Archive for Algorithms

James Carson | April 17, 2012
No comments

Why I Believe Debating Authors is Totally Moot (Like Most SEO Debates)

So my last post focused on my Brighton SEO presentation, I Believe Authors are the Future, and I want to give my opinion potential debates that will likely end in a load of hot air, wasted energy, and people taking sides about how important it is. I think we’ve wasted enough time talking about inbound already!

The reason I’m writing this is to counter posts such as Why I don’t believe authors are the future (which I actually rather enjoyed as an alternative point of view) and variants on this theme. I think it’s fair enough to have an opinion about the subject, but here’s why debating online authors is a moot point, and something we should simply forget about:

Online Identity is Serious

Online identity is a serious topic and social services are becoming identity cards. More and more stuff is going to be favoured if it is properly identifiable. This is clearly expressed in Eric Schimdt’s views. My web experienced is greatly enhanced through online identification such as Facebook Connect. Not everyone thinks that way (and those people normally have big voices in the Twitter loving media) but over 1 billion people are on networks, and that number is growing. I don’t really believe people care as much as the media suggest about web privacy. Most people are totally passive about it. For an interesting counter view to online identity, here’s a great podcast with 4chan founder Chris Poole.

There’s Nothing Wrong with Easy

Authorship is so easy to set up. Who cares if it doesn’t have a major effect right away? You’re going to spend minimal development time (particularly on WordPress) setting it up.

Toilet Author Rank

It Already Exists

Authorship is a small enhancement to what should already exist on sites – author profiles. A lot of news sites have them without markup simply so readers can build a rapport with who is writing stuff. Adding social elements to these profiles can only build up social rapport. Also, Google patented the system 5 years ago…

Complete Overhype!

It’s getting ludicrously overhyped. I know we like to talk about slight changes in SEO as if the Russians are marching through Berlin, but the reality is the algorithmic changes of this simple markup are a long way off. I would consider 18 months for a noticeable effect – I could be wrong, but it’s taken 6 for Google to even recognise the markup – not to mention it’s changed the goalposts FIVE times! Adoption rates are not high, and as I considered in my second ever post, the rollout won’t be that easy. For starters, currently it assumes that all authors sign up to Google+, which is nonsensical. If this doesn’t change, it will take absolutely ages for it to have any effect – because adoption rates of Google+ aren’t that big compared to other services (20% of Facebook usage at best). It would take ages for Google+ to usurp Facebook. If it ever does.

It Only Effects Authored Pages… Or Does It?

People are inevitably (as Why I don’t believe authors are the future does) going to say that authorship won’t effect homepages, product pages etc. Okay – if you’re a tricycle selling website, it would be ludicrous to have your product pages to be written up by a prize winning journalist so your site can rank. Not to mention the whole scheme makes no sense – is the technology section of the Argos catalogue written by Charles Arthur? Nope. However, it’s not always about ‘who’ has written the content, rather who has endorsed that content. If you sold trikes and your product pages were endorsed by a journalist from askamum, that would probably have an effect.

Google is a Machine

Finally – Google is a machine, created by people. We are not quite at the days of Skynet, or iRobot or anything else. If you couldn’t beat a machine through some sort of loophole, I think we’d be in a weird place. Yes, you will inevitably be able to game author rank, I just don’t happen to think focusing on gaming things is ever a long term strategy.


My basic suggestion is: Set it up as standard and get on with it.

Now then, back to this I think:

Category: Algorithms, Authorship, SEO |
James Carson | September 19, 2011
No comments

Content for Links? What Ever Happened to Content for an Audience?

When it comes to content, I’ve got to say SEO rarely gets it right. Every conference I attend, speakers talk about great content that attracts links. The closest I’ve seen to an SEO conference really considering content with regard to audience, was at the most recent Brighton SEO, with Rosie Freshwater’s peta kucha Market Research: Informing SEO and Link Development (bottom of the page) describing how customer insight could inform building content which builds links. This entire guide by Distilled, while quite useful and comprehensive is also about links. Links, links, links, links, links. This is a great post about link bait from Search Engine Land, which explains:

“For SEO purposes, we are interested in sharing. You want people to write about your link bait on their websites, blogs and on their social media accounts. It would be a shame if 10,000 people saw your link bait yet none of them actually linked to it.”

That’s fair enough, but I don’t think it’s that much of a shame if 10,000 people still see your stuff. If 10,000 people saw this blog I couldn’t really give a hoot about whether or not someone linked to me. My audience would have gone up and I would have grown in community stature.

