops becoming a best practice.
Once users start to see things everywhere on the web, people start to ignore them. Think about parallax websites. You start to get numb after a while.
That’s why you need to radically change things - You’re operating under assumptions that are no longer effective on your audience
- Test multiple hypotheses
We tested 4 hypotheses
a) If we create an image variation that speaks more to the user, we’ll find higher conversion rates
b) Maybe we can test multiple different types of business owner images
c) By removing the image and giving more content, users will convert through the form better
d) If we clean things up and simplify the page, conversions will increase
The winner was the one without the images
- Separate desktop and mobile when testing
The use case for mobile devices is totally different
Desktop responsive version converted ok…
But maybe the mobile user needs a more customized approach:
Maybe users want to call in - let’s test various click to call cta
Let’s test different ways of showing the content
The winner had the prominent click to call cta with a form lower on the page.
Result: call volume and form completions increased
- Use your tools, break the rules
Just because something worked in the past, doesn’t mean it will work this time
User behavior changes so rapidly, you have to challenge the assumptions you are assuming to be CRO correct.
Users were not scrolling down the page on mobile. We gotta get people to scroll down further. What visual cues can we use to get them to do this?
But why would you want to force your audience to do that
We tested removing the content that was below the fold.
4. Run Radical Redesign Tests for at least 2 weeks
You want to try and go for a 95% statistical significance. But that significance, by itself, are useless.
You need to give your tests a long enough life span for radical results to normalize and level out.
Always follow up with iterative testing. Not every test can be radical.
Parole, Parole, Parole: Practical, Modern Keyword and Topical Research
Gianluca Fiorelli, @gfiorelli1 - International SEO and inbound strategist
Sometimes we feel lost when doing keyword research…sometimes I don’t have a budget.
Google has evolved. Semantics govern. So what can we do? What if we have a niche site, small budget, looking for content ideas? Especially if you’re working on a product or service, you know nothing about.
Let’s use google better.
I start at Google trends looking at content for a travel service:
I see seasonal spikes, but I was surprised by Japan.
Don’t be data driven. Be data informed. Data is for bots. Understand the story that the data tells you.
It turns out the searches for Japan were coming from Hawaii. That’s because there are a lot of Japanese there.
Use wordreference.com - whenever you have some doubt about language usage, use word reference to find out the differences between words. i.e. difference between “holiday” and “vacation”
So then I searched for…
Use the keyword planner better:
(don’t just look at search volume)
- Put your main topic
- Put your landing page
- Your industry
Use the Ad Group Ideas to find semantically relevant keywords
Use Google Suggest:
Google suggest is different in every niche. Google News suggests differently than other Google properties.
Use various web properties to see other autocomplete suggestions: Pinterest, Amazon etc.
Rebuild Wonder Wheel:
Start connecting all the related searches you have found from the research thus far.
Look at various Google SERPs to get more topic ideas. Rich snippets, carousels, knowledge base. These pieces of a SERP will also give you other hints about ideas.
If you look up destinations - Google displays results for the destination, categorized by beaches, nightlife, museums etc.
Carousels are not just a Google thing.
You want to know what your competitors are doing - try using SimilarWeb.
Look beyond organic search. Sometimes your competitors suck at SEO so their content is not so visible. Look at paid competition too.
Use SEM Rush to compare domain vs domain and identify keywords that everyone is using.
If you can, track the internal search on your website. This is great data if you can match it with conversion information. You can find out which keywords drive conversions (informational queries vs transactional queries etc.)
Talk like your audience:
- Go to forums, Quora, or advanced searches on Twitter - to find out what people are saying about a topic.
- check out Facebook communities and see what people discuss.
- Grab the URL of these communities and put in your Feedly to monitor the conversation.
http://itseo.org/knimekws ←- great tool for content analysis. TF-IDF Noes.
Use that information for internal linking.
Once you gather all your keyword research data. Use tools like Excel, Moz, etc. to aggregate the data.
“The greatest trick SEO ever pulled was convincing the world it didn’t exist.”
