Following the "Periodic Table of SEO" session at SMX Advanced today, was the Mad Scientists of Paid Search session.
SMX's description of the session says, "Our mad scientists emerge from their PPC labs, where they’ve been assembling keywords words, bidding strategies and other assorted parts into monster PPC campaigns. Attend this session to see how they’ve successfully enhanced paid search performance with rigorous testing and trial and error. Their creations are abnormally effective!"
The speakers list was as follows:
Moderator: Matt Van Wagner, President, Find Me Faster (@mvanwagner)
Q&A Moderator: Joseph Kerschbaum, Vice President of Client Services, Clix Marketing (@joekerschbaum)
Michael Behrens, SVP eMarketing, WebMetro (@MichaelBehrens)
Benny Blum, Director, Strategy & Analytics, eSearchVision (@bennyblum)
Ryan Hutchings, Director, Online Marketing, VacationRoost
Matthew Mierzejewski, Vice President of PPC, RKG
As usual, attendees tweeted their own takeaways and highlights from the panel. Here's the best of what was said:
Tip: Use longer keyword (match length) ** Adgroup + Use keyword with highest ad-rank (Quality x Bid)
Queries triggering more than one kw results in chaotic account behavior, data integrity, and accuracy of modeling & bidding
On average multi-mapping happens 10% across accounts
Match Type length Seems to be the Most over ridden rule
Force matching through negative keywords & choose match types carefully
1 query can trigger multiple keywords across adgroups & campaigns, making account control difficult!
Take back control by force matching through negative keywords
You know your business better than Google does (enough said)
Cross serving on a query is a frequent issue.
Use the Google Ad Preview Tool to identify which adgroup/campaign your keywords are matched too. **it's not always accurate though
Solution for wrong matching: insert thousands of negatives
Have a VERY large negative keyword list to force
I'm appreciating the sarcasm of speaker
It's not a bug, it's a feature (said Google)
Just because your ad is showing doesn't mean it's showing for what you want! -E
In most cases there is weak correlation between Quality Score and Conversion Rate
Matthew Mierzejewski from RKG.
Root of Google's CPC decline (it's not Mobile) it's Brand Queries & PLA's
CPC 11% down, because mobile traffic has double (at lower CPC) on last qt.
Avoid Adwords Adrank trumping your match types by granular structures with EXTENSIVE negatives
RKG's says they have seen a 51% increae in Brand query clicks and an 85% decline in CPC
Mobile: Lower engagement, less likely to convert
Tablet: High PPV for impulse and DR / Ecom, High bounce rate for service / software
Purchase intent and engagement varies across devices.
I'll be interested to hear if any Google reps respond officially to some of what was said today
Google rolled out new matching behavior because the system was already doing it..? Interesting thought from the scientists
Question: Best practice. To group Broad Match keywords with Exact Match - or not? Should we isolate kw's based on search demand?
Google incorrectly matches 10% of your queries due to not following their own processes
Microsoft wrote a blog post about the session on its adCenter blog, including a list of tips for "taking back control":
- Use negative keywords
- Force matching through negatives
- Conducive account structure
- Tightly themed ad groups
- Maximize ad group match length rule
- Easier negative implementation
- Choose match types carefully
- Phrase OR Broad (almost duplicate) ?
- Broad OR Broad Match Modified ?
- Remove duplicates/ambiguous
- Maintain campaign budgets
These were from Michael Behrens' presentation. Check out their post for a more complete recap.
Image: Aqua Teen Hunger Force (via YouTube)