Compete.com purports to report traffic statistics for any web site you specify. The problem is, their web stats are estimates, derived from the browsing habits of those Compete.com members who volunteer to download a monitoring application. And any time you derive data from a subset of your universe, the data will be inaccurate to some degree.
I was introduced to Compete.com by a potential advertiser to my newsletter, who showed me a Compare.com analysis of my web site ... where the numbers were about a tenth of the actual, bot-free numbers reported by my hosting company and ad-serving company ... and where they ignored completely our companion directory site, on which our clients' banner ads also appear.
Thanks to Compete.com, potential advertisers can get data on your radio station's web site without the expense of subscribing to a reputable paid service (like Arbitron Online Radio Services) or the bother of contacting you directly. Potential advertisers probably think the data are accurate enough. This hurts them, because they make buying decisions based on bad data; and it hurts you big-time - especially since even the busiest local radio sites have traffic that pales in comparison with national and worldwide sites.
Just for fun I went to Compete.com and plugged in the web sites from some top-rated New York stations. In most cases, Compete.com issued this warning: "We have little data for wxxx.com, so these are rough estimates." In other words, even a high-traffic major-market radio site has a small audience in Internet terms.
Myriad other services, including radio's own Arbitron, measure Internet traffic by means of statistics. But why, when exact usage data are available from the every web site's hosting company, do we use inherently-inaccurate estimates?
There are two concerns with the "exact" usage data:
- It can easily be inflated by the use of automated "bots" to hit a web site again and again.
- Web site representatives - whether inadvertently or on purpose - often misrepresent the data, confusing "hits," "page views, "visitors" and so on.
I have some suggestions to bring credibility and reliability to radio web site statistics:
- Establish a web site that has statistics for every radio station in the U.S.
- Derive the statistics from a formula including the stations' own verified server-generated numbers, combined with estimates from a reputable survey company (like Arbitron).
- Present all stations' statistics in a consistent format.
- Include definitions of all parameters presented ("hits," "page views," etc.).
At first, the task of developing such a site seems daunting, but once the protocols are established it should be manageable. Further, I submit that such a site, if not absolutely necessary, is vital to make online radio easy to evaluate and to buy.
Over-the-air radio is notoriously difficult to buy, primarily because there is no such coordinated industry-wide effort. We have the opportunity to do it right on the web ... and if it works, maybe we'll be motivated to replicate that success back here on Earth.