Toolbox Google activity feedback
Written by: Jonathan Briggs
October 8, 2008 [3182 views]
A big thanks to everyone who completed this activity. My goal is to give you rapid feedback as a group so that you can use your new knowledge in the rest of the module and elsewhere.
As a general comment I would point out that many of you left this to the last day. That’s fine but particularly for the next activity I would urge you to get started sooner.
Now lets go through some of the answers:
Part 1
Numbers of searches for Kanye West?
These ranged from under 4,000,000 to over 7,000,000. I wanted you to realise that Google is only able to estimate the number of pages about any topic. There is also evidence that you are actually accessing a variety of different Googles because a single machine could not cope with all of the requests. Keeping all of these in sync is simply a huge task so there is variety in the answers. The key message here is that Google does a very complex job very well but search is not a precise science and variations occur. I want you to keep that complexity in mind when building and designing your own systems.
Why did you think the US and UK results were different?
We have already seen that there is variation between searches on the same Google but you clearly noticed that the different countries give different results too. Part of the reason is the same as the one above. Other reasons may be much more complex – it is suggested that Google prioritises sites that are popular within a particular country (links from sites within that country). Google is highly secretive about its exact algorithms and that is one of the lessons I wanted you to learn.
Take a look at the theories of others by Googling “different results in UK and US Google”
But, a clear lesson to this sort of question is just that. If you don’t know something then ask Google to see what others think. The trick is to ask good questions but I find answers for most of the technical and business problems I have (or at least a discussion).
Who is advertising alongside Kanye West?
We will return to advertising in the future but I just wanted to make sure that you saw the ads and thought a little about how and why they are there. Google allows companies to bid on keywords and the adverts shown have all bid to be associated with the term “Kanye West”. The interesting thing (for the industry) is how easy it is to be associated with specific terms.
What was special about results beyond 1000?
This confused many of you because you thought it must be a trick question. Once again I want you to think about the complexity of the search problem and see that Google has simplified the problem by only showing the first few pages of results. Some of you suggested that the pages beyond 1000 would be less relevant and that’s what Google thinks too but we cannot judge this for ourselves. For popular topics this is probably not true.
Why do you think this is so?
Some of you believed Google’s message about the rest of the 4,000,000 pages being similar. This is nonsense; they may be less relevant but they are not likely to be the same. Google is a software product that has been designed to meet a market need. It simply does not have to store the rest of the results because very few people would ever use them (and that saves them money).
How could you search for pages published in the last six months?
This was the hardest question and many of you simply could not find an answer. I wonder how many of you tried to Google the answer? “daterange search query Google”. Here you will find techniques that allow you to use the search query to find exactly what you want.
Questions raised
- Is Google invading our privacy?
- Why are the questions in the activity so open ended?
- Will I lose marks for the “wrong” answers?
- What happens if I am a day late with my answers?
- Can we have normal lectures that just tell us what we need to know?
- Can we have individual feedback?
- Can we work in groups on the activities?
Part 2
I will discuss some of the multiple choice questions in the lecture but as some people have still to do the activity, I am not going to give the answers here.
How does Google work?
You had a good go at producing “friendly explanations” of Google and many of you pick up some of the key ideas. These include the role of a spider to visit pages online to create an index that is then searched when a user types in a query. More sophisticated answers include trying to measure pages for “relevance” (to keywords) and then sorting the results by “reputation” (PageRank). It is really important that you appreciate the computational complexity of all of these tasks.
Better than Google?
Some sensible suggestions but also a realisation of how difficult it might be. Faster indexing, better understanding of language, putting back human reviews, more advanced search options and greater focus on the personal needs of the searcher were all good ideas.
Questions about Google
- How did someone think of the idea?
- What do we know about the technical architecture?
- How did they become so popular?
- How much is Google worth and what are staff paid?
- Can Google see the stuff on my computer?
- Does Google read my Gmail?
- How does Google make money if most of its services are free?
- How does Google find new web pages?
- How does Google work so quickly?
- Could we design a search engine?
- How much of the web does Google search?
- Why is Google images not better?
- Whats in the pipeline?
What do you think?