Links vs. Audience

The SEO industry never talks about content as a means to build audience, which is how every major media organisation would consider it. Of course, content which builds links builds search rankings, which technically builds audience, but it’s not exactly a way of thinking that puts people first. Since I’m a proponent of loving the people f rather than worrying about algorithms, I’m keen to change this mind-set.

Writers are Not Link Builders

Creating content as a means the end of building links is not something a writer will particularly enjoy. I was a writer in an SEO agency before – we wrote content with the objective of building links and it totally sucked.  Our boss even told us to try and not be creative and noticeable, because then the clearly paid for links would be less likely to be found. I lasted eight months. Much more pleasurable is seeing that people are actually reading your stuff, and then a % of those people are sharing it as well. It makes your writers happier. Think about creating content around a niche to build an audience and a community. Don’t go about thinking that your content should always be attracting links. People first, algorithms second.

Benefits of Building an Audience

The benefits of building an engaged audience and community can be wide ranging (I suggest reading Groundswell: Winning in a World Transformed by Social Technologies), but building an audience who are engaged, brand loyal and thus more likely to come back are in the end more likely to buy your stuff. You should gain customers who repeatedly buy from your business, have a view on your business, and can help you (and are prepared to help) by providing insight. Once you have this, your advocates will recruit more customers through word of mouth. Think about how you can make your content interactive, how your audience can be involved with the creation of that content – build the audience and you’ll have more people who will share your stuff. Ultimately, you’ll get more natural links through this distribution.

Category: Algorithms, Content, Inbound Marketing, Link Building, SEO, Social Media | Tags: , ,
James Carson | September 15, 2011
2 comments

How to Rank on YouTube… (Including Channel Data)

If I really loved data, I might bang on about how YouTube is the 2nd largest search engine in the world, how more content is uploaded in 60 days than all the major US networks could air in 60 years and how 3 billion videos are viewed a day. I might do that, or then I might just suggest you go to YouTube and check these stats out for yourself.

I might also think that someone who is allowed to write for econsultancy wouldn’t miss out some important factors about the YouTube ranking algorithm, particularly when these are readily available in the YouTube  Content Creator’s playbook. I often feel most of the information we actually need to do our jobs is actually provided by Google/YouTube themselves, but for some reason we need to get the CTO of another company to completely miss out channel data from ‘How to Rank Highly on YouTube Searches’, even though it’s pointed out as a ranking factor by YouTube. Very strange. Alas, the document is below and you can see for yourself.


How to Rank on YouTube (a more comprehensive guide)

Now, because the post on econsultancy is incomplete, I’ll go through the ranking factors in a bit more detail. First of all, I’d like to point out a couple of posts that gave me some depth while doing the research – it’s just they were written near 2 years ago:

Now, we’ve these two post’s backing, a little common sense and what YouTube tells us, there are five key factors to YouTube search:

  • Keywords
  • Engagement
  • Time based
  • Channel strength
  • External

Now these can be broken down further so that we have a bit of a clearer understanding:

YouTube Ranking Factors

I’m slightly uneasy about the idea of inbound links to your video causing it to rank higher on YouTube. It sounds like the sort of Causation vs. Correlation debate you regular see in SEO, and I don’t really feel I’ve seen enough evidence. Additionally, I don’t think it’s that sophisticated. All the same, it may be a factor, but not one I’d pursue. I’d be most concerned with getting keywords right, making sure the video is engaging (pretty much the creative part) and getting it up quickly and thus making it topical. So, breaking it down, I really think (at the top level) the algorithm is something like this:

YouTube Algorithm

I haven't based this on cold hard data, more my own observations and a sprinkling of common sense - let me know if you disagree!

But I wouldn’t lose sleep over the makeup of this anyway. Video virality relies on strength of creative, distribution and a lot of luck. Search is a useful sub factor within distribution, but it’s not the most important thing in video. I don’t think ‘Charlie Bit My Finger’ really got where it was by having great keywords in the title. The same could be said with most of this list.

Category: Algorithms, SEO, Video, YouTube | Tags: , , ,
James Carson | September 7, 2011
No comments

Why I Love Panda: Part Two

So I had so much to say about this topic that I had to split it into two posts. Here’s the first one if you haven’t read it, now here’s the second part:

Becoming Over Reliant on One Traffic Source

Anyone hit hard by Panda (25% traffic lost or more) should consider their over reliance of Google as a generic search traffic source. I know that for most sites, Google is going to remain the number one traffic source for the considerable future – but what really drives search? Significant traffic is rarely pushed through generic keywords unless there is scale in the long tail and a website has a robust content strategy. Over relying on Google as a marketing channel is downright dangerous. You ‘d have to be pretty reckless to invest in a portfolio of one stock and expect not to be met with huge volatility and risk. You’d have to be pretty reckless to invest in a business model that relies entirely on one marketing channel, never mind one search engine.

