Companies spend billions of dollars online for paid links to branded search terms. Measuring the effectiveness of this marketing spending is hard. Blake, Nosko and Tadelis (2015) ran an experiment with eBay, showing that when the company suspended paid search, most of the traffic still ended up on its website. Can findings from one of the largest companies in the world be generalized? We conducted a similar experiment with Edmunds.com, arguably a more representative company, and found starkly different results. More than half of the paid traffic is lost when we shut off paid-links search. These results suggest money spent on search-engine marketing may be more effective than previously documented.