Summary of MN Search Summit
From the eyes of Happy Dog
This year was another record-breaking year for the MN Search summit. Just shy of 400 attendees is an awesome feat.
And yes, we like when you hold a mic like a rapper.
Mike truly did pack in a ton of content. He mentions that people are hesitant to do big changes and really don’t take action until Google puts an algorithmic gun to their head. The main theme of the opening keynote is that there are problems that we should address. He talked about 4 core problems.
There is not an open source of truth
People don’t measure success correctly (don't use traffic as the measure!)
There are not enough effective business cases
People can’t keep up with technical advances
By addressing all of these, truly addressing these, you are ready to take a big step forward.
Some key takeaways were:
- We should improve our pitches and make a strong presentation. Do it in person and invite them all! Thank goodness I will be attending Lea Pica's session later.
- Instead of shying away from technical SEO, we should embrace it. We should know how browsers work, We should understand the document object model, critical path, page speed, log file analysis, JavsScript frameworks and how SEO interacts with them, Chrome Dev Tools, Accelerated Mobile Pages (AMP) and HTTP2 for starters.
- Some technical tidbits he mentioned were using onpage.org for checking on page optimization and using rel=preconnect and prerender for your assets
He also dropped some good nuggets like really leveraging Google Analytics with filters, grouping, audience insights and then pairing that with user intent and the user journey throughout the site.
Slides at bit.ly/same-old-seo
Session 1: Neuro-persuasion: Brain-based strategies for online marketers
5 years ago, academia would call neuromarketing bullshit. Now, not so much. Roger Dooley dives into neuroscience and psychology with marketing and the web. In fact, 5% of decisions are conscious decisions. 95% are unconscious decisions.
There has been a lot of research regarding this topic. But studies show that these studies are unreliable. :). Test, test test. Just note that there is no single theory that works for all. However, Roger introduced the “Slide Theory” or “The Persuasion Slide”. It works as follows:
Gravity: Gravity is the initial motivations, needs wants and goals. “Do this for us” is not the way to market to people. It fights against gravity. An example of this would be a “fill out this form” CTA. With gravity, the thought process is “We will help you out”. A good example of that would be “I’ll show you how to make 10k within minutes” or “Make $1,000 per month”.
Nudge: Get the attention and grab the eyes of the intended person. An example of a bad nudge would be a tiny envelope icon with no explanation. A good nudge is prominent.
Angle: Steeper the slide, the more rapidly people move. It’s the motivation you provide. There are conscious and nonconscious motivators.
But include them both. Nonconscious motivators are really effective and usually free. Showing emotion and attributes that are shared with customers are nonconscious motivators. When you do this, it a good idea to put this on your homepage. There are other common elements of nonconscious motivators such as social proof. Social proof is action or preference based. Preferences are more of an endorsement. Also, you can utilize cost perception. Adding a more expensive product will boost sales to a lower priced ones. Set a high anchor first. Another motivator is the scarcity model. Phrases such as “2 seats left” or “offer expires at…” are examples of utilizing the scarcity model.
Friction: Friction is what stops you from going down the slide. Avoid friction! Some examples of friction would be form fields, steps in the checkout process, instructions, what am I supposed to do? Friction is anything in the conversion process that would stop a user is friction. Amazon is a near zero friction model. A sample of friction could be disabled autofill. Use autofill, promote it. Another example of friction could be super strong passwords or expiring passwords. Auto-logouts are friction. The other kind of friction is imaginary friction and cognitive fluency. How hard is it for your brain to process something. Hard to read fonts. Focus on your colors, use short text, make your content easy to read. If it’s not motivation, it’s friction!
To recap, simply you should align with gravity, provide a nudge, maximize angle and eliminate friction. That creates the best opportunity for you to guide your users where you want them to.
Session 2: Hacking Rankbrain
Rankbrain is the 3rd most weighted ranking signal and is currently leveraged for obscure long-tail queries. Machine learning is what search rankings will be based on eventually. Larry does his presentation as “John Connor, sent back in time to warn us all about machine learning”. What a funny guy.
