Testing creative messaging across channels
These are all front-end applications. At the back end, there are exciting opportunities for us to optimize the overall campaign based on better performing channels.
Many marketers often use search as a starting point to test different creatives for other channels. It is relatively easy to rotate different creatives at an adgroup level in most of the major search engines. Keeping the end goal in mind, based on either the click-through rate or conversion rate, we can see which creative appeals to prospective users. Once we have the winning creative, we can use that in other channels.
The reverse is also true. If you see search performance trending down, or being stagnant, but other channels (such as online display) doing well, then you might want to look into what the user experience is on the more successful channel — both creative and landing page — and replicate that in your search campaign.
We recently worked on a campaign for one of our clients in the prepaid wireless industry. For branded terms, search campaign creative was solely focused around the brand and we drove people directly to the brand's home page. In contrast, the display advertising component had a very offer-specific creative running. When people clicked on the ad, they were taken to these specific offer pages.
We orchestrated a test for search, where we ran three creatives for branded terms. The first focused on brand, which dropped the user to the brand's homepage. The second and third creatives were specific to the two offers that the client was offering at that point, and users were taken to the respective landing pages.
In the test results, we found that conversion rates on branded terms from these offer-specific pages were 50% higher than the conversion rate from the brand-specific messaging.
When thinking about cross-channel integration from a testing standpoint, first set client expectations and parameters. Define the test parameters for the client (examples include time, budget, keywords and key performance indicators) and also what the trade-off will be during the test period. For example, if it is a creative test, which involves different landing pages, then conversion data will only be available at the creative and not keyword level (unless each keyword is in its own adgroup).
Also, start small. Identify only a small set of keywords to run the test on. A perfect set of keywords will have moderate to high levels of activities. Run the test for at least two weeks. This will help level any weekly fluctuations, and will be enough time to gather statistically significant data.
Finally, be open-minded. Be curious about how other channels are performing. Look at integrated reports that are being sent to the client. Unless we know how the performance of other channels stacks up against search performance, we will not know what the testing opportunities are.