We all know the Pay Per Click (PPC) basics such as setting up your campaigns and ad groups correctly, keyword expansion, ad text best practices, and how to bid to a positive ROI, so in this advanced session we were taught how to multivariate test elements in campaigns, tap into long tail terms, use dynamic keyword insertion properly, and manage portfolios of keywords.
The first of four speakers was Eduardo Llach, Founder and COO of SearchRev. Eduardo was so hurried because each speaker had 10 minutes to present, therefore it made it hard to follow all that he was saying. And he had some great tips to share! His key message was the idea of multivariate testing elements in campaigns. He has developed a science around testing multiple elements such as syndication, geo targeting, keyword match type, and day parting down to the keyword itself. He gave examples of each:
1. Syndication: testing the conversion rate of content syndication against the other elements. Does content match perform better in other geographic areas?
2. Geo targeting: testing geographic locations. He mentioned that country mapping is very accurate, state mapping is 50% accurate and city mapping is even less accurate, possibly down to 25%. So watch those city campaigns closely. He also suggested grouping metro areas together to improve mapping accuracy.
3. Keyword match type: using all match types in campaigns. Does phrase match outperform exact for a particular keyword?
4. Day parting: bidding up or down during a particular time period. So if on Monday his click rate goes up 40%, he bids up 40%. He runs his campaigns in shifts such as Monday am till 4pm and 4pm till midnight. He said with day parting you can learn on Google and apply to other networks.
Next Jon Kelly, President of SureHits, talked about long tail keywords and geo targeting. He reminded us of why we should care about the long tail: small volume adds up, clear intention equals better conversion, and less competition equals lower bids. He took us through three action steps in order to optimize campaigns for long tail keywords:
1. Calculate the click value: for a low volume phrase the click value is the probability of conversion and value of that conversion. Look at the value of the a�?heada�? term versus the tail. For home loan, the click value may be $7, while the long tail term could be worth $15.
2. Reward the usera��s choices: customize landing pages as you will be rewarded with a higher conversion rate.
3. Watch your campaign data: look for outliers in the data. He reminded us that city phrases can also be brand phrases. Example: Portland Home Loans might be a company name. Bidding on such brand phrases yields a lower conversion rate since the user might be looking for the brand instead of a generic search. Also beware of homonyms. Example: Mobile home loans a�� Mobile, Alabama, and Mobile homes. Lastly beware of geo targeting a�� he gave an example of a campaign he was running in Texas that had a horrible conversion rate and it was because Houstona��s mapping wasna��t working correctly. Conversion rates increased 20% after he took out Houston.
Matt Van Wagner, President of Find Me Faster taught us that DKI ads can be a good idea if used in the right way. He gave some hilarious examples of DKI ads a�?gone wilda�?. I wona��t go into them here but see for yourself by searching: a�?used fisha�?, a�?used cigarsa�?, or anything useda��DKI is a great way to customize your ads and save yourself some time. They are excellent when you want an abbreviation, such as a state name, or correct spelling of your company name to appear in your ad. Wea��ve used them for years but never set them and forget about them a�� they must be watched and tested. All the main engines use a form of DKI a�� even Ask! His parting words were to test in ad groups where keywords are very focused and related, and only with phrase or exact match, never broad.
Michael Sack, Director, SEM Techology and Development for Idearc Media took us through the idea of managing campaigns like a stock broker would manage stock portfolios. He said ita��s best to use day parting when you are in a competitive market, have limited sales windows, identifiable demographics by time, can detect patterns, and you have the business capabilities to track and manage campaigns to this level. All the main analytic platforms have this capability, so just slowly start testing. He also told us about his Hi-Lo optimization theory which is based on analysis that suggests most phrases have a performance window and that you should focus on the hi periods and bid up during this time.
As for managing keywords in portfolios, the strategy is to minimize keyword management since campaigns are growing larger each day. By doing so, you minimize the risk of market fluctuations. As you can see, he thinks of keywords like investments: they have a cost, rate of return, there is a risk, performance can be tracked, and market behavior can be monitored. You must set the rules and objectives, measure the performance, and diversify.