This is the continuation of yesterday’s post featuring my deadblogging coverage of SphinnCon, the recent SMX-affiliated 1 day conference here in Jerusalem, Israel (where I’m currently living).
Ophir Cohen – Compucall Web Marketing
Shares Perry Marshall’s formula for PPC ad copy:
Dynamic Keyword Insertion (aka DKI) for the title
2nd line: Benefit
3rd line: Feature [I wonder where/if there's a call to action somewhere?]
Display: Brand or subdomain or folder.
Ophir: DKI bypasses Tm filters.
- To give greater % of impressions to one ad (eg a new split test to try and improve on a previous winner, without losing 50% of impressions to the expected loser), just make multiple copies of one ad.
- Use capitalization in the URL as a best practice.
- Price, numbers and signs boost CTR… and conversion rate!
- Free shipping / good offers boost CTR.
Dan Perach – PPC Proz PPC Management
Learned from Ophir at Compucall. Kudos! [Classy start!]
A/B testing is something to do consciously. Not just because it’s a best practice. Dan spends 1 hour optimizing ads in just 2-3 ad groups
“It’s not enough to come up with a new ad [at random] and start running it, because 50% of your impressions are then running a loser.” In other words, don’t just test a new ad/lander for the sake of varying – come up with a hypothesis as to what ad/landing page combination will outperform your current winner. [This goes back to what Ophir was saying about writing ad copy highlighting benefits, features, prices, offers etc.]
For a new split test, Dan will pause the current winning ad.
He clones the existing winner twice, and edits 1 of the new duplicates. This is to avoid the existing winner’s quality score and history affecting the test (presumably getting better position/cheaper).
Later in the preso, Dan also pointed out that for a true split test, you need both ads to have been running for the same amount of time. [And presumably during the same time period to adjust for day-parting, weekday vs weekend, and seasonal differences.]
A setting Dan likes to use in contrast to many others is to set AdWords to optimize to find CTR winners.
To the same effect, Dan user a tool called WinnerAlert.com, which syncs to AdWords and then sends you an email when there’s a statistically valid winner between your ads. [I think winners in this context still means winning on CTR, not conversion rate.]
Ad approval takes 2-3 days [I think he means 2-3 ordinary days, not business days as defined by Western schedules – Monday to Friday.] In B2B segments, start on a Thursday or Friday so you only lose 1 business day. [I think the Thursday reference is perhaps to Israeli B2B companies since the Israeli business week is Sunday – Thursday.]
For local analysis of different ads, Dan uses Google Global Local (I think that’s Dave Davis’ Firefox extension he’s referring to, or else I just didn’t write that down properly).
Another tool he uses for competitor analysis is Keyword Spy, which shows ads displaying most. Dan then analyzes the features and benefits most prevalent and aggregates the data. This is better than spot-checking (eg manual checking ‘on the spot’), according to Dan, because spot-checking is [more] limited by day parting and other methods of ad targeting.
For composing new ads, Dan follows a simple formula.
Figure out a unique feature or benefit, and add that to a call to action. He creates a variety of these in excel, uses excel’s tools to mix-and-match, then copies to AdWords Editor to upload. He doesn’t find space for both a benefit and a feature.
If you enjoyed this post, come back tomorrow for coverage of SphinnCon’s content network advertising tips (those Ads by Google you see on so many sites) and more . Or more easily, just add my rss feed to your reader for more of this good stuff (that link also lets you subscribe by email).ppc, SMX