Hitchhiker's Guide to Serverless JavaScript

steven-faulkner-nodejs.png

Although Serverless has become a bit of a buzzword recently, Steven Faulkner, Director of Platform Engineering at Bustle, disagrees that it is just a fad.

The Linux Foundation

Although Serverless has become a bit of a buzzword recently, Steven Faulkner, Director of Platform Engineering at Bustle, disagrees that it is just a fad. In his talk at Node.js Interactive, he points out that much of the faults that people find in Serverless -- that it doesn't scale, that it is not production ready, that it’s too slow and expensive -- are provenly false.

Faulkner has made Serverless the backbone for Bustle's content since it became a thing a couple of years ago. This means he is in charge of all things back-end, but does a lot of front-end work, too. At Bustle, one of the largest women’s media properties online, they use serverless for 99 percent of the requests. This translates to between 10 and 20 million calls per day. Serverless has a latency of less than 200ms average, and it costs less than than their EC2 (Amazon Elastic Compute Cloud) base setup.

Serverless 101

Faulkner also addresses some of the misconceptions surrounding Serverless. The first is, admittedly, down to the name of the concept itself. Serverless is a misnomer because, clearly, there are still servers deep down somewhere.

However, Serverless is the result of answering the question "What if we abstracted all the servers away?" -- that is, what happens when you create a service that relieves the developers and admins from having to worry about the underlying server? Serverless implements "anything and everything as a Service," which is not a new concept.

Faulkner says Serverless is really about Functions as a Service (FaaS). The core tenants of FaaS are that code is only run when needed, you only pay when code is run, and you are not bothered with the details, in fact, your app is a function that sits between the request from and the response to the user:

request
yourApp()
response

There are several providers offering Serverless-Function as a Service. For example, Microsoft Azure has Azure Functions and Google has Cloud functions. IBM has an open source service called OpenWhisk. Faulkner, however, Faulkner thinks Amazon Web Services is ahead of everyone else in this arena and described some examples in his talk.

Why Use Serverless

Faulkner says there are several advantages to deploying web apps to a Serverless service. Despite what many think, Serverless does not mean NoOps; it means Less Ops. Serverless also scales, in the sense that fewer things change between when you have 1 request per second, and when you have 1000 requests per second.

Price is another factor. You only pay when somebody is using your code. This means you can maintain legacy services that you are not sure you should turn off quite yet, or you can deploy services you are not sure will be successful at zero cost.

But even when your code is running, it is still cheaper. Faulkner says that an API he uses to trail all the user tracking and event tracking on the Bustle website and then route it where it's meant to go costs $2500 month on EC2. On Lambda/API Gateway only $400 month.

For developers, it is easier to iterate at function-based level rather than on your whole application, and Serverless allows for single function deployments. This means you can do things very quickly by re-writing one single function and you can deploy it without affecting the whole. Amazon's service also makes for quicker deployments because you can deploy 50 functions and alias them all at once.

Why Not Use Serverless

There are, however, also several reasons why you may not want to use a Serverless service like Amazon's. There is, for example, the risk of becoming locked in. Migrating from a Serverless FaaS can be a minefield, but, according to Faulkner, from Amazon it is not that hard. He has a chunk of code 9 lines, an Express server, that calls the Lambda functions if you have to migrate.

Another problem is running cold functions. If you're doing something outside the function, like loading a bunch data before the first time your function runs, this can delay the first run an unacceptably long time. Amazon solves the problem by running all your functions within 3.5 hours of the upload.

Testing is yet another area that can cause problems. Testing individual functions is not too difficult according to Faulkner. But when you want to do integration testing across all of your functions, that is still painful.

Serverless at Bustle

At Bustle, the front-end serves up HTML and JavaScript through the API gateway. The application reads some static assets through S3 and uses Lambda for SSR (Server Side Rendering) to generate the actual pages.

On the back-end, Bustle uses GraphQL to fetch most of the data. The JSON generated by GraphQL is processed through API Gateway and Lambda, which are talking to dynamodb or redis, with some PostgreSQL thrown into the mix.

Tools

Despite all its advantages, if you are deploying more than two functions, it is very painful to deploy your stuff manually. That's where the tooling comes in. Faulkner mentions Serverless Framework, node-lambda, apex, claudia.js as tools to check out. But, he also recommends his own open source tool, shep, which will allow you to build and deploy your serverless applications to Amazon's Lambda with a few simple commands.

