Part of my job in managing the marketing and product strategy for AlertRank is watching the Twitter stream for mentions of Google Alerts. Although we aren’t connected with Google in any formal way, I have found myself in the unofficial position of Google Alerts product support. I see what upsets people about Google Alerts, and try to help them solve their problems. Some of the complaints are legitimate problems with the way Google Alerts works, but others are due to the mistaken belief that Google actually understands what the user wants and who they are.
1. The most common complaint is about Gmail treating Google Alerts as spam. Most comments on this recognize the irony of this mistake.
I agree that this is silly. Gmail should automatically assign email from Google Alerts to its whitelist. The fix for this problem is easy. I’ve told lots of people how to fix this, and it always solves the email spam issue. This is a problem that Google should be capable of fixing. Why they don’t is a mystery.
2. The age of some Google Alerts is another big issue. Google Alerts can deliver results that are old, sometimes very old.
There are lots of reasons why this can happen, but the most common cause is splogs that collect other posts from other sites and republish them, so the item Google Alerts is delivering is actually a new copy. Another problem is that there is no clear standard for identifying the date of a web page or item in a feed. There are lots of possible clues, like server modification date, and a variety of feed date formats, but from my experience these are laughably inaccurate. I think this is largely a problem of garbage in, garbage out. But maybe I am being too generous to Google here. This problem occurs often enough, that there should be something they can do to eliminate some of it.
The larger problem, however, is the mistaken belief that Google understands what it is doing in some way. Most of the complaints are phrased as if Google is being stupid, and should know better. Sadly, Google doesn’t know or understand anything. It is a just a computer program that matches text.
3. A related complaint is based on Google Alerts delivering results about the person who created the alert.
This seems to be based on the assumption that Google Alerts should have known that the person who created the alert is also its subject. There are two flaws in this complaint. First of all, Google doesn’t know who you are. It may seem like it, but when you create an alert, there is no way for Google to relate that back to the fact that you also happen to own a particular website that will be found in the alert. The bigger confusion is thinking that Google not only knows everything you do, but that it understands your intent. It is supposed to decide that since you created the alert, you don’t want to know about items it finds that you already know about. So Google not only knows what you know, it knows what you want. I guess some people really do believe that Google is God, or maybe Santa Claus.
4. People also assume that Google does a virus scan of all pages, and recognizes which ones are “dangerous” in some way.
I suppose it would be cool if Google did evaluate all pages for hacks and viruses, but the problem with this is the assumption that the people who created the traps on their pages can’t detect the fact that Google is looking at their page and adjust for that. Google always makes it clear through something called a user agent that it is a Google process that is reading a page. This is easy to detect and program for. So any virus author can make sure that his pages look fine to Google, and only deliver viruses to other visitors. This falls into the Heisenberg problem. The observer changes what is observed.
5. The final complaint is actually pretty funny. The entire tech punditocracy has decided that indexing Twitter in real time is the holy grail, and that Google will fail if it can’t achieve this. Over the last few months, Google has been getting better and better at this, and there are many mentions on Twitter about Google picking up tweets now. Only now a new complaint has appeared. Some people find Google Alerts about Twitter annoying.
This is definitely in the damned if you do, damned if you don’t department. Luckily, this is easy enough to fix. Just add -site:twitter.com to the end of any Google Alerts search term, and tweets will be excluded.
Related Posts
{ 3 comments… read them below or add one }
Google is a computer – yes. But there is also people programming the computer. And since Google is teaching all of us how “important” new content is for our websites to get notices – it’s ironic, and stupid (!) that the Alert function renders results that are months or even years old.
Googlebot presumably frowns upon content not being added within the last week, but Google still sees the need to “Alert” us about stuff that are months in the past? In what way is this not stupid? Or maybe it’s just a crappy tool
I’ve had some of my Google alerts I’ve set up actually disappear. I’ve had to re-add the one I rely on most just today after not receiving alerts for 3 days. Very frustrating!
Last week I stopped receiving all of my 58 google alerts. I have no idea why.
Help. I rely heavily on my alerts.