What is machine learning? It is Google as we know it. It is a manual hand coded algorithm currently, but eventually, it is moving to a machine-coded and machine-learned algorithm. It basically has to guess and check if it provided the best result to the searcher. Google already uses machine learning everywhere else, for years. Images, videos and other assets of Google already. AdWords has been machine learned since day one, for 15 years. Facebook news feed and ad targeting use this. Even stupid Twitter ads are like this.
What is Rankbrain?
Rankbrain is query interpretation that affects rank. Did the search result make the user happy? If so, show it again. If not, move it down the ranks. It tries to match the intent of the user and provide the best result.
Rule of thumb, every 3% of improved click-through rate, you rise one spot. This is big for any PPC marketer. CTR is huge to focus on. But what do we prepare for with Rankbrain and other machine learned algorithms?
First, realize they won’t tell you everything. Then realize your content is auditioning. It’s auditioning for Rankbrain. As for crushing your auditions, that’s not as easy it is sounds -- it’s tough nowadays.
First, get a donkey/unicorn detector. Is your content a crappy donkey or a remarkable unicorn? You must build unicorns, not donkeys. Then find your click-through rate, and you can do this through Google Webmaster Console. SERP differs greatly by industry, so it is not as cut and dry. But the best you can do is analyze your own data.
Donkeys are your bottom 10% performers, typically 3x worse than average. Unicorns can perform twice as good as the average. Can you convert a donkey to a unicorn? Yes! The SEO title is a good place to begin for that. If you do PPC, use dynamic keyword insertion for “ok” results, but unicorns don’t really rely on that. To get unicorns, leverage emotional trigger. Then write the headline from personas.
Then format it with, insert an emotional hook and work a little magic on it. Make sure your content type and topic are all set and you are on your way to a unicorn ad. Unicorn headlines are the top 10%. But the odds of finding a unicorn out of that is 1:10. So, write 10 headlines for the intended unicorn. Then test, and try and find your unicorn. Bid to the same position. It’s okay to test on a cheaper target audience. Use a broad match keyword match type.
Now we look at an engagement rate unicorn/donkey detector. For that, we look at the bounce rate vs. organic position. 76% and lower seems to be the natural rate for spots 1-4 and the lower spots typically have 78%-88%. Also, the dwell time is factored. So, do they just view your page and quickly leave? Google is checking that out. Average time on site for the top spots seems to be roughly 500ms based on some initial studies. Next, let’s look at some great data for conversion rates:
But believe it or not, conversion rates are related to click through rates. So focus on those. And don’t do this by focusing on little changes. Do big changes to increase conversion rates.
Another “weapon” is Facebook ads and Google Ad network. Advertising dramatically affects organic click-through rates. Promote your unicorns through these ad networks through retargeting. Once they see your brand, even if they don’t do a branded search, they are biased towards you now. Remarketing is great. You can do unicorn video ads and remarket that. After that, you can delete your “bad neighbors”. Delete your low, bouncy, spammy neighborhoods on your site. Old tactics just don’t work anymore. Don’t just wait for Rankbrain and machine learning algorithms. Be proactive. Get in “Unicorn Land”.
Session 3: Make your remarketing more than an echo
Remarketing has become a standard practice when it comes to PPC and it makes sense. On average, remarketing has positive ROI, increases branding and awareness and can also have a lower CPA. But too often, we just launch it and walk away.
Why is remarketing essential to paid search? Thomas Smith estimated (in 1885) that people will look at a product 20 times before they buy it. In the 30s, there was the “Rule of 7”. 7 meaning the number of times needed for a message to “stick”.
We should learn to remarket without being “creepy”. The first thing we look at is “Why didn’t they click in the first place?”. Did they just not need it or need to do more research? What do you want them to do next? We start with choosing the remarketing type. There is standard and dynamic, and dynamic is killing it.
