« Measuring Transparency | Main | Great post on measuring ROI of Events »

March 24, 2009


David Phillips

Ah, Katie, you fall into the evaluation trap.

As I made quite clear, this is not a measure of AVE's or anything remotely like it.

It is a software tool that is designed to looks at hypothetical value markets and exchange rates as part of a much bigger research programme.

Long ago (20 years?), you may remember that I was one of the people that fought against the idea of AVE's and Media Measurement, when I was its owner and MD, did not and would not offer AVE's for all the reasons that have been put forward in more recent years by people like yourself.

I guess you will have read my book (published in 1995) which gave voice to my opposition to the silly idea of using advertising as a measure of editorial.

All that the software I am developing does is to find broad data that will give indicative data on advertising value opportunity for reach of web pages.

Alexa provides a useful API to help identify reach and comparative data using online reach of known and audited reach-data which I use to create an algorithm to normalise reach data. As I did explain in my post, I have worked the numbers back to find out how close the algorithm can get to normalised reach and to normalised value.

Having done that, I have worked on the extent to which reach, in other forms of online presence, provides exposure that is comparable between sites using advertising and sites that do not.

Fortunately we can test some types of web site using the bid price of advertising using services from organisations such as MSN Labs and (less helpfully) Google.

There is a lot more work to be done but results are not so far out that such data is not unusable (and the compound effect is within tolerable levels).

From this we are beginning to see that in advertising there are two markets and that they have a recognisable exchange rate between the two different forms of advertising value.

This gives us a potential approximate advertising opportunity cost for (some - not all) web pages.

I claim absolutely no more. So far, the results seem good and no doubts the online community, when I eventually publish the widget - hopefully this week - will identify flaws.

Now for the next part of the research.

Knowing that there is the potential for reach to be an element in value (not $ value but a value based on number n), we can look at proofs of reach as being of significance.

Now, lets look at your example as to how much a mention on influential blogs like Shel Israel's blog is worth, or the value of a Tweet by GeekMommy because they don't have published advertising rates.

No they don't but they do have something that might be called 'n' that could be a 'wow' value.

So, put away your measurement in terms of $ and start thinking about something called 'wow'.

In our research we are looking at different forms of wow which I explained in my paper in JCM called 'Towards Relationship Management'.

We find from that paper that there are Tokens that have values (lets call them 'wow's for the sake argument).

The work of Bruno Amaral this year is showing us that we can use latent semantics in texts, word densities and hyperlinks to triangulate online content in identifying tokens and values (wow's) in online social (network) groups.

I am curious to see if reach and/or advertising opportunity cost can offer us an ability to develop an idea about the value of tokens to and in groups.

To do this we need software (Bruno's Masters research used A Latent Semantic Analysis engine I have) because the sample sizes are absolutely huge.

As always, I have made the software available to the community and blogged about it.

It does not matter to me if the PR evaluation industry has some strange application for it and if its some form of AVE measurement - so be it. Such work has a market and the companies involved have to justify their products to customers.

Such applications have nothing to do with my research.

However, for researchers, such a tool may have value for the work they are doing which is a different thing.

One thing that this form of research tells us is that it is all too simplistic to imagine that SEO-PR generated links represent x number of sales. In the case study you refer to, the Golden Ruler Award for Excellence in Public Relations Measurement
“You Are Now Free To Link PR and Sales”, the range of factors (variables) are too big for such a generic claim.

Its rather like claiming that editorial is worth three time the advertising equivalent (which Burston found was right for a specific campaign 30 years ago).

I might be a little more convinced if you can replicate the data when you try it for the KDPaine's blog.


The comments to this entry are closed.