Can SEO be Automated?

33 comments

Matt Malden from Yield Software noted that they are out of stealth mode. Their goal is to automate the SEO/SEM process from keyword selection, to bid pricing, to landing page optimization. From their homepage:

It drives the most relevant traffic to your site and helps turn your visitors into valuable customers. It does this by simplifying and automating the process of search engine optimization (SEO), pay-per-click bidding (PPC) and landing page optimization.

The question is, how much can you automate SEO?

Comments

you can't ... period

You can't automate writing good content.

Yes, you can

But it does take some serious effort. PM me if you're interested in knowing more.

No way

dude I wouldn't trust my clients to some automated program... it's just not right on multiple levels...

No problem

If you think you know all about it already, you won't find me trying to convince you otherwise.

..

Ralph, I think the key word here is "good".
You can automate up to a point, you can even create readable understandable text that conveys an idea or thought, but then the subjective human brain takes over to decide what is "good" or not.

This is and has been an interesting concept, I see it as kinda equivalent to playing Chess or Poker with a program.

One of the good things about being human is our capability to think illogically, jump to conclusions and most importantly to make leaps of logic based on a 'hunch' or a 'feeling'. In my opinion, SEOs will only be in real danger of extinction when programs like these are able to mimic those mental processes.

Depends what you want to do.

Depends what you want to do. If it's ranking for chosen keywords, you can automate it easily.

If you want to optimize a page it's got to be hard to automate. How can an automated program scan a page for a topic and match it to the heaviest traffic relevant keywords without a hand check? For example I have a site on paints and the page is all about blending the paints. What if the software decides it wants to rank me for food blenders and puts those words in key places on the page? I guess you can build a semi automatic system, but it might be quicker to just manually decide on those page titles...

Quote: One of the good

Quote:
One of the good things about being human is our capability to think illogically, jump to conclusions and most importantly to make leaps of logic based on a 'hunch' or a 'feeling'. In my opinion, SEOs will only be in real danger of extinction when programs like these are able to mimic those mental processes.

You need content to automate but you need to automate as all those "hunches" and mental processes eventually form the algorithmic decision. Small tweaks in an automated environment can mean big gains.

SE's and SEO's really think alike, its all about clasifying your document(s) in their classification system.

Not talking AI here

As academic AI research has historically suffered from an overkill of targeted features trying to machine-recreate the entire human language.

And yes, I do mean "good" content - freshly generated, though, not optimizing any set page. Agree 100% with Sly here - don't see how that could be automated seamlessly. (Machine aided editing may be another story, of course.)

But if you want to create text that a) makes sense, b) is pretty well written (I mean let's face it: How many humans will actually produce good let alone excelling text?) and c) will stand human editorial muster - yes, that can be done and actually IS being done. And performing well in terms of SEO, too.

Ime soory

SEO taken as a whole is a NP complete problem

FFS is some program be able to dig into the crap that is the average site a seo gets lumbered with to improve it.

I may be a relative newbie SEO but I bet i have more years coding than 99% of the alist Guys this is vapor ware.

thers a technical description but I was expalined to me as problems that would be esier to solve if we had true hard AI.

SEO is their side-show

I'll bet these guys do verrrry basic stuff both in terms of SEO and SEM. Anyone with the pitch they're making is targeting novice online advertisers, for which level of SEO/SEM solution sophistication requirements are low. SEO for them probably will mean what so many SEO me-too shops mean = crapola.

Reminds me a bit of the early days of PPC

when newbie Net entrepreneurs where being conned by pseudo-SEO artists hawking "Top 10" rankings which were nothing but (thentime) cheap GoTo PPC ads.

Not saying that's their only business model but their site does smack too much of a PPC biased outfit to reasonably expect a lot of deliverables on the organic SEO front.

From a scientific perspective

..I'd love to get deeply into auto content creation, because I think it really could be done well enough to fool most of the people most of the time.

..On the other hand the moral dilemnas about making money from it are another matter. Is it that far removed from brainwashing???

Unless you're being ironic here...

I don't really get the "moral dilemma". Would you say the same about press releases, product announcements, ads, commented statistical tables and other forms of corporate droidspeak? And if not - why not?

I mean, it's not as if the Web as a whole were particularly dominated by high end literary prose, deeply suggestive well crafted poetry or similar feats of human creativity.

Here's a little experiment for you: Go over to elance and hire ten different copywriters for an identical article project (topic, keywords, etc.) without making them aware of each other. And let it be something fairly specific, i.e. not too generic. E.g."Learning to play golf", "Teach your parrot to speak", "How to give up smoking", etc., you get the drift. Let them write 10 different articles each.

