|
|
|
b. The number of pages served should be thought of as a minimum, because statistics do not count re-readings or re-servings from local computer memory. Nor do they count sharing of pages once downloaded.
c. Most but not all graphics served are excluded from the figures to cut back on increasing "page" averages just by calling up various graphics. Without specifically programming this, a casual analyst will overcount web traffic thanks to graphics.
2. Tracking where computers are from is bound to have some error as many wise users intentionally surf the web and leave a false trail. On the other hand, we doubt that such error means that non-U.$. readership is lower than reported, given that we do not count .com, .net and .edu readership.
|
|