Google AdWords is one method. It requires familiarity with code, so hopefully you love code like we do :). Start with the Dynamic Quick Start. But retail is a little different. Google Docs / Sheets can be used here to make it easy. Next, look at custom audiences and Smart Lists. Also, cap your frequency. 5-7 times per day is enough. Test multiple layouts. As Larry would say, you can find your unicorn this way.
You can try separate campaigns for text and image ads. But test, test, test. Give the best performers more of a budget and snip out the bad performing ones. Use the ready to go lists that Google provides. Use managed placements for extra control. They also have a conversion optimizer -- use it! Lastly, there is target & bid vs bid only. Bid only, typically performs well for shopping. Target & bid works well for the other methods.
Next, figure out how you are going to segment your audience. Also, keep in mind you should set a CPA for each audience. Then use demographic and interests to narrow automatic placements. You can do this with larger lists like “all site visitors”.
Next, focus on the “buy in”. You need to insert code for this, so pick an awesome company like Happy Dog to do this for you. You can utilize this to increase conversion from cart abandoners. If you do this, retarget cart abandoners within 24 hours. It tapers off after that.
Then, continue the conversation with RLSA. Bid on more generic/head term keywords. Exclude or bid down converted users. Also, shepherd behaviors to different landing pages. Then combine for unicorns! RLSA and retargeting. It’s like a waterfall.
Facebook Dynamic Product Ads can produce great results. And it’s cheaper. It is very similar to Google, just not as large of a platform. The right rail converts a little better than the newsfeed on a desktop. But use caution. You still need a first time engagement. It also cannibalizes other channels since it has a last-click attribution. So, use your overall revenue numbers, not just Facebook’s.
Finally, you wait. 2 days generates the most recent. 7 days was second most. 28 days was third then 14 was fourth. Why? No one really knows. But wow, what great data!
Tips and tricks
Have each placement have its own ad set.
Start bigger. Narrow down from there.
Watch the size of account/campaigns.
Power Editor is your friend.
Is there a time when remarketing isn’t for you? Not really, but there are a few times:
Unable to add code/tags
Fast sales cycles
Really long sales cycle
Privacy or sensitive verticals like mental health, 12 steps
Rand Fishkin | Moz and Wil Reynolds | Seer Interactive
These two well-known figures in digital marketing came together to debate about SEO trends and strategies. Here is a recap of what went down:
The first question asked sparked a debate right away. What about SEO that isn’t discussed enough? Wil took a strong stance on UX, which Rand took a stance on high-level strategy. Wil emphasized that the person still needs to make an action. Rand focused on not being like “crack addicts” on SEO tactics and one-off tips, and not implementing a better high-level strategy. Wil rebuts and agrees with Rand, and pulled it back to what the CEO or CMO would want to know, how do we retain clients? You almost have to stop SEO to smell the roses, analyze and figure out how to solve a potential problem. So, if a client typically leaves at the 6-month mark, perhaps studying that and targeting them at that point.
We are moving away from keystrokes and moving towards voice search and other methods in how we search. The gentlemen are asked about the context of searches now. Rand starts talking about conversational search. With Alexa and probably others for sure, the general sense is people wait and respond as it changes. Rand admits to being a little bit hesitant to ever invest in issues like “mobile search is going to change everything”. Yes, be aware, but the doom and gloom that was around for years really irked Rand. The tactical thing to take away is to just keep an eye on things. Wil responds with an emphasis on Mobile and context. Using “strategery” to sort out what is content that will fade away with “OneBox" and find things that will be searched longer. What impresses Wil are people that are “past mobile”. Looking to what’s after mobile. He gave an interesting example of how their wi-fi doesn’t allow you to connect unless you like them on Facebook. He also looked at contextual innovation, such as buttons that allow you to order, sans website.
Is there a possibility that we see these entities basically take over? Like Amazon or nothing? Wil makes people raise their hands if they are Amazon Prime members. It was amazing, like 95% of the audience raised their hands. Wil says to get on the “Amazon Train”. More people went to Amazon to search for a product, vs. Google. Wil adds that Amazon isn't 100% of where he shops. He mentions that sometimes he buys things from different places for different reasons. But, with the absence of a story, Wil is going to Amazon for shopping. Rand is not as fond of Amazon. The most interesting thing to Rand about this, many of the primary young engineers have driven the prostitution rates in Seattle up. He is short term pro-Amazon, long term not. Throughout history, companies seem unstoppable and they get crushed. He sees many many businesses we are the best at one thing, and they all collectively chip away from them. Etsy, E-bay and other entities are already doing this.