Then, when they've submitted all 100 articles, compare results. How "different" or "non-duplicate" do you think they will be?

While this stuff may not trigger any dumb search engine duplication filters, five will give you twenty that it'll be nothing that couldn't be generated by a fairly sophisticated automatic text generator at a fraction of the cost and time.

Now where's your moral dilemma in that scenario?

to continue

I'm not worried about triggering search engine dupe filters, for the most part all the articles I've had written have only ever saved me the time in writing them myself (there seems to be a total lack of understanding of SEO by most article writers).

And as I say auto generated articles could well get to a stage where they fool most of the people most of the time. (If they aren't already there)

I suppose for me the real moral dilemna comes when the spark of originality is taken out of the equation. Take a situation where autogen content has got to the stage where it is almost indiscernable from human generated content, and that enough people have access to autogen content creators, you can have hundreds of pages created from one initial creative process.

At this point you have a number of moral issues that are raised:

1. Does the autogenerated content of one point of view completely obliterate that of any contradictory points of view.

2. Does consistent reinforcement of one point of view constitute brainwashing (and of course is this any worse than what is already practiced by many religions)

3. Could autogenerated content lead to a content monopoly, in the same way that Google has a search engine monopoly? (On the other hand would it just play back into the hands of organisations like Rupert Murdochs News Corp which pretty much had a content monopoly before the rise of the internet)

4. Any proof of concept solution produced by a small organisation will immediately be bought by less scrupulous large organisations. (On the other hand at least some small organisatioons will benefit in this situation)

5. Election propaganda, Religous Propaganda - where does this lead?

I'm just trying to think of where this is heading? - and I'm not sure I really like where it is going long term.

Or to put it succinctly

There is nothing morally wrong in producing a nuclear bomb, but the issues raised when you actually come to use it are another matter.

Conversion Rates

Yield automatically and continuously optimizes your web pages based on what works and what doesn’t. This includes optimizing phrases, images, layout, colors, fonts and much more.

Why is my website bright magenta, why is all of my text blinking 72pt comic sans, and why has my company's logo been replaced with a picture of Britney Spears?

I had a chat with Matt from

I had a chat with Matt from Yield Software. He said they do not intend to automate content creation or link building...they aim to automate more of the testing and labor intensive stuff outside of that.

Assisted Automation is the Way to Go

>>>they aim to automate more of the testing and labor intensive stuff outside of that.

No reason not to.

If you have the resources, knowledge, and skills you can build some advanced tools to assist in automating the process of optimizing your content. I think automated assistance is better than complete automation. Top notch quality content, the type that is produced in magazines and newspapers, will be very difficult to automate. But if you have a large scale site with content constantly entering the system it may be beneficial to build a system that assists those entering the content to optimize the pages. It can analyze the pages for entities and concepts, cross reference with keyword databases, and suggest keywords. It can build tags based on templates. You can build automated QA systems. You can define logic to determine which pages link to each other. The list goes on.

At a certain point it comes down to the level of control you want and can afford to have, how precise you want to be with your optimization, and the amount of content you're managing. If you have hundreds, or thousands, of pages being added to your site each day do you think you'll be able to optimize each of them manually? Not unless you have some serious resources. Instead, you may have to settle for training those who enter the content to optimize it or creating tools to assist them.

Long story short. There are a lot of tools and automated systems that SEOs can and should utilize to assist in their site management and daily work.

..Hmmm

Quote:
Top notch quality content, the type that is produced in magazines and newspapers

Personally I see very, very few pieces of 'Top Notch' content in newspapers.

Magazines tend to be slightly better, but all are aimed at mass market consumption so tend not to pull in readerships who are deeply interested in their subject.

...

Quote:
...He said they do not intend to automate content creation or link building...they aim to automate more of the testing and labor intensive stuff outside of that.

Well this is a far cry from full automation. This sounds more like another SEO tool.

Quote:
I mean, it's not as if the Web as a whole were particularly dominated by high end literary prose, deeply suggestive well crafted poetry or similar feats of human creativity.

LOL, for the most part I agree with your sentiment.

However, I do believe that once in while a human does create something 'good' with words. When a program can compete on a subjective scale of what is 'good' with the likes of Mark Twain, Jack Karawack or George Orwel, just to name a few, I will be looking for a fully automated solution to SEO... Till then though I will enjoy using tools to help me accomplish my SEO goals.