For more examples and details, watch the complete presentation below:

If you are interested in speaking or attending Node.js Interactive North America 2017 - happening in Vancouver, Canada next fall - please subscribe to the Node.js community newsletter to keep abreast with dates and time.

Why Engagement Might Be The Most Important Metric For Musicians

Why Engagement Might Be The Most Important Metric For Artists

DEC 19 2014

Whenever the topic of social media is raised in any forum of the music industry, a choice discussion is always the practice of purchasing followers or views, allowing artists to game charts, rendering social activity unreliable and therefore less significant.

But often overlooked in this dismissal of the validity of social activity is the fact that building a significant social following is fairly pointless for an artist unless they successfully leverage that following at critical moments. To reach fans with a new lyric video, to spread the word about a tour, to gauge interest in an upcoming album release – ultimately to interact with fans and get instant feedback on what you are creating.

While a vast social following can be a valuable tool, only legitimate and loyal followers that engage with the content you provide through social channels, and ultimately your work, provide you with the reach and engagement you are seeking as an artist. For this very reason, it is important to look beyond totals, and start to measure the relationship between daily activity and the number of fans you have – a measure of engagement.

Artist manager Robbie Lackritz has been working with Leslie Feist – known only as Feist – for close to a decade, and has a strong sense of how to build valuable engagement with fans. Feist, who is a Canadian singer-songwriter with more than four solo studio albums under her belt as well as countless collaborations with other artists and bands, has maintained a successful career as an artist for many years now.

Her most commercially successful album came in 2007. The Reminder (the making of which is chronicled in her 2010 documentary Look At What The Light Did Now) earned her four nominations for Grammy awards in 2010, including Best New Artist. About her latest album Metals and their release strategy, Lackritz says “it was a lot more about finding a more engaged core of followers.”

With just shy of a million page likes on Facebook, 127,000 followers on Twitter and more than 20 million video views on Vevo, Feist is classified as a Mainstream musician. But for the artist and her management team, building the largest online following, and simply hocking social content for attention, has never been the goal.

While they see the value in cultivating a social following, and Lackritz regularly employs analytics to understand how fans are reacting online, the artist posts fairly infrequently online. Her Instagram account totals only 15 snaps and two videos, she rarely puts out more than one piece of content per week on Facebook, and her last tweet was in November.

For each individual post Feist has in her Facebook feed, there are hundreds, if not thousands of interactions from followers. But given that she is nearing the end of her current album cycle, and the relative infrequency of social activity, her audience engagement is defined as Moderate.

Next Big Sound identifies four levels of engagement: Occasional, Moderate, Strong, and Passionate. Audience engagement is defined by the relationship between daily activity and the overall size of an artist’s following, and then performance is measured against that of every artist in the database. Artists that are defined as Moderate, such as Feist, range from the 16th to the 84th percentile.

Engagement is not contingent on having a massive following. For instance, Taylor Swift’s audience engagement is also defined as Moderate, despite the fact that she has more than 77 million Facebook page likes, close to 50 million Twitter followers, and has just released her latest album to great fanfare. This measure is valuable for artists at any stage, to help them understand how they are reaching fans, and what content is reacting.

But be weary of engagement for the sake of engagement. “If we’re posting photos of cute cats, people are going to engage,” he says, “But when it comes down to the marketing of Feist, and what we put on her pages and why, we really try to center around where her head is at, and what she is working on.” The fans who follow and interact with Feist on social, attend her shows, enjoy the art of her album, and will help her sustain a career for many years to come.

“Her types of social media campaigns are far less active in volume of posts, but try to be much more curatorial about content,” Lackritz explains. “That lends itself to the type of career that she aspires to have – which is not about being the most gigantic artist on the face of the earth, but about having fans that really appreciate the deep cuts on the record, as much as they do the songs that radio might consider playing.”

“As nice as it is to have these social mediums of marketing to be able to engage people,” Lackritz says, “Don’t compromise the end result of why you do what you do, and what’s special about what you do.”

This article originally appeared on Forbes.com. Liv Buli is the resident data journalist for music analytics company Next Big Sound. She takes a data-driven perspective to covering the ins and outs of the music industry. Buli is a graduate of New York University’s Arthur L. Carter Journalism Institute and her work has appeared in Newsweek Daily Beast, Forbes, Billboard, Hypebot and more.

http://blog.nextbigsound.com/post/105623023398/why-engagement-might-be-the-most-important-metric

Forrester: Mobile Is The Anti-Channel

• by Sarah Mahoney, January 2, 2015, 8:42 AM
• Recommend (8)
• Comment

While most marketers continue to think of mobile as a channel — one that runs somewhat parallel to its Internet presence — a new report from Forrester Research says it’s time to reverse that perspective. “Mobile eliminates the notion of channels by blurring the distinction between the physical and digital worlds,” writes analyst Julie A. Ask. And brands like USAA, using mobile as a way to sell expanded services; Starbucks, which is continually finding new ways to streamline purchasing, or Guinness building passionate beer communities, are finding that “mobile generates new revenue, improves customer relationships and reduces costs throughout all channels.”