Can an SEO specialist nowadays truly have their heads around all of these things? How do they augment all of this vast industry? Rand said it’s a T-shaped paradigm. If you are a web marketer, choose one vertical, but have enough knowledge to know how the other channels function. These other channels are like the top of the T. Rand admits he is not a specialist in every aspect. But others are. So, in short, it is absolutely fine to be a specialist and not try and know everything. Wil adds to that for the first time in 12 years the paid division will generate more revenue than organic SEO. Align your skills where the money flows. Go where it is easy to track your work to revenue. If you are able to do that, client retention can be crazy. In the same report, organic SEO mentioned dollars once, paid search mentioned it 18 times. Subsequently, the client shoved all organic SEO to paid search. Rand rebuts, stating Wil is not wrong, but he says the best among us should not take his advice. Putting more $$ into paid search for more revenue is not a competitive advantage. SEO can drive it much higher and use it as a competitive.
If you are going to be disruptive, embrace the challenge and 80% of the clicks go to organic. He finished with a store about some of the best link builders in the world, left an agency, build their own websites, did some link building on those and worked it for a huge organic SEO bonanza.
Wil finishes and adds that strategy is what is crucial in any SEO situation.
What is the importance of links? Wil says he doesn’t know. He tries to not care. He focused on revenue and user experience. There are a lot of other things to worry about besides links. Trying to explain it is tough. Focus on what's important.
Rand on the original question, links won't demise right away. In the future, yes, big toss up. But SEOs have an obligation to know how this works. We should learn how to teach clients how it works. But we need great answers to this and be able to tell clients and team members. Rand thinks proactive outreach and link earning can work well, some should focus more on content. Content alone doesn't earn links, so get a strategy to do so.
To Wil, outreach is fun when you believe in your content. It’s when you know your content can help people. If you think it is really that good, pay for people to see it.
Do you find it easier to work with companies with their own SEOs? Should all companies get an SEO? Wil's answer is somewhere in between these two questions. SEOs can be critical to keep owners and shareholders in line. For large clients, it's a good idea. Rand states that he works with startups. Every company should have an in-house marketing team. Whether to hire an SEO is if search rankings matter long term. You can work with an agency as well as have an SEO team member.
What do small companies do with handling SEO strategies? Rand chooses to be controversial. Do less, better. Never half ass two things, whole ass one thing. It is okay to “punt” to focus on other things that are important. Wil actually couldn't agree more. It's hard to let something burn. If you do SEO you may have to stop do something else. Fix the biggest problems you have.
How can you prove SEO value? Wil says track revenue to organic measures. Track revenue to content and assets. That makes it an easier sell. You have to learn how to push back when asked the value. How much is a visit worth? How much is a call worth? How much is a page view worth? How much is a newsletter subscriber worth? Rand adds to it: in his experience with SaaS organic clients are typically worth more and retain longer versus paid customers. The nature of that customer is the value. Wil says then you can compare what they’d pay per click versus an organic click.
The "Repandacans" and "Demolytics" parties concluded with a lively discussion about subdomain vs subfolder, which is the best method for getting content out quickly. Basically, they conclude that if you have the ability to, do a subfolder, although sometimes it's quicker to just put it out there on a subdomain.
Get their attention: How to present results to get results
What do you do when someone wants campaign results? Avoid the data black hole. Get their attention.
Why do bad things happen to good data? It has to do with your audience's brain. You have about 8 seconds to retain their attention. With mobile, it's even shorter! Repetition commits information to long-term memory. Shareholders only care about moving the business forward. Maintain attention and be memorable. So make your data stand out.
