Forum Moderators: martinibuster
Not my first idea, but one that I think might stand the test of time, is to use the following formula. I’d like to coin a new three letter abbreviation – the LTQ ratio:
Link Traffic Quality (LTQ) = (traffic per month from referring domains excluding engines / links into the domain as defined by Yahoo)
It seems to me that this would give you a pretty good idea of just how spammy you (or your client) has been in getting links to date. More positively, this could be described as a measure of how effective your online branding message has been recieved.
I would have thought that a site that has bought poor quality links would have an incredibly low “LTQ” Ratio, whilst a site that has only got one inbound link, from a great story on the BBC will get traffic. The LTQ would be an interesting KPI to measure. For one, Google can’t measure it (unless you’ve given them your analytics) and I don’t believe you should be measuring things just because you think Google likes a particular statistic. For another, it is a metric that anyone can find for their own site with relative ease and you can see if it is getting better over time. If it is, then EITHER you’ll be getting more traffic from your referring domains OR you’ll be getting higher in the search rankings under relevant terms – which in turn should improve your organic search traffic quality.
Is this a useful metric? I think so. But it is by no means the only one. One big disadvantage of this is that you cannot easily compare your own LTQ with that of a competitor. Again, simply measuring inbound link numbers against a competitors' rarely gives any meaningful information. It's the quality of those links that count. We have, for a while, all been looking at some of the obvious: ODP entries, Yahoo directory listings and lately Wikipedia mentions as a barometer and these help, bit they really don't go far enough in assessing what strategy a competitor has been able to do successfully that you, perhaps, have not. You need to be able to filter the wheat from the chaff. We've been able to do that, to a certain extent, but not in a way I can show I am afraid, but I can say this: Any two competing sites should really have a similar "set" of associated authorities associated with them. If you define these authority sites. By authority I mean decisions like "The institute of accounants" would be an authority for "accountancy sites" not "PR X counts as an authority.
Once you have a list of sites, you then compare the inbound links from those to the two competing links and call these links "trusted". these are ones that you can reasonably suppose will be bringing in good traffic and be positively contributing to the sites' respective reputations. This spearates the wheat from the chaff.
Not perfect metrics, but two quite good ways of defining or comparing link quality.
[edited by: Receptional at 11:09 am (utc) on May 15, 2007]
Hypothetical for you R, The Institute of Accountants adds some paid footer links, would that impact their LTQ?
If you are hypothesising the LTQ of the Institute site
Presumably it wouldn't affect the LTQ of the Institute of accountants in the short term, but might hurt their brand and therefore their own LTQ in the long term as less people link to them as an authority as their reputation takes a hit! If Matt Cutts has his way, of course, that reputation might take a turn for the worse sooner rather than later. I'm not going to second guess that wretched algo though. Life's too short :)
If you are hypothesising the accountnts' LTQ
Afain, not directly (unless the client bought the paid footer links instead of getting decent "editorial" traffic). But if the authority decreases its own relevence, then - slowly - less targeted people would go to the Institutes' site and hence less would click through. So yes, in the end, there's a net down side.
I define link 'quality' in three different ways:
1. Conversion. So a that sends 1000 uniques a month with 10 conversions (1%) rates a link ROI quality of 0.01 while another that sends 10 with 5 conversions (50%) rates a 0.5.
2. Reputation (branding plus). A link from the New York Times with a positive comment would get a brand quality rate of 1. A nasty comment and link from some little bloggart: 0. Most somewhere inbetween. Subjective but worth doing.
3. (SERP) Trust. More of a moving target than the others as each SE algorithm is unique and mutable. This one seems to be in line with the OP.
* Authority (natural, i.e. trade organisation, government; grown, i.e. ODP, Wikipedia, expert blog).
* Traffic volume referred.
* Referring sites backlinks: quality and number.
Rate all between 0 and 1.
Links have many values. Those that Google can't or won't measure are priceless.
I rarely pay much attention to competitors. I do like your method of comparing several sites backlinks to pick out potential 'authorities'. Will be interesting to see how many I am missing.
Are you are referring to link 'quality' in the 'Google likes it sense'?
Not directly (you know I couldn't possibly advocate link building for Google! ;) . I mean quality - in a broader sense - but encompassing and encouraging Google's view of a site as a positive part of the site's reputation.
I agree of course that measuring converting traffic is always better than measuring visitors. But a small - niche - site sending only three visitors a week will take a long time to build up enough conversion data to be statistically helpful for a top line measurement. The LTQ idea is a metric that is meant to be used at a macro level to be able to give a good indication of the site's past link strategies without having to spend a large amount of time analysins individual links. The measurements you are suggesting are probably more valuable, but extremely labour intensive to assess and therefore not scaleable to large (big brand) sites.
OK. Me, Myself, I, and Our sites already know our history. An outside SEM service coming in cold would certainly need to build a fast efficient (good enough for agency work ;-)) overview. For that your ideas have great merit.
The initial information available could vary significantly depending on client's prior logging and analysis programs. You have my sympathy, Yes, I have no clients today!
I see your ideas as a base to build upon - I certainly would track conversions and after a month or so expect to reconstruct link prior conversion histories.
Automation would be key. Which means careful design consideration especially of possible human subjective input. Build out capability over time.
The measurements you are suggesting are probably more valuable, but extremely labour intensive to assess and therefore not scaleable to large (big brand) sites.
Even the apparently subjective ones can be broadly objectised. Anything objective can be automated. Automation can scale well.
Would you differentiate Slashdot or Digg type traffic spike backlinks? This goes somewhat to your BBC analogy: the quality of a 'story' may be huge but short lived, while a content site reference may be small but constant. Also month to month, year to year traffic comparisons may (likely will) yield very different 'quality' results for any given backlink.
Note: I have run a number of competitors backlinks and isolated some interesting commonalities. Nothing earth shattering but certainly a new perspective. Thank you.
Don't forget the pie charts. Pie charts are second only to profit potential.
