We create best practices around what works for most accounts. However, there are times you should abandon best practices. You might due this due to resource constraints, simplifying complex problems, or just failures of a best practice to account for your specific situation. In this presentation, you will look at some bidding and attribution management scenarios where abandoning best practices improved the account’s performance.
3. Answers:
1. It’s amazing - I use it all the time
2. I use it sparingly
3. It’s terrible - modified broad is the way to go
@bgTheory
Poll: Who things Broad Match is Wonderful?
AdAlysis
5. 1. Simplifying complex problems
2. Resource constraints / Non-Defined Processes
3. Failures of a best practice due to exceptions
@bgTheory
Why to They Fail?
AdAlysis
9. @bgTheory
Funnel: Generic > Specific > Brand
0
200
400
600
800
1000
1200
Number Spend (10s) Conversions Assisted Conversions
Keywords with conversions Keywords without Conversions
Article on managing assisted conversions: http://adalysis.com/blog/3-ways-to-handle-assisted-co
AdAlysis
10. •Best Practice (debatable):
Choose all languages your consumer speaks
Don’t mix languages (ads to landing pages)
Only choose languages you support
Language targeting is a minimal setting; don’t optimize for it
Note: no stats from AdWords/Bing on conversions by language
@bgTheory
Language Targeting
AdAlysis
11. @bgTheory
Local Philly Company: Just Started Adding
Languages
Overall CPA Conversions / Month
English $56 192 (baseline)
Russian $55 197 (+5)
Korean $52 203 (+6)
Chinese $55 204 (+1)
Italian (China removed) $53 206 (+3)
Polish $54 207 (+1)
Spanish $59 211 (+4)
More added… $54 228
Final Tally of 11 Languages $-2 / conversion +36 (16% Lift)
AdAlysis
13. @bgTheory
Travel Company: Created Language / Country
Combination Campaigns
AdAlysis
Country Language CTR Conv Rate
Italy Italian 3.1% 2.1%
Italy English 3.9% 7.9%
Italy German 4.1% 8.1%
Germany German 2.9% 3.1%
Germany Italian 4.2% 7.6%
German English 5.6% 7.7%
14. @bgTheory
Optimizing Modifiers To Conversions
0
20
40
60
80
100
120
140
160
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Calls Per Day
AdAlysis
15. @bgTheory
Applied Modifiers Based on Conversion Rates
0
20
40
60
80
100
120
140
160
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Calls Per Day
Pre Modifiers After Modifiers
AdAlysis
16. @bgTheory
Sunday Evening Was Their Best Research Time
0
200
400
600
800
1000
1200
Sunday Monday Tuesday Wednesday Thursday Friday Saturday
Research vs Action
Company Details Found Calls
AdAlysis
17. @bgTheory
Best Practice: Bid to ROAS/ROI for Ecommerce
0
100
200
300
400
500
600
700
800
Ad Hoc Bidding ROAS Bidding
ROAS Orders/Month
AdAlysis
18. @bgTheory
The Complicated Data
Overs over
$10,000
Orders over
$500
Orders under
$500
Orders 102 1098 4502
Unique Keywords
Leading to Orders
87 119 245
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
How Predictable Are the Large Orders?
AdAlysis
20. @bgTheory
Location Targeting
2011: Company examined credit card zip codes
Only targeted high revenue zip codes
0
50
100
150
200
250
2011 2014
Conversions / Month (Spring/Summer)
CPA
AdAlysis
21. •Location targeting is more accurate
•More people are searching on phones
•DC is:
Worst commuting traffic
2rd highest public transit rates
@bgTheory
Three Problems
AdAlysis
22. @bgTheory
Added Most Work Areas: Conversion vs CPA
Balance
0
50
100
150
200
250
300
350
400
450
500
2011 2014 2015
Conversions / Month
(Spring/Summer)
CPA
AdAlysis
25. •Airline company (rounded data to protect company info)
• Fly from 40 cities to 40 cities
• Cities encompass 6 countries & 5 languages
•How do you create this account?