But few companies are prepared. Forrester estimates that less than 4% of businesses are ready to take advantage of the opportunity presented by mobile, and even then, fail to take into account its rapidly expanding definition. It’s not just phones, she says, but watches, cuffs, glasses and even lighting systems. 

“We still see most of our clients cramming PC experiences onto phones when they should be thinking in terms of mobile moments instead,” she writes. Among the executives it surveyed, 47% planned to spend $500,000 or less externally to build up their mobile services. “This budget barely pays for the analytics, messaging, testing, and marketing solutions on which most enterprises depend.” And only 45% have at least five developers in-house.

The best strategies, she writes, are those that find ways to interact with consumers at specific stages in the six-step customer journey:

*Discover Getting consumers who are not yet customers to engage isn’t easy, but financial brands, such as Intuit or USAA, “manufacture mobile moments by offering consumers tools to manage their finances, cars, and homes — and opportunities to buy new services in those moments.”

*Explore With 43% of American smartphone owners saying they’ve researched a product via their device in the last few months, savvy marketers are finding ways to make mobile research easier, driving total transactions, not just digital sales.

*Buy In 2014, U.S. consumers bought about $24 billion in goods via mobile and used their devices to pay for $52 billion, “whether they are in-store with a mobile-enabled associate, paying remotely, paying bills, or engaging in peer-to-peer payment methods.”

Giving customers mobile access to both pricing and inventory increases their shopping confidence.

*Use Increasingly, consumers expect support through apps, updates and notifications. Basis, for example — a fitness and wellness wearable that gathers information on steps taken, sleep, and heart rate — comes with an app which walks consumers through behavior changes. 

*Ask Mobile can be a low-cost self-service tool for customer support, and “is at its best when it streamlines customer interactions, saving them time and saving you money.” One example: American Express uses [24]7 to resolve potential fraud through an interactive mobile phone experience, instead of a regular voice call. In addition to saving millions in operation expenses, “nine out of 10 consumers gave the experience at least four out of five stars.”

*Engage Sharing reviews, comments, and tips allow customers to create their own communities, fostering brand loyalty and customer passions, like the Guinness Pub Finder app. Mobile extends those communities by making it easier to post photos and videos. Marketers benefit “from crowd-sourced content and a brand affinity that paid media can’t buy.”

http://www.mediapost.com/publications/article/240998/forrester-mobile-is-the-anti-channel.html

How to choose the best predictive lead scoring vendor for your company

Generating a bounty of leads doesn't mean your work is done. What's much more important is whether those leads will ultimately convert—and predictive lead scoring is one way B2B marketers can determine if a lead is good or if it belongs on the trash heap.

Predictive intelligence platforms can help B2B marketers prioritize existing prospects or identify unknown prospects through lead scoring. Basically, the technique uses data from both CRM and marketing automation solutions and unites it with data from third-party sources to detect a usable pattern. Lead scoring can boost sales productivity, improve marketing effectiveness and encourage sales/marketing alignment.

"Predictive lead scoring offers a more accurate outcome that B2B marketers and sales reps can rely on and gives you a competitive edge," said Jill Stanek, research analyst for the technology practice at B2B research firm SiriusDecisions.

Predictive lead scoring technology is in its infancy—just a handful of companies offers tools that can be used in the B2B market and the standards for crunching that data are not consistent across the field of providers. But there is a hunger for this capability among B2B marketers. According to SiriusDecisions, there are almost 14 times more B2B organizations with predictive lead scoring today than there were in 2011.

"Using predictive analytics to model the propensity of prospects to buy has been something that's been an important part of B2B marketing for many years now, but we simply didn't have enough data in B2B in prospects and in sales results to take advantage of this [lead scoring] technology," said Kerry Cunningham, research director of demand creation strategies at SiriusDecisions. "But now that we have so much data in B2B, and it's growing all the time, I think the adoption cycle is going to be pretty rapid."

FIRST STEPS

With so few companies offering predictive scoring capabilities, how can B2B marketers go about choosing the right one?

