Now, I know n=1 is not a good sample size, but to downplay the significance of kidnappings by "someone the child knows" or that "it was a family member" is not reasonable.
Also, a study on the effectiveness of amber alert messages on the freeway be extrapolated to give data on results from a 404 campaign.
I can't estimate whether or not internet amber alerts are a good or a bad idea. Just tossing in my 0.02.