The 3 pillars of enlightenment:
1. Be your audience
Put ourselves in the audience shoes. What do they really want? Why are we invited to a meeting? That context helps us become their partner, not a data fountain. Call them if you work with them and ask them. Find out what is the big story. Go from statement to insight.
2. Use your tools wisely
We are visual creatures, so don't drown out your audience. Powerpoint can do just that. It came out in 1987. Wowzer. There is a phrase “Death by PowerPoint”. But really, it's not PowerPoint’s fault. It is a people problem.
If you do slides, unload the bullet points. Bullets kill. It exposes all information at once. Why do we love them? Our slides are being used as a crutch so that we don't have to care. Your slides are for your audience. They are not a script. Create ideas, not slides. Break up bullets into slides.
Resist the fluff. Logos, watermarks, and doodads can create unnecessary fluff. Fluff is distracting. Instead, design like a designer. If you don't have any, get some design training. And for Pete's sake, do not use clipart.
Instead, harness the power of real imagery. Vision is the strongest sense. Images create recall and stir emotion. Use it to tell a story. Use the right photo and use the photo right. Match your image to your message. Use the whole slide, don't float your images on part of your slide. Data counts as imagery.
3. Maximize data absorption
Choose the best visualization for the data you are presenting. You can use Chart Chooser. Leapica.com/chartchooser. Pie charts seem to be our favorite, but far too often they are misused. Basically, ditch pie charts. Assorted bar charts can be much more powerful. Of course, use it right. Kill 3d, reduce visual noise data labeling and strategic coloring is highly impactful. Put your pertinent information in a McKinzie Title. “Mckinzie Title” means communicate your message right away at the top.
Remember, you are your PowerPoint, not your presentation!
The conversion equation
What is the future of marketing? What are the common mistakes?
Machines will end all of this mania with automated marketing. Automated megavariate testing of marketing models figures out which version works best on what pages for conversions. That connects to cohort analysis for the highest lifetime performance.
Right now, we are at the conversion equations. Heuristic analysis, rapid experiments conversion, research conversion data and video conversion and engagement data are all conversion equations. The conversion equations show us what to optimize and where to optimize it.
This fast-moving presentation was chock full of a ton of slides and examples. One humorous example showed a landing page that had tons of distractions. By removing all of the clutter besides one button conversions increased. Explain what you do, not how good you are at doing it.
Ditch the hype. Be unique and present yourself at unique. Lose the jargon and there even is a Chrome extension for that. Readability is hugely important too. See how many acronyms you have. Also, motion facility and the typographic golden ratio should be observed when checking your readability.
Visual clarity is important as well. A picture of Larry King should not be what sells a financial institution's services. Oli is killing it. Expectation. Set the expectation for the reader. “See Moz tools” lets the user know that they are going to see Moz’s tools. Crystal clear.
With a design problem, get ADD. Attention Driven Design. Design matters. It creates joy and reduces confusion. It reduces germs when used on new age bathroom toilets in men’s rooms. The first rule of thumb, have a f#@!cking CTA. Have a call to action. Without one, what would the user do? Every website has conversion opportunities, and every conversion opportunity needs a CTA.
Videos can be hugely powerful if done right. By changing words, for example from nothing to click here improved 390%. CTAs and annotations are opposites. In the future, CTAs could have an autocomplete feature for what would work the best! How great.
Social proof is also a big influence in getting someone to complete your CTA. Do others use it? Do you have testimonials? Use video that loads in a lightbox for a testimonial for better results on testimonials. Leveraging social proof will gear a user to make the conversion. Security is also important, you must build trust with your customers. Different audiences trust different identities. Trust seals are important but place them based on the audience that will be on the site.
Just being honest, I’ve been enthralled with Oli’s presentation and typing this is a bit of a chore. The main takeaway- it is really easy to miss the mark, confuse your audience, and lose them. Conversions are so important. There are a lot of things to consider when you do your CTAs. Get out there and do some great work.
Thanks everybody for reading our recap! We hope you enjoyed the MN Search Summit. If you didn't attend this year, maybe you should consider going next year!