One of the first things to do, Cunningham suggested, is to run trials with several vendors. Have those companies provide you with a set of set of scored leads to test. "There are a couple of ways to do a trial," he said. "One is actually to run a small pilot and test those leads in the field. If you don't have the resources to actually send your salespeople out to try to sell those leads, then at least have salespeople look at them to assess whether or not the leads that this model is saying are good make sense and are leads that they would also agree are good."

Kerry Cunningham

Stanek said the next critical thing to consider is the vendor's design and deployment of a scoring system that has been specifically tailored to your needs. "Some of them actually can do it within twenty-four hours and for some it can take up to four weeks." She also said to make sure to ask if a lead scoring vendor actually sources net new contacts; some vendors only score existing contacts. Securing customer references is also a must.

Another important factor for B2B marketers looking to hire a predictive lead scoring vendor is how well that vendor is able to support delivering the leads in a format and in a way that it's going to work well for the ultimate lead recipient. For example, since predictive lead scoring vendors are ultimately going to be providing sales leads that go to either teleprospectors or salespeople, it's important to understand how the scores will be presented to those folks. What's the format? How much info will they get explaining to them why it's a good lead?

"There's not yet a right answer for what's the best way to present it back or what are the best pieces of information to present back to those teleprospectors or salespeople," Cunningham said, "but for each organization there will be some ways of doing it that are better than others. It's going to be pretty individual to the company, but it's extremely important, maybe more important than how good the model is itself."

To help its clients properly evaluate the lead scoring vendor marketplace, SiriusDecisions recently released a guide detailing the pros and cons of several vendors. The report focused on seven vendors—6Sense, Fliptop, Infer, Lattice, Leadspace, Mintigo and Salesfusion.

Jill Stanek

According to the report, 6Sense may be the platform of choice for high-tech, manufacturing and other B2B enterprises focused on identifying new leads in the early stages of the buying cycle; the company tracks content consumption across thousands of B2B publisher sites, blogs, communities, forums and business directories to help it source leads. In contrast, Fliptop aims its platform at SMBs and larger organizations that use Salesforce.com and examines a company's existing database to identify leads. Infer may also be a good fit for rapidly growing SMBs that require a lead scoring platform that can be easily hooked up to a database of existing prospects.

For B2B marketers with strict security requirements that need multiple scoring models across different business units, or that are focused on account-based marketing, Lattice should be considered. For those organizations, large or small, that leverage multiple buyer personas, have multiple products and need new leads outside an existing database, Leadspace should be on the short list.

If data hygiene is not the best, Mintigo provides data remediation during onboarding; the company also offers insights to marketers interested in using lead scoring to drive content delivery. However, Mintigo requires more work than other similar vendors during initial setup. And, finally, Salesfusion is an affordable option for SMBs that do not currently have a marketing automation platform and are looking for standard marketing automation functionality and predictive lead scoring based on their own data.

Both Stanek and Cunningham emphasized that their recommendations may change due to a vendor landscape that is in flux. For example, since this report was released a few weeks ago, a couple of new vendors are now on the research firm's radar. But change is now here, and both analysts predicted that more and more B2B companies will use predictive lead scoring as new vendors pop up and older ones are acquired by large marketing automation vendors and integrated into their solutions.

"SiriusDecisions estimates that fewer than five hundred B2B organizations are actually using [predictive lead scoring] technology today," Stanek said. "And as a developing market that's rapidly growing, the reality is that not that many companies are actually implementing it. But we definitely estimate significant growth because most organizations that buy predictive lead scoring platforms already have a marketing automation platform in place."

http://www.fiercecmo.com/special-reports/how-choose-best-predictive-lead-scoring-vendor-your-company

What’s Missing from the Industrial Internet of Things Conversation? Software

• BY MATT CICCIARI, MERIDIUM  
 
• 11.20.14  |  
 
• 3:14 PM  |  
 
• PERMALINK

• Share on Facebook

  97
 
 
 
• inShare196
 
• 1

 brunkfordbraun/Flickr

These days, you can hardly have a technology conversation without talking about the Internet of Things (IoT). And when that conversation shifts its focus to the industrial sector, including energy, Oil & Gas, Power & Utilities, and petrochemicals, among others, the discussion changes to what is being called the “Industrial Internet of Things” (IIoT).

