A detailed look at what can (and can’t) be automated in SEO

robot_flower.jpgIn my post about SEO automation, SEO expert Halfdeck expressed concern about the possibility of customers preferring a sophisticated SEO tool over a highly-trained SEO professional. I am sure many of my peers had the same thought running through their heads when they saw my post on Sphinn. The post even made it to the front page!

The reality is that we should embrace software progress in our field. We should look at software improvements and innovations as tools that can help us scale our own business and be more successful. Is it bad to use SEO software that does all the tedious work and helps us serve 20 clients per month instead of 10?

Man vs. Machine

If we look at the last 100 years, man has achieved an incredible amount of progress. Every invention brings a new level of comfort. You can do many things far easier, faster and more efficiently than before. There is an irreversible momentum in this direction.

The fear of computers replacing the work we do is not unfounded, of course. Machines are replacing human labor everyday. The more repetitive the work, the easier it is to get a machine to do it for you. On the other hand, the more creative and more social the task is, the harder it is for a computer to replicate.

There is software for almost any professional activity that we do, yet we (the professionals) are still in high demand. Nobody is going to trust his or her business to a computer blindly, after all. If something goes wrong, who would they complain to? 😉

So the question becomes: do we resist change, or adapt and thrive?

SEO processes that can be automated so far

Search engine friendliness. Most SEO experts start optimizing sites from keyword research. I start by establishing a baseline for the web site so that I can identify problems and measure progress. This includes:

  • Checking the current rankings and index penetration

  • Checking and fixing broken links, missing tags (title, meta description), etc.

  • Checking search engine penalties, duplicate content problems, etc.

It is not trivial, but it is perfectly possible to automate all this. The harder part is fixing the problems automatically once found.

Keyword Research. It is hard to do keyword research without a tool. On the other hand, advanced keyword research involves more than just selecting a group of keywords based on their KEI. In my keyword research process I focus more on the organization of the keywords, measuring relevancy, understanding the intention of the searcher, the real value of the keywords based on the potential traffic, as well as how much they are worth in the PPC marketplace. I also include in this process the selection of which pages should be optimized for which keywords. Again, this is not trivial but it can be automated.

Competitive Analysis. Studying and following a top-ranking competitor is one of the most fundamental, bulletproof ways to achieve high rankings. We normally rely on tools to perform the analysis, so clearly it is possible to automate this.

Site Architecture. If the site is built with a content management system, it is possible to make changes to the CMS to allow for site architecture changes that improve the site’s crawlability and navigation.

On-page Analysis. Extracting the relevant metrics from a particular web site also requires the use of tools, so at least extracting the raw data is automatable. Acting on that data to create an improved version that is still comprehensible to a human being is not, however.

Link Analysis. Extracting the links and anchor text of a particular web site efficiently requires tools, so this is automatable too.

SEO processes that cannot be automated so far

On-page optimization. In order for a computer to make changes to a page and still keep it readable for a human being, the computer would need to comprehend it. If you look at how badly automatic language translation tools work, you know that we are still a little bit far from that. I can’t say that it will not be possible in the future, though.

Link building. Aside from identifying link targets, automatically submitting your site to directories and social bookmarking sites, or doing massive link exchanges, link building is probably one of the most difficult tasks to automate. Why? Because those easy links I just mentioned are not worth much. Useful links are generally placed by other webmasters and they usually don’t like automated spam asking for links. Effective link building requires creativity and a personal touch.

That being said, here is an interesting idea that might (or might not) work. As I discussed in my previous post about advanced link building, giving tests and rewarding test takers with badges is a really good way to build links. I think that this specific link-building method could be automated. You create the web application that gives the tests, generalize it enough so that you can plug it into any of your sites or your clients’ sites. I am imagining here a test-giving engine that takes a feed of questions and correct answers. You could buy several trivia books to feed the engine with groups of related questions and let the visitors do the link building for you. You can also automate the initial promotion of the tests by submitting to social sites or e-mailing your popular friends 😉

What do you think?

  • SEO can't really be automated. Sure – you can use tools to help you out but like you point out, at the end of the day you still have to make the on-site changes yourself, or fix broken links yourself, or linkbuild by hand.

    Automated tools can be a great help and can save you a lot of time. But SEO is still a task for humans.

    • It is possible to fix broken links automatically 😉

  • Its interesting to compare how the process that can’t be automated or only automated to a degree are the ones that search engines use as major factors. Bar domain age, which of course has its exploits.

    I think a lot of the processes that you mention as automatable (is that a word?) are overlooked. Maybe due to the fact that a lot of SEOs are either really bad programmers or not programmers at all. I think these people could really benefit from some of the features that RankSense packs. Naturally, its not going to make the difference between page 100 and number one on page one, but it could be that little extra that makes the difference between coming up on page one and coming up on page two or shaving a month off results because you were able to better evaluate you possible courses of action.

    P.S. There is another slug in the URL. This seems to be causing a bit of trouble with Stubmle Upon. I have Stumbled it but its having problems registering it and its having some conflicts with character encoding, ala 404 error.

    If its any help, I wodged this PHP function into some other blog system a while back to stop the same error:

    function strtourl( $str ) {
    return strtolower( str_replace( ' ', '-', preg_replace( '/[^a-z d]/i', '', $str ) ) );

    • David – Thanks for the heads up. I am in the middle of moving the blog to a new more powerful server and a newer version of WordPress. This is an old server (250m ram) and crashes too often. My blog is getting more popular 🙂

  • Pingback: 100 Posts Old » SiteLogic Web Development()

  • I don't think you can completely automate SEO simply due to its competitive nature. If there was some incredible tool that you could apply to your site, and just click a button that says "Optimise my Site!" then everybody would use it so they would all be using the same tool, the same algorithm but there will only ever be one number #1 spot.

    Therefore we will always need to push forward with new and creative ideas and those people who do that will rise to the top. I would say that you test idea isn't so much automating the seo process, but using an automated tool to do something creative. If everybody caught on and did the same thing it would probably lose its effectiveness.

    • Caroline – I have to agree with what you are saying. Tools can give us advantages, but as soon as everybody is using the same tools we're back at square one. Creativity is a human trait and is the most important and determinant factor

  • Absolutely many repetitive functions can be automated. I also agree Hamlet that many more functions can be automated if you've got some programming expertise in-house. In the end, anything that makes us more efficient in our jobs, gives us more time to be more effective too.

    I do agree that not all roles can be automated … but certainly there is lots of potential!

    • Thanks for your comment, Jeff. As a developer, I see automation as a competitive advantage. The more efficient my business the more profitable I am.

  • Pingback: SiteMost’s Weekly Blog Recap 2/10/07 at Brisbane SEO Blog()

  • Hamlet- great post. I agree with everything that you've mentioned. You can automate certain tasks, of course. But organizations need to look at this as tools to help Search marketers, not to replace a search marketer – that's just impossible. You still need a human element to think outside of the box and apply creative strategies to an SEO campain.

  • Derrick – Thanks for your visit. I'm glad we are on the same page.

  • I'm new to all this SEO stuff and found this article very interesting – even though I stumbled on it months after written!

    The point I'd like to make is that the vast majority of webmasters / bloggers have little to no interest in SEO and are simply trying to get their message out / make a living online. They (read me!) then get embroiled in a subject that is too technical for them to fully grasp and find themselves losing the Will to live! If there is a way to automate the bulk of the tasks then surely this is great for the vast majority who wouldn't employ a professional SEO in any event.

  • Great article. I think there needs to be a balance between the two as well. Thanks for the clarification on this issue.