The Psychology of Social Media
Courtney Seiter, @courtneyseiter
- We mimic expressions - this is called social contagion
- This has created a new brain pattern within us
- When people use an emoji they feel the same way as the emoji feels
- As emojis become more prominent, slang falls - changing our language pattern
- This is important because emojis are a sign of social media power and influence. People who were rated the highest on competence use the most emojis online
- You can use an emoji as a hashtag
- Branded emojis - coca cola, ikea and comedy central have done this
- Not seen as advertising but as part of self-expression
Social media is at its best when you remember it’s about the humans
Astoundingly Useful Applications of Facebook Search for Marketers
David Mihm, @davidmihm
Context: Why you should care about Facebook Search
Facebook is getting a billion search queries a day.
Google’s search volume is declining in growth
To continue to grow our audience, we’re going to have to start looking at other platforms to generate traffic.
Apps continue to dominate the mobile web. Facebook is dominating the most used app on mobile phones.
Parts of a Facebook Search:
- Set Relationships
Professions can be subjects or objects.
Interests as subjects or objects
Well known people can also be used as objects
Example: “people who like Senator Jim Inhofe and Environment”
Syntax + Vocabulary - Constructing undefined queries
Essentially you are looking for keywords on a particular page type.
- Example: looking at professional profiles that people have created who mention “blogging”
Use Cases for Content Marketing
- What should I write about?
Search: “Interests liked by people who like [insert your page here]”
2. Newsletter blog content
Search: “Recent photos taken at [your business location here]”
3. Google Display/ Ad Network Targeting (Guest Articles)
Search: “Pages liked by people who like ______”
4. Local Partnerships
Search: “Pages liked by [geo-location] residents who like [Your business name]
5. Journalists/Bloggers at particular outlet
“Journalist who live in [geo-location] that like ______”
6. Use a similar process to find out Who’s writing and interested in your topic and live in your area
7. Find thought leaders in a specific area
” Presidents who like ______ and live in ______”
8. Good candidates for reviews
“People who used Yelp and like [your business name]”
9. Where should I sell my stuff?
Search: “ars in [geo-location] liked by people who like [company name]
10. Finding an internal cheerleader
11. Targeted offer outreach
Search: “Singles who like [company] in [location]”
You can plug individual Facebook profiles into the FullContact API
Need Help doing Facebook Search?
- Move beyond the page like
- It doesn’t do you much good to have page fans (less than 2% see your content)
- shared content is what can be searched and give you long term presence in Facebook.
- Algorithm is not dependant on link relationships (link profile)
- Drive App Authentications
- Make sure your app is indexable
Google Drives check-ins and reviews right now.
If you do a search of “bankruptcy lawyers in Portland”, you’ll get a set of relevant results.
It’s likely that check-ins and reviews will drive a large portion of the ranking algorithm.
Facebook was the #2 preferred place to leave business reviews.
Facebook reviews now show up as rich snippets in Google SERPs
Onsite SEO in 2015: An Elegant Weapon for a More Civilized Marketer
Rand Fishkin, @randfish
I love making Google more transparent.
Remember when we had one job? It was to perfectly optimize a page.
Links were greater than content.
By 2007 link spam was rampant. Even in 2012, it felt like Google was making liars out of the whitehat SEO world.
Over the last 3 years, things have changed. The hardware and software infrastructure improvements. Algorithms were rolled out that wiped out spammy sites and poor content.
Google got better at understanding intent. They started to look at language, not just keywords. They predicted when we wanted more diverse results. They learned that we wanted fresh information.
They began segmenting navigational vs. informational queries and can figure out what we want to find much better.
They learned to identify entities of knowledge and connect them to a topic and keyword. Brands became one of these identities that could be tied to topics.
These advancements brought Google back in line with its public statements of how SEO should be done. The “right thing to do” started to become more important to ranking.
During these advances, Google’s search team underwent a revolution. In its early years, Google rejected machine learning in its organic ranking algorithm.
They expressed concerns about machine learning in the algorithm.
In 2012, Google published an article about using machine learning to predict PPC click through rates. It would predictably learn.
and then in 2013, Matt Cutts talked about how Machine Learning (ML) was being used in organic search.