So what is all this hype about? The convergence of cheap processing, unending storage, massive bandwidth, near-ubiquitous connectivity, and cloud-based applications is driving new capabilities for gathering information and changing the way we interact with machines and services. The data generated by sensors in this network of connected devices is being collected and analyzed, spawning the growth of big data analytics and applications. And the resulting analytics are being used to improve business efficiency, better serve customers and disrupt old business models.

There will be nearly 26 billion connected devices by 2020, according to Gartner, which is an increase of 30 fold over 2009. Impressive, but what does it really mean to have all these devices connected to the internet? Is there any real value involved? Just because you can, does it mean that you should? We all remember the technology boom and bust of the late 1990s and early 2000s (aka Dot-com Bubble) that was fueled by an excessive amount of wishful thinking or “irrational exuberance.”

But the IIoT can and will be different – if we keep our collective eyes on the prize. Unlike consumer-based IoT that is trying to devise a way to make your world a better place by telling you when your washing machine needs service or letting you control all of your home’s systems while you are away, the IIoT is working to make our collective world a better place by improving the monitoring, control and safety of everything around us. In short, reducing risk and improving the reliability of our massive industrial systems. No small task given the amount of data that will be generated by all of those devices running around the clock.

But there is an important component missing from many conversations that will glue it all together: software.

Think about it: How good is your computer without a good operating system to make all those programs function properly, or your smartphone without a million apps to connect you and your friends or coworkers and the information you share? Large hardware vendors recognized long ago that it is the software that makes the hardware more valuable. Software supports the human interaction side of the equation by helping to turn data into information. But not just basic information, I am talking about critical, insightful, influential, and actionable information, without which we run the risk of potentially dangerous outcomes.

This next-generation software should be cloud-based because, let’s face it – on-premises solutions require significantly more time, money and resources than you would like to afford while trying to remain competitive. Advanced software should easily connect the existing enterprise asset management (EAM) or computerized maintenance management system (CMMS) with the field-based industrial devices and equipment. It should provide early, rapid insight into the volumes of data and information that mere mortals are expected to evaluate to find business opportunities and make recommendations as part of a larger asset performance management strategy. But more importantly, software should make it easier for your organization to achieve its highest objective: operational excellence.

Even in today’s increasing connected environment, industrial organizations do little more than make use of a minimal set of vendor-supplied analytics related to machine-specific solutions – automated instrument calibration, instant vibration analysis of pumps and compressors, or other equipment related algorithms. The analytical capabilities of affordable processing and storage capacity in new software solutions are just beginning to be tapped.

One key to discussing the software needs of the IIoT is to include asset performance diagnostics. More specifically, to take advantage of all the useful information out there and the connectivity promised by equipment and software vendors, we need to discuss asset performance diagnostics; the ability to quickly assess a plant’s equipment/asset status or ‘health’. This ability can be critical for merger and acquisition activity versus building new. This implies that we must be able to check across many equipment types, not just one or two manufacturer-specific reports, in order to understand where our metrics stand in relation to our industry peers.

Additionally, with next-generation software, organizations will gain enhanced asset reliability and comparative analysis for operational goals. Operators can perform better and faster analysis on collected data to report to the C-Suite executives, who then have better context for defining strategic goals. Comparative analytics provide the ability to internally compare the status and performance of the equipment or assets to “get your own house in order” and really ignite your continuous improvement initiatives. Next-generation software must help organizations “take it to the next level.” To get there, you need to benchmark your operations against your peers and competitors. That capability will help you move to the pinnacle of continuous improvement, and closer to overall operational excellence.

There is an enormous amount of data generated today and as new sensors are introduced in global industrial facilities, big data will become even bigger. Organizations still struggle with how to leverage that data operationally and strategically to better manage reliability strategy, operational risk and facility operations, which is why we need to insert software into our discussion of the IIoT.

Matt Cicciari is Associate Director of Solution Marketing at Meridium.

http://www.wired.com/2014/11/industrial-internet-of-things-software/

More Panda 4.0 Findings: Syndication, User Engagement, Indexation & Keyword Hoarding

Glenn Gabe, June 4, 201432 Comments

•SEO Evolution: Sell, Discover, Deliver & Report on Highly Converting Keywords by Krista LaRiviere, gShift

Now that the dust has settled, it's clear that Panda 4.0 was a powerful update that impacted many sites across the web. And that impact has been fascinating to analyze.

Once I heard the official announcement from Google's Matt Cutts, I began to dig into sites that had been impacted by Panda 4.0. That included clients that saw recovery and new companies reaching out to me with fresh hits. During the first week, I analyzed 27 websites hit by our new bamboo-eating friend.

