'Would it ever actually take that long for link-building to have an effect? I thought that links were supposed to lose value over time, and so are likely to peak pretty soon.'
You can base performance related pay on achieving specific search engine positions and link building tends to target a small cluster of specific search phrases. However, this is a very shallow look at what SEO does for a site.
With personalised and universal search results, measuring positions is becoming not only difficult, but pointless. We can measure a search result at a specific position, but 30% of your users might see it in a lower place and 40% in a higher place. Search engine positioning is becoming a pointless metric.
Also, where consultancy is concerned, with technical improvements to the site, you will see improvements across all positions the site holds in the search results. I for one, am not going to put a performance model together based on the overall average improvement of 100,000 search phrases; it is a reporting nightmare and would take more time than the actual work being done.
We tend to focus on traffic. Specifically, non-branded organic search derived traffic. Non-branded search has an impact on branded search, so there should be indirect uplift there too, but we largely ignore that.
Our typical aim for an SEO campaign (tech & content consultancy and link building) would be to increase non-branded search traffic by X%; these are our usual KPIs and we sometimes tie performance models to these.
However, we can conduct six or twelve month programme of work on a site and then walk away. The traffic benefits of this work will exist long after contract end (until they tear the site down and replace it), so performance-related pay based on traffic can be a little unfair, especially if the client wants a cost-per-click model so that they can compare SEO ROI against that of PPC. This is a common request from clients, but there are fundamental differences between SEO and PPC activity; traffic gains from SEO work does not go away when the spend stops (like PPC does).
A lot of what is written is 'SEO books' is out of date, or just shite. I have never read one that is worth the paper it was printed on, apart from as a basic beginner's guide (but I wouldn't even give one to a trainee as they can learn faster from our internal training programme).
A lot of our knowledge came from trial and error (not on client sites, but testbed sites we control). A lot of it was driven from hypothesis and testing, and we do use control pages; it is a little hard to do proper blind testing on this kind of thing though. We use testbed sites to confirm or refute other SEOs' assertions - we don't much like taking things as read, we need to see some impact before we take methods to clients.
We tend not to publish our results very often - this is the stuff that gives us a competitive edge over our competitors.
Link building is fairly bread-and-butter work, and to some degree, content optimisation and keyword research is as well; I tend not to get involved in that very often (the people that do this earn a lot less and have a lower rate-card cost). I tend to stick to the in-depth tech stuff and the overarching strategies and supporting new site builds.
Everyone in my business is self-taught or trained internally. You start out by reading everything you can find. With experience and testing, you learn what to discard as conjecture or nonsense. This is where a lot of the best people come from. Sometimes we cannot satisfy someone's ambitions and they leave to become a 'head of' at an agency that is keen to build a search offering and start training their own team. It isn't an easy industry to break into on your own.
Of the people I would field to handle any question or problem a client could generate (this includes myself), I would say it took at least three years of on-the-job learning to get to that point. But like development, it is a constant requirement to stay on top of changes in the industry, and this isn't just they way search engines rank pages, it is changes in HTML, CSS, CMSs, server platforms and other elements of the technology stack (you don't want to know the problems an Endecca information access platform with guided navigation introduces).
(,
Fri 25 Jun 2010, 10:51,
archived)
You can base performance related pay on achieving specific search engine positions and link building tends to target a small cluster of specific search phrases. However, this is a very shallow look at what SEO does for a site.
With personalised and universal search results, measuring positions is becoming not only difficult, but pointless. We can measure a search result at a specific position, but 30% of your users might see it in a lower place and 40% in a higher place. Search engine positioning is becoming a pointless metric.
Also, where consultancy is concerned, with technical improvements to the site, you will see improvements across all positions the site holds in the search results. I for one, am not going to put a performance model together based on the overall average improvement of 100,000 search phrases; it is a reporting nightmare and would take more time than the actual work being done.
We tend to focus on traffic. Specifically, non-branded organic search derived traffic. Non-branded search has an impact on branded search, so there should be indirect uplift there too, but we largely ignore that.
Our typical aim for an SEO campaign (tech & content consultancy and link building) would be to increase non-branded search traffic by X%; these are our usual KPIs and we sometimes tie performance models to these.
However, we can conduct six or twelve month programme of work on a site and then walk away. The traffic benefits of this work will exist long after contract end (until they tear the site down and replace it), so performance-related pay based on traffic can be a little unfair, especially if the client wants a cost-per-click model so that they can compare SEO ROI against that of PPC. This is a common request from clients, but there are fundamental differences between SEO and PPC activity; traffic gains from SEO work does not go away when the spend stops (like PPC does).
A lot of what is written is 'SEO books' is out of date, or just shite. I have never read one that is worth the paper it was printed on, apart from as a basic beginner's guide (but I wouldn't even give one to a trainee as they can learn faster from our internal training programme).
A lot of our knowledge came from trial and error (not on client sites, but testbed sites we control). A lot of it was driven from hypothesis and testing, and we do use control pages; it is a little hard to do proper blind testing on this kind of thing though. We use testbed sites to confirm or refute other SEOs' assertions - we don't much like taking things as read, we need to see some impact before we take methods to clients.
We tend not to publish our results very often - this is the stuff that gives us a competitive edge over our competitors.
Link building is fairly bread-and-butter work, and to some degree, content optimisation and keyword research is as well; I tend not to get involved in that very often (the people that do this earn a lot less and have a lower rate-card cost). I tend to stick to the in-depth tech stuff and the overarching strategies and supporting new site builds.
Everyone in my business is self-taught or trained internally. You start out by reading everything you can find. With experience and testing, you learn what to discard as conjecture or nonsense. This is where a lot of the best people come from. Sometimes we cannot satisfy someone's ambitions and they leave to become a 'head of' at an agency that is keen to build a search offering and start training their own team. It isn't an easy industry to break into on your own.
Of the people I would field to handle any question or problem a client could generate (this includes myself), I would say it took at least three years of on-the-job learning to get to that point. But like development, it is a constant requirement to stay on top of changes in the industry, and this isn't just they way search engines rank pages, it is changes in HTML, CSS, CMSs, server platforms and other elements of the technology stack (you don't want to know the problems an Endecca information access platform with guided navigation introduces).