3 ms·

I'm tired of people conceding that Alexa is next to worthless and then going on to use an Alexa graph to make a point. It doesn't work that way. It can't be eit

by jm4 17y ago

I'm tired of people conceding that Alexa is next to worthless and then going on to use an Alexa graph to make a point. It doesn't work that way. It can't be either unreliable or of some small value depending on when it's convenient. Extrapolation from too little data is fraught with peril. The author of this article made this mistake. How is your argument any different? The Alexa stats you cite are an even smaller data set than what is used in the article! I do agree that we just need to wait this out. Even with good stats there's no way to know whether Bing is a worthy competitor. It's obvious that Bing hasn't been a complete flop in the way Cuil was. So far it's good enough that people are going to try it out for a while, but what's really important is whether they're still using it in 6 months or a year.

jacquesm 17y ago
Alexa is good for relative figures and larger trends, not for detail. The 'compete' graph has exactly 3 data points, the 'alexa' one has 30 and shows a completely different graph (and imho a more realistic one). The lack of accounting for the influx of MSN traffic is a major omission, so my point is - and I think this is a big difference from the original article - that there is better data available than that used in the article which would suggest a different conclusion and that no matter what source of data you use you have to be careful to extrapolate at this point in time. Better like that ?
- jm4 17y ago
  I agree with most of what you're saying. MSN definitely should have been taken into account and it's way to early to call Bing a success. Where I disagree is with the criticism of the data used for the article and then use of equally unreliable data to prove the point. The Alexa data has more points, but it's looking at the same period of time so I'm not buying into the idea that it's more realistic. Not to mention day to day numbers are a microscopic view of a much larger picture. We've got two measurements of the same thing that seem to be drawing two different conclusions. I don't think anyone could reliably come to the conclusion that one is more accurate than the other based on the information we have.
  - jacquesm 17y ago
    A common trick to show an upward trend is to show just the first and last point of something that has gone up in a certain period without showing the data in between. Like that it always looks like you are going 'up', no matter what. If you look at the datapoints in between you get to see more detail, in this case a meteoric rise over the first couple of days and then a fairly steady decline, which is a more realistic picture. So, in this respect I'm more inclined to believe alexa.com over compete, not because they give me a story that I'd like to hear but simply because their data offers more transparency. In the long run (over many month) there will be more convergence between the shapes of the graphs, but for the moment alexa gives a more realistic image to extrapolate from, simply because it shows you more intermediate points. If alexa had just shown the first and last datapoints then the graph would have been the same but the conclusion would have been completely different. Of course you're entirely free to disagree with that :) For the record, I use my own site (ww.com) as a benchmark whenever I look at other sites to try to figure out the bias of the reporting agency but with sites as large as msn, google and bing that is not a useful technique because my own site is so much smaller that it is simply a flat line.

by jm4 17y ago

jacquesm 17y ago

Alexa is good for relative figures and larger trends, not for detail. The 'compete' graph has exactly 3 data points, the 'alexa' one has 30 and shows a completely different graph (and imho a more realistic one). The lack of accounting for the influx of MSN traffic is a major omission, so my point is - and I think this is a big difference from the original article - that there is better data available than that used in the article which would suggest a different conclusion and that no matter what source of data you use you have to be careful to extrapolate at this point in time. Better like that ?

jm4 17y ago
I agree with most of what you're saying. MSN definitely should have been taken into account and it's way to early to call Bing a success. Where I disagree is with the criticism of the data used for the article and then use of equally unreliable data to prove the point. The Alexa data has more points, but it's looking at the same period of time so I'm not buying into the idea that it's more realistic. Not to mention day to day numbers are a microscopic view of a much larger picture. We've got two measurements of the same thing that seem to be drawing two different conclusions. I don't think anyone could reliably come to the conclusion that one is more accurate than the other based on the information we have.
- jacquesm 17y ago
  A common trick to show an upward trend is to show just the first and last point of something that has gone up in a certain period without showing the data in between. Like that it always looks like you are going 'up', no matter what. If you look at the datapoints in between you get to see more detail, in this case a meteoric rise over the first couple of days and then a fairly steady decline, which is a more realistic picture. So, in this respect I'm more inclined to believe alexa.com over compete, not because they give me a story that I'd like to hear but simply because their data offers more transparency. In the long run (over many month) there will be more convergence between the shapes of the graphs, but for the moment alexa gives a more realistic image to extrapolate from, simply because it shows you more intermediate points. If alexa had just shown the first and last datapoints then the graph would have been the same but the conclusion would have been completely different. Of course you're entirely free to disagree with that :) For the record, I use my own site (ww.com) as a benchmark whenever I look at other sites to try to figure out the bias of the reporting agency but with sites as large as msn, google and bing that is not a useful technique because my own site is so much smaller that it is simply a flat line.