4 ms·
That theory can easily be tested by turning off the ads in question and seeing what actually happens. Can always turn them on again later if needed.
by arbuge 6y ago
That theory can easily be tested by turning off the ads in question and seeing what actually happens. Can always turn them on again later if needed.
- chillacy 6y agoEbay did this 7 years ago, not sure how it turned out though https://hbr.org/2013/03/did-ebay-just-prove-that-paid https://hbr.org/2013/03/did-ebay-just-prove-that-paid
- sosborn 6y agoWould you be willing to make that decision knowing full well that it could reduce revenue that is 100% traceable to that decision? I mean, obviously there are ways to make this politically acceptable in your organization, but you better be completely confident that it won’t bite you in the ass.
- arbuge 6y agoYou can just turn the ads off in one small geography and/or for a short time. Seems to me the potential benefits far outweigh the risks here. If your organization can't see that, then you have bigger problems.
- austinpena 6y agoGoogle Ads also has a built in report for seeing uplift of ads + organic and organic only
- arbuge 6y agoI'd take that report with a big grain of salt and run my own tests independently. Google is more than slightly biased here.
- austinpena 6y agoYou’ve got a point. Plus attribution modeling isn’t for the faint of heart. One popular method is a geolocation test. Turn off the test campaign (usually remarketing) in certain geos and see what happens.
- arbuge 6y agoExactly what I suggested earlier... see my other comment reply.