TQR — Metasearch Puppy Piles vs. Lone Wolf Search Engines

Posted bylara

All search engines are not the same and all search engines do not return the same results. Ask each of the big four search engines the same ques­tion and you’ll get four very dif­fer­ent sets of answers. Well duh… 

But just how dif­fer­ent are the results? Different Engines, Different Results, pub­lished in April by the good folks at Dogpile gives some hard numbers.

Over the set of all returned first page results 88.3% of first page results are unique to the search engine that returned them and only 1.1% of the first page results turn up in all four engines.

This is a great study for fer­ret­ing out lit­tle nuggets that will astound and amaze the naive searchers. Even the DH who has to lis­ten to me bab­ble on about this stuff and who, being an ubergeek him­self, ought to know bet­ter thought that there was­n’t much dif­fer­ence in the results of one SE vers­es another.

Another favorite num­ber is from Figure 2.

Google’s index­ing cov­er­age of the “Total Web”(1) is giv­en as 69.6%.

Do the math. Google does­n’t cov­er 30.1% of the Web.

That said the study is prone to a bit of an apples and oranges con­fu­sion. Sometimes they sep­a­rate out spon­sored and non-sponsored results and oth­er times they don’t.(2)

Cum gra­no salis. Though they cred­it researchers at Queensland U. and Penn State this research is in aid of a par­tic­u­lar com­mer­cial ven­ture — Dogpile. (One of my fav agre­ga­tors, though.)

(1) Actually the term used in the fig­ure “Total Web” is mis­lead­ing. It should be “Total Visible (to Search Engines) Web” There’s a whole ‘nother web out there that the search engines can’t see. The authors get it right in the text.

(2) For exam­ple in Figure 5 they list the per­cent­ages of first page results (of those returned by all four SEs) that you would miss if you had only used one SE. (All around 70%.) But they don’t call out spon­sored vs. non spon­sored results. So what? So… spon­sored results can only appear in results pool of the SEs that the spon­sors have cho­sen to pay for. By def­i­n­i­tion, there are cer­tain results that can not appear in mul­ti­ple search engines.

(Why is all the good stuff in the footnotes?)