Online Monitoring Cyborg Style

The magical formula for online monitoring is determining who has "influence." If you want to advertise on a blog, do you go for traffic, or do you look for someone who will "influence" the way other bloggers and the audience to take action? Traffic is easy to measure - you decide on uniques, pageviews, length of visit, and demographics, and buy an ad.

Influence is harder to measure. You have to really do the homework, trust your instincts, and filter the link-baiters from the loyal, trusting audience builders.

So far, no one has done it. The problem is one of complexity. Considering the size of the web, and the structure of websites, and the lack of proper data, it's near impossible to write an algorithm that can shuffle through the internet and define the variables that equal influence.

Many companies sell this a pitch, but an overload of data and a lack of comprehension of what that data means leads the smart companies to mix the human ability of pattern recognition with the hard number crunching of computers.

Nathan Gilliatt, who is composing a guide to the online monitoring softwares around the world, writes about the different levels of human-computer interaction in the field, and, he even provides a nifty illustration. A commenter on his post asks the following instructive question.

What is the real signal-to-noise ratio out there in these consumer exchanges, and how is that being addressed? Where does the automation start and stop with regard to this source identification process? And then of course, the larger question, what is the current state of the balance between Automation and Accuracy in digitally-directed research?

The answer, is its being addressed poorly. There are three stages to gathering information effectively, and I'll label them as initial data, detailing, and analysis.

The human factor in each of these stages is necessary to prevent bad data from corrupting the entire process, and so far, it's my belief that a human brain is better at determining what is influential and what is garbage.

• Influential: Good writing, positive conversation flows, high-interaction communities
• Garbage: Splogs, blatantly commercial sites, untargeted sites, dangerous sites (NSFW, language, or charged political sites).

Computer Assisted Human Filtering is still the best bet for accuracy, but it's difficult to price correctly. Online Monitoring can be as expensive as a $20,000 setup fee and $10,000 a month for reports. If the technology exist to knock this report down to a few hours of work, do you charge just for the computer time, or do you separate the costs like the larger vendors do, charging extra for analysis of the data?

The real question is a simple one. Who do I need to engage to promote, defend and enhance my brand? And what are you going to charge me to do the work for me?

Do you base the fee on the hours it would take a human to compile the information, o