As ML takes over more of Google’s algorithm, the underpinnings of the rankings change.
Google is public about how they use machine learning in image recognition and classification:
Jeff Dean’s slides on Deep Learning are a must read - link in slide deck.
Machine learning in search could work similar to this:
Based off deep learning, we’re talking about algorithms that build algorithms.
Humans will no longer feed in ranking factors to the algorithm. The machines will develop them themselves.
Query success metrics will be all that matters to the machines. They don’t care what the ranking factors are, they just want to make sure they get the SERP right.
Now we’re going to be optimizing less for inputs and more for outputs:
The future of on-page SEO will be focused on those outputs. We can’t forget the inputs, we’ll still have to do both.
But today….do those outputs already affect SEO?
Remember our “Queries & Clicks” Test from 2014? What happened: Everyone at MozCon did the same search Query and then clicked on the same page. The result was that it got boosted to the top of the SERP.
Repeating that test and getting the same result is much harder now.
Google says using clicks to determine SERP rankings no longer works.
What about if we use long clicks vs short clicks?
Short click: click through have a short visit and hit the back button
Long click: click through and spend a long time on the page
So we ran another test and got a bunch of people to search the same query and long click on a specified result while short clicking on another specific result.
Result: the pages that got long clicked moved up in rankings.
BUT, after 8 hours it went back down. That’s because the machines can tell that this was an anomaly. All the clicks came within hours - it wasn’t valid.
We must now choose how we balance our work
The old signals (inputs) vs the new signals (outputs)
[classic on-site image]
5 Modern Day Tactics of onpage seo
- Punch above our average CTR?
- Average Title, meta description and URL a little bit for keywords, but optimized even more for clicks.
- Every element of the SERP listing is going to count
- Do you get the brand dropdown?
- [every element counts image]
- Consider repeated publication on the same topic, to see if we get better at nailing what our users want
- Try driving up CTR through branding or branded searchers
- Google search patent says: “The % of people who do branded search influences that brands visibility for non-branded searches”
- Paid search can secondarily influence organic success metrics because of brand recognition.
- Pages that get a lot of social shares, few links - seem to outperform
- Google says they don’t use social signals directly, but that seems suspicious
- [show example of low DA ranking]
- Rand thinks Google looks at engagement data (which is highly correlated with social shares)
- It’s likely that Google will take different factors into consideration when rankings pages in different verticals
- Google wants to see shares that result in engagement
- New KPI: shares and links per 1,000 visits
- Unique visits /
- Total visitors /
- Knowing what makes our readers share is going to help us make 10x content
- Don't’ ask “How can we make something as good as that?. Ask “how can we make something 10X better than these”
- 10x content is the only way to stand out in the future. There’s far too much noise already.
- None of our old SEO tactics will give us the 10x content
- Beating out your fellow SERP neighbour
- An SEO’s checklist for engagement
- Filling in the gaps in your visitors knowledge
- Google is looking for content signals that demonstrate that a searcher is getting all their needs fulfilled so they don’t need to search again.
- If your content doesn’t fill the gaps of your searchers needs, you won’t be seen as relevant.
- Example: your article is about New York, but you don’t talk about Brooklyn, Manhattan, Queens etc.
- Earning more shares, links and loyalty per visit
5. Get visitors to accomplish their tasks faster
- If Google sees people who perform a variety of searches, and see that the majority of searchers end their query on a specific page -that’s the page that they see as valuable - it will rank the best.
- Google wants to send people to sites that will help them complete their whole mission
We are in a 2 algorithm world
Algo 1: Google’s input
Algo 2: Subset of humanity that interacts with your content
“Make pages for people not engines” ←awful advice
You gotta do both! Take care of the outputs and inputs
Algo input vs human input
- This is what I mean when I talk about 10x content: bit.ly/10xcontent
- Monkeylearn created a tool just for MozCon to help with topic modeling keyword extraction and comparison: Bit.ly/monkeylearnseo
Thanks for reading. Until next time! J