@bgTheory
Flight data organization
AdAlysis
26. @bgTheory
Organization is From the Ad: Not Keywords
Ads CTR Conv Rate
To <city> 2.3% 1.90%
From <city> 1.8% 1.5%
<city> to
<city>
4.1% 3.75%
Keywords:
• Orlando to Munich Flights
• Munich to Rome Flights
• etc
Repeat for airport names, codes
Ads CTR Conv Rate
To <city> 2.9% 1.8%
From <city> 2.3% 1.5%
<city> to <city> 4.1% 3.6%
Geo-targeted campaigns to city:
Keywords:
• Flights to Munich
• Cheap airline to Rome
Result: Every city should have 14,400 ad groups
AdAlysis
27. •Airport names > City or airport codes: need airport name ad
groups
•Airport codes > City names or airport names: need airport code
ad groups
•Using city based campaigns outperforms national campaigns
when only destination is included in query.
•Perfect management:
Minimum: 360,000 ad groups
•Resources: 4 people
Result: “Best Structure” overwhelms resources
@bgTheory
More Testing
AdAlysis
28. @bgTheory
Final Initial Organization Due to Resource
ConstraintsCampaign Type # of cmp type Keyword Type Modifications Total Ad Groups
Country - Specific 6<city> to <city> None 9600
Country - Specific 6<airport code> to <airport code> None 9600
Country - Specific 6<airport name> to <airport name> None 9600
Country - Specific 6From <airport code> None 240
Country - Specific 6To <airport code> None 240
Country - Specific 6From <city> None 240
Country - Specific 6To <city> None 240
Country - Specific 6From <airport name> None 240
Country - Specific 6To <airport name> None 240
Country - Specific 6<city> to <city> Cheap / Deals 9600
Country - Specific 6<airport code> to <airport code> Cheap / Deals 9600
Country - Specific 6From <airport code> Cheap / Deals 240
Country - Specific 6To <airport code> Cheap / Deals 240
Country - Specific 6From <city> Cheap / Deals 240
Country - Specific 6From <airport name> Cheap / Deals 240
Country - Specific 6To <airport name> Cheap / Deals 240
Country - Specific 6To <city> Cheap / Deals 240
City - specific 10To <airport code> None 400
City - specific 10To <city> None 400
City - specific 10To <airport name> None 400
City - specific 10To <airport code> Cheap / Deals 400
City - specific 10To <city> Cheap / Deals 400
City - specific 10To <airport name> Cheap / Deals 400
City - specific 10Generic flight words None 10
City - specific 10Generic words (cheap flights) Cheap / Deals 10
Country - Generic 6Generic words (cheap flights) Multiples 150
Brand 1Brand Multiples 11
Total 23 53,461
AdAlysis
29. •If city has more than 500/conversions year – it gets a
campaign.
•RLSA is bid modifier only
•Using dynamic remarketing
•All ad groups labeled by match type to ensure proper structure
Cannot use anything but exact and phrase in <city> to <city>
campaigns.
@bgTheory
Ongoing Rules
AdAlysis
30. @bgTheory
Close Variant Matches
Query Type Conv Rate Average Order
Value
Blue Widgets (plural) 2.3% $673
Blue Widget
(singular)
2.9% $81
Small Accounts > Often ignore as managing plurals & singulars are
very difficult and increases management time significantly with little
returns.
Large Accounts > Take advantage of this and split out singulars &
plurals by campaign for bid purposes.
AdAlysis
33. @bgTheory
Why? Improper Definition of Hero Term
A Hero Term (for them) is:
• 35 impressions / month
• 2 conversions / month
For 3 consecutive months
AdAlysis
41. 1. Enter name
2. Enter zip
3. Choose data type
4. Enter email
5. System ‘looks up
data’
Showed benefits of data
while ‘looking up info’
@bgTheory
Speed To Conversion
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
6.00%
7.00%
0 15 30 45 60 75 90 105 120 135 150 175 190 205 220 215
Seconds Waiting
Conversion Rates by Seconds Before Conversion
Option Displayed
AdAlysis
44. •Long tail
•Small GEOs
•“Target & Bid” RLSAs
•“Target & Bid” Customer match
•Capturing everything possible
•Research
@bgTheory
Good Broad Match Uses
AdAlysis
45. •Best practices are great starting places
•There are many exceptions:
Most are due to simplifying complex problems
They may (rightfully so) be ignored due to resource constraints
There are some (but not many) actual exceptions to ignoring them
completely
@bgTheory
Wrap-Up
AdAlysis
46. Thank You! Hero Conf
Let’s Connect
@bgTheory
in/ewhisper Powerful Ad Testing Made Si
Ad Testing
Software
Learn PPC
PPC Training & Tools
@bgTheoryAdAlysis