I posted my initial Panda 4.0 findings on my blog, based on heavily analyzing recoveries and new Panda victims. During my research, it was incredible to see the highs and lows of Panda as I jumped from recovery to damage.

From a recovery standpoint, I had some clients absolutely surge after making a lot of changes (which I call "the nuclear option"). In addition, I saw industry experts rise in the rankings, I noticed a softer Panda, and I saw several Phantom victims recover. Again, you can read through my initial findings to learn more.

And on the flip side, I unfortunately saw many companies get pummeled. And when I say pummeled, I'm referring to losing more than 60 percent of Google organic traffic overnight. Some websites I analyzed lost closer to 75 percent. You can see two very different situations below.

Panda 4.0 recovery:

Panda 4.0 fresh hit:

A Quick Note About the Rollout of Panda 4.0

Panda 4.0 was officially announced on May 20, but I saw websites begin getting impacted the weekend prior (Saturday, May 17). So, it looks like Google began pre-rolling Panda out a few days prior to the announcement (which I have seen in the past with major algorithm updates).

But I also saw another interesting trend. I started seeing more volatility the following weekend (beginning Saturday, May 24). And then more jumps and decreases on Monday, May 26 and Tuesday, May 27.

Others saw this too, and several of us thought it could be the start of a Penguin update (or some type of major update). But Google denied Penguin (or some other spam effort) was rolling out. I had a hard time believing that, based on the additional volatility I was seeing.

Since we've been told that Panda can take 10 days to fully roll out, I was wondering if the Panda tremors I noticed a week into Panda was the tail end of Panda 4.0. I asked Cutts via Twitter, who responded yesterday, saying that Panda 4.0 actually rolled out a bit faster than usual (and that it didn’t take 10 days to fully roll out).

Wow, now that we know it wasn’t a 10-day rollout causing secondary volatility, what in the world was it? Maybe it was Google tweaking the algorithm based on initial findings and then re-rolling it out. Or maybe there was some type of secondary update, separate from Panda. It’s hard to say, but the data reveals that something was going on.

From an analysis standpoint, I saw websites that surged increase even more the following week, and I saw websites that took a hit decrease even more. I didn't see mixed trending (up then down or down then up). That's another reason I thought the volatility early last week was Panda 4.0 still rolling out (or getting re-rolled out).

Here's a quick example of the secondary surge:

More Panda 4.0 Findings

In my first post about Panda 4.0, I mentioned that it wouldn't be my last about the subject (not even close). I had a lot of information to share based on my research, and from across industries. In addition, I've now analyzed 14 more websites hit by Panda 4.0 so I have additional findings to share.

So, if you are interested in Panda, if you have been impacted by Panda 4.0, or if you are just an algo junky, then sit back and grab a cup of coffee. This post will provide several more findings based on analyzing fresh hits and more recoveries from the latest Panda update. Let's dig in.

Syndication Problems

There were a number of companies that reached out to me with major Panda hits (losing greater than 60 percent of their Google organic traffic overnight). Upon digging into several of those sites, I noticed a pretty serious syndication problem.

Those websites had been consuming syndicated content at various levels, and much of that content got hammered during Panda 4.0. The percentage of syndicated content residing on the websites ranged from 20 to 40 percent.

Each site that experienced a serious drop was in a different niche, so this wasn't a specific category that was targeted (like I have seen in the past). After analyzing the sites, it was hard to overlook syndication as a factor that led to their specific Panda hits.

When reviewing the syndicated content on the sites in question, none of them had the optimal technical setup. For example, using rel=canonical to point to the original content on another domain (via the cross-domain canonical tag). Some content had links back to the original content on third-party websites, while other pieces of content did not.

Also, the content could be freely indexed versus noindexing the content via the meta robots tag. That's always an option and I explain more about managing indexation soon.

From a Panda standpoint, I've never been a fan of consuming a lot of syndicated content. If it can be handled properly from a technical SEO standpoint, then some syndicated content is fine. But knowing that Panda can target duplicate content, scraped content, tech problems that create quality problems, etc., I've always felt that webmasters should focus on providing high-quality, original content. That's obviously the strongest way to proceed SEO-wise.

I'm not saying to nuke all syndicated content, but you need to be very careful with how it's handled. In addition, you need to always manage how much syndicated content you have on your site (as compared to original content).

Side Note: Indexation Numbers and Syndicated Content

By the way, I have a client that called me a year ago after previously being impacted by both Phantom and Panda. They also had been consuming syndicated content. Actually, they were consuming a lot of it (greater than 200,000 pages). It's a large site, but that's still a significant amount of third party content.

