4 ms·

Display Advertising Works, But It Works Differently Than Search

16 points by terpua 18y ago

DenisM 18y ago
For once a nugget or real research amidst marketing voodoo. Good stuff.
- fredwilson 18y ago
  yeah, but they could really do a lot more with this work, like look at how targeting impacts lift, or how frequency and duration impact it. there's still so much to be learned about online advertising. a lot of the comments to the post are great.
stanley 18y ago
The study backs the type of impact we have seen on many of our campaigns. What's missing, however, is the technology for accurately measuring the various data over a long timespan.
netcan 18y ago
This may be true, but it misses a general point that relates to the first paragraph. Separate the world into two types of good ads: the ones you can see working & the ones a PHD can see working. A pretty girl soliciting people outside your restaurant vs an ad in the paper; Search vs banners. That's not new. That's always been the case. But online, the latter had massive failings & the former had a lot of success. This is in contrast to the experience of the other big media's experiences. For most advertisers the nespaper ad/banner is voodoo. If the average ROI is outstanding, then they might punt. But the other type, the search ad/tourist tout is measurable & reliable & understandable. So in bad times you lose confidence in the things you can't see. The voodoo looses out. But you probably don't fire the tourist tout until things get really bad. Theoretically, for mega brands with mega budgets to do focus groups & hire PHDs to test the effects of their ads over years, there is no difference qualitative between a tourist tout & sponsoring a tennis player. I don't really know, but I think that the theory doesn't pan out. Even campaigns in the 10 of $Ms are really reading these kinds of studies, following 'best practices' & guessing. My evidence for this is that their ad budgets go down in bad times & up in good times. BTW, it runs both ways. Good times make companies spend more on 'brand' advertising. Most Search campaigns are always linked to returns somehow & those don't change that as much as group psychology. So 'brand' advertising is susceptible to bubbles & crashes.

16 points by terpua 18y ago

DenisM 18y ago

For once a nugget or real research amidst marketing voodoo. Good stuff.

fredwilson 18y ago
yeah, but they could really do a lot more with this work, like look at how targeting impacts lift, or how frequency and duration impact it. there's still so much to be learned about online advertising. a lot of the comments to the post are great.

stanley 18y ago

The study backs the type of impact we have seen on many of our campaigns. What's missing, however, is the technology for accurately measuring the various data over a long timespan.

netcan 18y ago

This may be true, but it misses a general point that relates to the first paragraph. Separate the world into two types of good ads: the ones you can see working & the ones a PHD can see working. A pretty girl soliciting people outside your restaurant vs an ad in the paper; Search vs banners. That's not new. That's always been the case. But online, the latter had massive failings & the former had a lot of success. This is in contrast to the experience of the other big media's experiences. For most advertisers the nespaper ad/banner is voodoo. If the average ROI is outstanding, then they might punt. But the other type, the search ad/tourist tout is measurable & reliable & understandable. So in bad times you lose confidence in the things you can't see. The voodoo looses out. But you probably don't fire the tourist tout until things get really bad. Theoretically, for mega brands with mega budgets to do focus groups & hire PHDs to test the effects of their ads over years, there is no difference qualitative between a tourist tout & sponsoring a tennis player. I don't really know, but I think that the theory doesn't pan out. Even campaigns in the 10 of $Ms are really reading these kinds of studies, following 'best practices' & guessing. My evidence for this is that their ad budgets go down in bad times & up in good times. BTW, it runs both ways. Good times make companies spend more on 'brand' advertising. Most Search campaigns are always linked to returns somehow & those don't change that as much as group psychology. So 'brand' advertising is susceptible to bubbles & crashes.