malcolm coles

Google ran an interesting experiment - attempting to predict the results of the 2009 Eurovision contest based on search popularity.

But while it got the winner right, the rest of its predictions were a bit poor, as this table shows.

Actual results vs Google predictions

Nick Burcher had pointed out two reasons in advance why this experiment wasn't likely to be all that successful: "the jury vote [which accounts for half of the votes] can't be predicted and significant search volumes in Eastern Europe occur on non-Google search engines like Yandex".

And while he updated his post to claim "Norway wins a landslide Eurovision victory, Google gets it right!", I think that's pretty kind - Google predicted a narrow win over Turkey in 2nd, as opposed to a thumping one over Iceland in 2nd.

Interestingly, Norway were red hot favourites at the bookies - with Greece 2nd favourites and Turkey 3rd. So they're no better at predicting it either ...

Still, one thing you could predict was that Andorra gave Spain 12 points - they had 0 until then. So some things never change:

• How does Europe Make Its Mind Up? Connections, cliques, and compatibility between countries in the Eurovision Song Contest
• Comparison of Eurovision Song Contest Simulation with Actual Results Reveals Shifting Patterns of Collusive Voting Alliances.
• The maths of Eurovision voting

Screenshot of Google predictor

Google's Eurovision predictions based on search volumes

You might also like

1. Google's European election predictions based on wrong search terms
2. How Google foresaw BNP election wins
3. New google sprite: July 2009
4. New Google sprite: September 2009
5. I'm a Celebrity 2009: ITV neglects its SEO