No Pulitzer Prize stuff

and not even Nobel Prize class. (Now what the heck did he mean by "not even"... heh.) No competition with the really creative stuff. Just some good old "content is king" kind of material. As snowy white and squeaky clean as any "organic" SEO could wish for.

Just good, average, readable content that, while it won't get you all excited for want of more, won't make you realize it's autogen after ten seconds of digging into it either. Or after ten minutes, for that matter.

In short: Output that will be unique and will hold up against any basic Turing test kind of setup, nothing more, nothing less. No detectable footprinting, that's all-important for SEO, of course.

Ian: I wouldn't expect this to turn into any major problem monopoly wise. After all, it's GIGO all the way. Whatever is put in will come out again in one form or another. Plus, if plenty of people have access to these kinds of tools, there'll be no Murdoch like situation, quite the contrary in fact.

What it might mean indeed, however, is the death of the search engines as we know them today. Because once you start deploying such text mills in earnest, i.e. in droves and fire them up to generate, say, billions of unique versions of the same basic texts, it will only take a handful of early adopters to double, triple, quadruple etc. the overall volume of the Net every other month or so...

Whatever is put in will come

Whatever is put in will come out again in one form or another.

With the exception that it will typically be biased towards misinformation that is thought to have high profit potential (in one form or another)...likely moreso than the current marketplace already is because it will syndicate / recreate the most profitable chunks at low or no cost.

..I think this is the key

Quote:
Plus, if plenty of people have access to these kinds of tools

If enough people have access to the tools then the monopoly situation is avoided - which woud be good, but as with all automation some people are going to be better at developing the tools than others. Will this lead to an arms race where those with the most money to put into their tools will get the most return or will this just lead to people having so much choice that no single organisation can get a dominant position. (you could site search engines as a perfect example of an organisation with massive resources effectively developing their tools to a position where they have a near effective monopoly.)

Typically biased towards misinformation?

Only if you assume that the same applies to all marketing copy. Which might be a bit of an overstatement.

In any case it's that old "what's in a knife?" carrot again: Any tool can equally be "used" and "abused", though determining which is which will generally be a question of which side you happen to be on.

Quote:
Will this lead to an arms race where those with the most money to put into their tools will get the most return

Most likely it will. The spectrum of tools available today ranges from the quite-ingenious and mildly canny to linguistically naive if not entirely useless Markov chaining or scraper stuff.

But don't let anyone be fooled by their respective developers' hyped promises: If you want to achieve to serious quality output in this field, there'll have be a no-too-trivial investment in terms of development costs.

While potential returns are truly huge (the three digit million range being an extremely conservative estimate - more likely it'll prove to be in the two digit billions), as far as we've investigated the matter to date, driving it to the proof-of-concept stage, getting this rolled out properly will require financing of 3-5 million. So this isn't anything for your average small time developer players to dabble in - but neither is it so exorbitant as to prove inconceivable.

As for who will actually get access to what and in which form, this will obviously depend on the various business and marketing models adopted.

More importantly, however, as with most genuine innovationes it will most probably be a question of who catches on first, fully grasping the potential involved. Though the SEO industry seems to be a fairly obvious contender, based on our experience so far I'm not too confident about this. At this point in time I'm more inclined for various reasons I won't go into here to expect this particular challenge to be taken up by some outside party.

Back to the orig Q - Can SEO be Automated?

SEO can be automated to a large degree. As many of us are aware, some of the many basic ranking signals are fairly easy to fix, though many large corporate sites struggle with sufficient focus to address them properly.

At the company I recently joined, Netconcepts, one of their larger areas of focus is their GravityStream product - a webservice which allows for automatic optimization of large, dynamically-generated websites. It's what I call a "near turn-key" service that automatically improves a number of ranking parameters (including individualizing with unique TITLE tags, META descriptions, H1s and making URLs spider-friendly) -- with all content handled by proxy and served up on the site's own domain name.

In addition, the service enables considerable administrative interfaces to allow customization of individual site pages, so either we or the client can have keyword optimizers and seo copywriters add in customized content into pages.

While I can't mention all of our clients, our service was recently highlighted by DMNewsfor a major retailer website, and we have partners such as Performics who perform SEO for their clients via our backend infrastructure, adding on their own unique keyword/content optimization processes and additional tracking/reporting tools.

(Pardon my tooting our own horn, but I'm enthusiastic about the product. As part of my role, I'm working on developing it out even further, and I assist with the day-to-day ops.)

Yield Software

Looking through their website, the only telling statement I can find is in their definition of SaaS (Software as service).