We worked hard to identify the content across the site, how attribution was being handled, and then drilled into the performance of that content SEO-wise. Then we had to make some hard decisions.

We decided that nuking risky content now would help them win long-term. Their indexation has dropped by approximately 70 percent over the past year. And guess what? The site is performing better now in Google than when they had all of that additional content indexed.

That's right, 70 percent less pages indexed yielded more Google organic traffic. It's a great example of making sure Google has the right content indexed versus just any content. And since Panda 4.0, Google organic traffic is up 140 percent.

More Syndication Problems – Downstream Considerations

During my analysis, I found some of the companies that were consuming syndicated content were also syndicating their own content to various partners. And much of that content wasn't handled properly either (attribution-wise).

To make matters worse, some of the partners were more powerful SEO-wise than the sites syndicating the content. That led to some of the partners outranking the websites that created the original content. Not good, and now that Panda 4.0 has struck, it's obviously even worse.

If you're going to syndicate your content, make sure the websites consuming your content handle attribution properly. In a perfect world, they would use rel=canonical to point back to your content. It's easy to set up and can help you avoid an attribution problem.

In addition, the sites republishing your content could simply noindex it and keep it out of Google's index. They can promote and highlight the content on their site (so it can still be valuable for users), but it won't be indexed and found via search. In aggregate, handling syndicated content properly can help everyone involved avoid potential SEO problems.

Even More Problems – Copying/Scraping Content

It's worth noting that during my analysis, I also found original content that had sections copied from third-party content. In this case, the body of the page wasn't copied in its entirety. There were just pieces of the content that were copied word for word. I saw this across several sites experiencing a steep drop in traffic due to Panda 4.0.

The solution here is simple. Write original content, and don't copy or duplicate content from third parties. It was clear during my analysis that content with sections lifted from third-party content experienced serious drops in traffic after Panda 4.0 struck.

Strong Engagement Tames the Savage Panda

Let's shift gears for a minute and talk about user engagement and Panda. If you have read my posts about Panda over the years, then you know that user engagement matters. Strong engagement can keep the Panda at bay, while weak engagement can invite the Panda to a bamboo feast.

It's one of the reasons I highly recommend implementing adjusted bounce rate (ABR) via Google Analytics. ABR takes time on page into account and can give you a much stronger view of engagement. Standard bounce rate does not, and is a flawed metric.

High standard bounce rates may not be a problem at all, and I know that confuses many webmasters. By the way, if you are using Google Tag Manager, then follow my tutorial for implementing adjusted bounce rate via GTM. If you implement ABR today, you'll have fresh data to analyze tomorrow.

Well, once again I saw poor user engagement (and bad user experience) get sites in trouble Panda-wise. As more and more companies reached out to me about fresh Panda 4.0 hits, I could clearly identify serious engagement issues once I dug in.

For example, pages that were ranking well in Google prior to Panda 4.0 had thin content, horrible usability, affiliate jumps so fast it would make your head spin, downstream links that were risky, stimulus overload, etc. There are too many examples to cover in detail here, but if you've been hit by Panda, take a hard look at your top content prior to Panda arriving. You just might find some glaring issues that can lead you down the right path.

One site in particular that I recently analyzed was driving a lot of Google organic traffic to pages that didn't provide a lot of solid information based on the query. The user interface was weak, the content (which was thin) was located down the page, and if you did click on anything on that page, you were taken to an order form.

I like to call that the Panda recipe of death:

• 1 part confusion
• 2 parts thin content
• 1.5 parts horrible usability
• 2.5 parts deception
• And you might as well throw in a bamboo garnish.

This recipe is guaranteed to drive low dwell time and a nasty hangover. There's nothing that attracts Pandas like frustrated users and low dwell time. Avoid this scenario like the plague.

Keyword Hoarding – Should Your Website Be Ranking For That Many Keywords?

While analyzing more fresh hits, I came across a few situations where smaller websites were ranking for thousands of keywords in a competitive niche without having enough content to warrant those rankings (in my opinion).

For example, one site had less than 40 pages indexed, yet ranked for thousands of competitive keywords. In addition, the homepage was ranking for most of those keywords. So you had essentially one page ranking for thousands of keywords, when there were clearly other pages on the web with more targeted content (that could better match the query).