You need to invest in a wider marketing model – one search engine (big though it may be) is too volatile!

We Can All Learn from Car Insurance Comparison Sites

Often the most numerous and most valuable visitors come through branded search, which will be largely unaffected by Panda (although there will probably be some knock on effect on branded search as a consequence of disappearance on generic terms). Check it out on the Keyword Tool – in the car insurance market, all of the brands are now markedly more voluminous than the generic term ‘car insurance’. How have they done it? Largely through TV advertising and traditional techniques reinforced by social media. In one case they’ve even created a keyword based entirely on a creative campaign – hats tipped to Compare the Market. Okay, they’re effectively affiliates – but they’ve also created a system (original product) that makes car insurance easier to buy and more accessible – they also don’t even really need to rank higher than the providers themselves for the generic term.

SEO isn’t a substitute for marketing – Panda should make us realise that.

Pre-Panda Makes Me Angry

I’m going to sign off this longish two part post with a story. Bauer is a company with brands that exist and have huge engaged audiences from the strength of their content. They invest millions of pounds a year in journalists to make sure the content is of top notch quality. But then blogs came along and could just take magazines and scan them, then publish them on the Internet; thus devaluing the product. Check out a search for ‘fhm’ on Google Images. You’ll be hard pushed to find any image hosted on FHM.com. Most of the images listed are so without permission. I have done a fair bit of optimisation on FHM.com to get us ranking on image search to little avail. I actually don’t think image optimisation is particularly worth it in contrast to other SEO disciplines such as link building. Google doesn’t even rank images from the FHM.com domain for a search for FHM! What we end up is a complete travesty of economics whereby publishers pay expensive rates to create content, then this content is ripped and published on third party websites which often rank higher than the content creator for certain keywords. This effectively parasite website then can earns revenue via Adsense because of this traffic. You’ve seen this example in image search, but it’s been like that for the main SERPS for too long.

Panda, bring some sense to the world – make the original creators of content and the original suppliers of products rank higher! That’s what this is all about!

Category: Algorithms, Google, Lifestyle Publishing | Tags: , ,
James Carson | August 28, 2011
No comments

Why I Love Panda: Part One

Such a big post I thought it’s best to split it up into two – so check out Why I Love Panda: Part Two just here. If you haven’t checked out Pandalized.com yet, then take a look. Even if you’re not strictly an SEO it’s a great data set showing the big hits taken to content farms in the wake of Google’s Panda Algorithm updates. I find eHow.com’s graph particularly interesting. Three seemingly mega hits in succession, reducing it to pre-2010 levels.

eHow.com

Seeing this data, you can only really say that Google’s approach has been nothing short of brutal. Across the board falls come rapidly and almost instantly from updates. There’s clearly not been much of a warning. Some of these ‘businesses’ must be getting seriously worried about their future. But do I feel sorry for them? No I don’t. There are plenty of reasons why – so many and I had to pour out so much that I’ve had to split this over two posts! Here we go with the first three reasons of why I love Panda.

Gaming Algorithms and Reducing Content Cost

Demand Media have profited for quite some time by building a system that basically worked out many uncompetitive long tail search terms, and created content for a vast variety of long tail searches. Having a long tail strategy is fine, but I feel doing it on this scale devalues online content; particularly when an amateur is getting paid the pittance that Demand award. Consequently, journalism as an online paid profession is now at serious risk.

Check out TechCrunch’s recent article – Now Can We All Agree That The “High Quality Web Content Experiment Has Failed? The economics of professional online journalism in comparison to how print operates largely doesn’t work. Content is not sold, thus there is a huge chunk of revenue removed from the model. The only viable option for online journalism is currently where it’s used to sell products (ASOS.com, web editors of the world salute you).

Demand Media style models squeeze the possible profit margins on online content. Content farms inevitably drove the value of properly researched professional content down. I agree wholeheartedly that you can get very useful insight from a site like HubPages, but more often than not you have to go through quite a lot of tosh. I’m wondering, for instance, why any writer would like to become a daddy on HubPages rather than on their own publishing platform, unless they had the ulterior motive of self-promotion. I would assume the majority of people writing on this site are doing so to promote their own agendas and sites rather than for love of their topic. Hey, I’ll even admit it – I deliberately achieved a 75+ Hub score to link back to some sites and have the rel=”nofollow” removed. The content was ‘quality’ enough for this to happen.

