Over the years, the The Google Display Network (GDN) has been knocked for poor performance. Advertisers have long complained about low-quality placements and low conversion rates.
To be sure, performance on the GDN isn't the same as search, in that the ads are shown on content sites and matched based on content and topics, rather than keywords. Advertisers should not think of display as an extension of search, but rather as a distinct channel.
Still, low conversion rates are often a reality. But low conversion rates on the GDN alone can be deceiving. In fact, the GDN can boost performance in your search PPC campaigns. Here's a case study illustrating how:

Background

Our client had recently launched a new suite of products and services in the financial services industry, and wanted to generate leads. We ran a GDN campaign for 12 weeks, using both text and image ads. Campaigns were targeted based on keywords and in-market segments. While we knew display would act primarily as an awareness play, the primary goal was still lead generation.

Results

The results for the display campaign weren't great on their own. In the 12 weeks the campaign was live, we only got 10 conversions:
gdnchart2
The campaign generated a lot of impressions and clicks, but precious few leads. On the surface, it looks like a failure.
Let's look at results for search, which ran for a total of 23 weeks: nine before display launched, and two after it was paused.
conversionsbyweek
Search conversions peaked when display launched, and fell off after display ended. And conversion rates decreased after display was paused.
Here's a closer look at weekly averages before display launched, while it was running, and after it was paused:
gdnchart1
Impressions increased by an average of 45 percent while display was running, and fell off when display was paused to about the same level they were before display. This indicates that more people performed searches while the display campaign was running. It's possible that the display campaigns influenced more people to search for the client's product after seeing a display ad.
Conversion rate also increased dramatically while display was running - by 52 percent. The display campaign may have attracted searchers who were more qualified, and converted at a higher rate. Even more interesting is the fact that conversion rates fell significantly when display was paused, even though the search campaign had been optimized heavily by this point.

Confounding Factors

Of course, no longitudinal case study like this is without confounding factors. Seasonality could have played a role in influencing conversion rates. The campaign ran from January to June, and display was live from mid-March through May. It's possible that the second quarter of the year is just naturally the busiest time for this client. We don't know, since this was the first time they'd used PPC, but it's definitely possible.
Also, as mentioned earlier, we were constantly optimizing the search campaign. The heaviest optimization was done during the first few weeks - before display launched. Would conversion rates in the search campaign have risen anyway, just due to optimization? It's possible, but the fact that search conversion rates fell off when display was paused leads me to believe that display was part of the impact.

Conclusion

If we had looked at display on its own, we would easily have decided it was a failed experiment. A 0.12 percent conversion rate is low by anyone's standards, and the cost per lead was high. When combined with a search campaign, though, overall results improved greatly. Without display, we would have had fewer overall search conversions, based on the data. Display appears to have positively impacted the search campaign.
The next time you're thinking about pausing a low-performing display campaign, look carefully at its impact on your search campaigns. You may be pleasantly surprised.