Once Panda 4.0 rolled out, that site's traffic was cut in half. The content on the site is definitely high quality, but there's no way the homepage should have ranked for all of those keywords, especially when there were articles and posts on other websites specifically targeting torso and longer-tail terms. The site had been teed up to get smoked by Panda. And it finally did.

I saw this situation several times during my research, and to me, it makes a lot of sense. Google is going to want to match users with content that closely matches their query.

If a user searches for something specific, they should get specific answers in return (and not the homepage of a site that targets the category). I'm assuming that had to be frustrating for users, which could impact bounce rate, dwell time, etc.

Again, low engagement is an invitation to the mighty Panda. Unfortunately, the site only had standard bounce rate set up to review and not adjusted bounce rate. See my points earlier about the power of ABR.

The Answer: Develop a Stronger Content Generation Plan

Since Panda 4.0, the site I analyzed is naturally getting pushed down by third-party content targeting specific queries. My recommendation to anyone who is experiencing this situation is to expand and strengthen your content strategy. Write high quality content targeting more than just head terms. Understand your niche, what people are searching for, and provide killer content that deserves to rank highly.

Don't rely on one page (like a homepage) to rank for all of your target keywords. That's a dangerous road to travel. High quality, thorough content about your category can naturally target thousands of keywords.

Recommendations and Next Steps

The scenarios listed above are just a few more examples of what I've seen while analyzing websites impacted by Panda 4.0. So what can you do now that Panda 4.0 has rolled out? There are some recommendations below.

If you've been impacted negatively by Panda 4.0, then you need to move quickly to analyze and rectify content quality problems. And if you have been positively impacted, don't sit back and assume that can last forever. I've seen many sites drop after once gaining traffic due to Panda updates.

Therefore, I have provided two sets of bullets below. One set is for those negatively impacted, while the other is for those experiencing a surge in traffic.

If you have been negatively impacted:

• Have a Panda audit completed. Hunt down low-quality content, technical problems impacting content quality, engagement and usability problems, etc. I've always said that Panda should have been named "Octopus" since it has many tentacles. Thoroughly analyze your site to find major problems.
• Move quickly to make changes. Don't sit back and over-analyze the situation. Find low quality content, understand the best way to proceed, and execute. Rapidly executing the right changes can lead to a faster recovery. Delaying action will only keep the Panda filter in place.
• Keep driving forward and act like you aren't impacted by Panda. Keep producing high quality content, keep using social to get the word out, and keep driving strong referral traffic. The work you do now while impacted by Panda will help you on several levels. And the content you produce while being impacted could very well end up ranking highly once the Panda filter is lifted.
• Since Panda rolls out monthly, you technically have a chance of recovery once per month. You might not see full recovery in one shot, but the more changes you implement, the more recovery you can see. It's another reason to move as quickly as you can.

If you have been positively impacted:

• If you've taken action to recover from Panda, then you were obviously on the right track. Keep driving forward with that plan. Don't stop just because you recovered.
• If you haven't taken action to recover, and you have seen an increase in Google traffic after Panda 4.0 rolled out, then don't assume that bump will remain. You should still understand the content quality risks you have in order to rectify them now (while you have stronger traffic). I've seen sites yo-yo with subsequent Panda updates. Band-aids do not yield long-term Panda recoveries. A strong SEO strategy will.
• After seeing a recovery, review your top landing pages from Google organic prior to Panda being rolled out, look at user engagement, and analyze that content. You might find a strong recipe for "high quality" content based on your audience. Then you can use that as a model for content creation moving forward.
• Review the keywords and keyword categories driving traffic to pages now that Panda 4.0 has rolled out. You might find interesting trends with how Google is matching queries with content on your site.
• Continually crawl and analyze your content to pick up quality problems and technical problems. I do this for some of my clients on a monthly basis. And there are always findings that need to be addressed (especially for larger websites). You would be surprised what you can find. For example, I just found 100,000 pages that were once blocked and now can be crawled. That can be fixed quickly if picked up during a crawl analysis. Technical SEO checks should be ongoing.

Summary – Panda 4.0 Was Significant

Again, Panda 4.0 was a major algorithm update. There were many examples of sites recovering, while also many examples of sites getting crushed.

I plan to keep analyzing websites on both sides of the equation to better understand Panda 4.0 and its signature. And I'll be writing more posts covering my findings in the coming weeks.

Until then, good luck with your Panda work. Remember, the next update is only a few weeks away.

http://searchenginewatch.com/article/2348110/More-Panda-4.0-Findings-Syndication-User-Engagement-Indexation-Keyword-Hoarding