Panda slams the models that game search engines – it celebrates authoratitive authorship built up over time.

Link Building via Content & Further Devaluation

I’ve previously worked for an SEO agency for more than a year where I was a ‘copywriter’, but the only real point in my copy being published was so that links to clients could be inserted into it. I certainly learned a lot about airport parking, credit cards and car insurance, but the knowledge that my writing was created to be linked and not seen was soul destroying. Writing that is used to game search engines and content as a commodity needs to go. Content should be used to build audience and engagement, not fill up the web with keyword heavy titled documents that answer huge amounts of search queries, linking to a client who may not deserve those links.

Panda sees content not as a commodity, but as something written by a human being.

Affiliate Problems

That SEO is affiliated with spam is nothing new. Indeed it’s often affiliates who are responsible for SEO spam. Affiliates – those people who don’t have any of their own products to sell. I’ve got no problem with affiliates; I am one myself on an extremely small scale. But it’s curious that you can expect to sell a product indirectly and rank you higher than your partner who does sell it directly. It’s their system, their business, which you’ll be using. They’re the ones who have taken the risk in setting up the system and invested in it. I’m weirded out by affiliates that rank higher than the original product. Yes, marketing is far more than just selling products; it’s about how you sell them, but people who create original products deserve to rank higher than those that piggyback them. Make something original, rank higher – that logic should mix with product as well as content. If you create a system that allows better comparison of products, then certainly you should be allowed to rank high as an affiliate, but someone who is taking a product and repackaging needs to be wiser.

Panda should help smack affiliates, and get the original stuff closer to the top.

Check out – Why I Love Panda: Part Two

Category: Algorithms, Authorship, Content, Google, SEO | Tags: ,
James Carson | August 24, 2011
No comments

Facebook EdgeRank – A Simple Guide

I’m not that much of a fan of algorithms. While it can be satisfying fun to crunch the data associated with algorithms and find out something new, I never really find that data particularly actionable. It’s often small beer. My answer to: “We’ve found that site speed now accounts for 2% of the Google algorithm,” would often be, “Who cares? Build links rather than spend money on a developer – it’s still a more cost effective way to get up a search engine.” I’d prefer SEOs to take a more human touch when working. Although, that said, this post is about Facebook’s algorithm – EdgeRank. Basically, it’s the algorithm that works out what should appear at the top of ‘Top News’ (the default for desktop, not mobile) on www.facebook.com.

Sources:

Before we get going, I’d like to point out some sources. Buddy Media’s Guide – Facebook’s EdgeRank: How to Make Sure You’re in the Newsfeed (PDF) is probably the most actionable White Paper I’ve ever read. It doesn’t go into depth about what EdgeRank is, just how to rank. Next up – Kelvin Newman’s longish post on the subject of EdgeRank. There’s also a few more things to feast on at the bottom of my blog Facebook Like Does Count as a Ranking Factor (I Reckon)

The Algorithm

This can neatly be summed up in triangular fashion thus:

Facebook EdgeRank Algorithm

Now let’s get to grips with this. The Newsfeed Item called ‘Object’ is just the Status update. Now we have the exterior triangles to tackle:

  • Affinity: This is how close you are to that object. If you engage with it, it becomes more relevant. If your friend or another object (like a page) you are close to on the Open Graph interacts with it, then it becomes more relevant. If you haven’t interacted with anything associated with that object for a while, then you’re affinity will be weaker. For instance, I see very little information from ‘friends’ I never interact with.
  • Weight: I don’t think Kelvin got the definition of this right on his post. Rather than it being the ‘weight’ of the object, I believe as a ranking factor it is the weight of the engagement. Engagements are Comments, Likes and Shares. A Comment has the most weight, then a Like, then a Share. The premise is simple – the more Comments and Likes a post gets, the more Facebook will give it more Weight, deem it relevant to you, and shunt the Object up the newsfeed.
  • Time Decay: The older posts are, the less likely you are to see them at the top of the feed. The default for www.facebook.com is Top News – that means posts have got to be current. You’ll rarely see a post more than 24 hours in top spot unless you have very few connections.

What’s the Weighting of Each Element?

I don’t really know – I would presume they’re all pretty similar. Certainly you can optimise for Facebook posting by putting an emphasis on weight – so asking questions and being explicit in requesting fans Like the page – check out Facebook Posts Should Meet 5 of the Following 10. You can’t really be as tactical with Affinity and Time Decay just decays. If there’s one you can worry about – it’s Weight.

Category: Algorithms, Facebook, Social Media | Tags: , ,