They state:
"All you need to do is let Yield know your web site address and who your competitors are. Yield will then guide you step-by-step through the process of modifying your web pages for tracking and analysis. After that, just sit back and relax."

So basically, they look at your competitors, copy/steal their optimization phrases and keywords, and then paste them into your website in the so-called SEO tag areas.

H'mm it's possible, but without an exact duplication of the competitor's body content I don't think they can get all the long-tail phrases that really create the conversions. I'm guessing what they're doing is very very basic SEO which any average joe can do.

If they have access to a handful of PR7+ URLs that could handle the Google listing/linking requirements, but it's not going to drive significant traffic. Perhaps their PPC program is assigned that requirement.

Nah, SEO when done right is a custom product for every website. If I ripped off my competitor's websites and left it at that, I'd clearly be doing a great disservice for my clients. There's a huge difference between SEOs that optimize for common keywords and those that optimize for conversions. I've always been in the latter group which takes a lot of creativity in the site design and keyword research.

No offense Silver, but it sounds a little like your product as well. On the positive side, what your both producing is definitely better than nothing.

...

Quote:
So basically, they look at your competitors, copy/steal their optimization phrases and keywords...

Steal optimization phrases and keywords! How horrible...

What... did I hear someone say... You can't steal common words?

WTF???

You can't own common phrases or words... What the hell? Who made these stupid rules?

You mean I can't own the word [Sex] or the phrases [Buy Games] or [Buy Software]... What kind of world is this where I can't own a common word or phrase???

And if I can't own a common word or phrase... How in the hell is anyone going to steal it from me?

>>>added - you can't get conversions if don't get the traffic first. SEO 101

steal

I used the term "steal" to mean grab the keyword from another website without doing any prior research on whether it's appropriate for your client.

Sure you can go ahead and mirror all of his keywords. Be my guest. Whether those keywords generate the kind of sales and conversions your client is looking for is a completely different matter.

Uh-Oh, looks like I'm dealing with another run of the mill SEM that believes 10,000 hits with one conversion is better than 5000 hits with 500 conversions. Traffic does not, I repeat..DOES NOT = SALES. Go back and play with your Adwords and their dismal results if you want traffic.

..

First, I never use adwords or any of googles 'free' tools. Why go for Penney's when you can make the big money.

Second - I was preaching conversions over traffic before you ever heard of SEO and before the term SEM even existed.

Third - when you say someone is stealing... it is rather obvious what you mean. Changing the meaning of the word after the fact still does not change the fact you accused someone of stealing, when they in fact could not possibly be guilty of stealing something like a common word or phrase.

Sorry you were offended by my sarcastic response, but throwing around accusations like "...basically, they look at your competitors, copy/steal their optimization phrases and keywords, and then paste them into your website..." Is in my opinion very offensive and maybe even legally libelous. If you meant "grab" that is what you should have written.

steal2

I stand by the word. If you can't comprehend how relying on someone else's marketing instead of getting off your butt and doing your own research is the proper use for the word then I can't help you.

Your second point totally escapes me since you don't know who I am.

Your final statement..MAY BE legally libelous..come on..I didn't just fall off a turnip truck. I'm shaking in my boots.

Let's quote you instead, "you can't get conversions if don't get the traffic first. SEO 101"

Dude you need another job, seriously. Traffic does not come first. Traffic doesn't even enter into the equation. You optimize for thousands of high conversion keyword phrases and that's it. These phrases create first page placements because the rest of the SEOs like you don't have a clue that they exist...And your client sits there amazed that all the other websites on the Internet are optimized for incredibly generic low conversion keywords because they drank the "high traffic" Google Koolaid.

To conclude..here's a dollar, go play in traffic.

..

Quote:
To conclude..here's a dollar, go play in traffic.

so you wanna get in a pissing contest do ya?

Fine by me, I like pissing contests, as I usually win. :-)

Quote:
Traffic does not come first. Traffic doesn't even enter into the equation.

Your kidding right? Or are you just regurgitating things you don't understand that you read somewhere...

I checked on a lot of your posts here about SEO, I have really good idea who you are...

Tell me there jboy, just how in the hell are you gonna get a single sale if don't have a single person showing up at the site?

I am waiting to hear, how you can create a single sale without anyone coming to the site...

"Traffic does not enter the equation" - stupidest thing I have heard in years...

I think this was originally

I think this was originally a thread about automating content.

I think this was primarily

I think this was primarily a yarn about mechanizing happy.

Hmm. Needs work. ;)

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.