PROGRAMMATIC ADVERTISING
How can machine and human
work together
to predict your next purchase?
AVOID DISTRACTOR METRICS
If an ad loads in a browser and no one is
there to notice it, does it really matter?
PLAN BY GOAL, NOT CHANNEL
Traditional media planning in 1930’s
Remember this guy?
Publisher
Time
Plan by mindset
Awareness
Engagement
Leads
Direct Response
Acquisition
Think like a customer, not like a publisher
Awareness
Engagement
Leads
Direct response
Acquisition
Online
sales
Call centre
Leads
nurturing
Store sales
Mobile
video
YouTube
Facebook
video
Instagram
Premium
programmatic ATL Referral
I’m becoming aware of
the brand
Facebook Native Display SEO SEM
Direct
traffic
Non-branded keywords
I’m interested in the
brand
Facebook Native Display
I’m considering the brand
Facebook Display
SEO SEM
Direct
traffic
branded keywords
I’m showing intent to buy
the brand.
Facebook Display
I’m purchasing from the
brand
SEO SEM
Direct
traffic
branded keywords
Engaged Visitors
Leads
Quotes
Completed Views
Sales
SEGMENT YOUR AUDIENCES
Listen to Audience Signals
What product was I looking at?
What product was I interested in buying?
Product
What media source did I come from?
How well does this source usually convert?
Source
What stage of the buying process did I get to?
Did I add the product to my cart or enter
personal details?
Stage
What topics have I demonstrated interest in?
What other pages have I interacted with?
Interest
How long ago did I interact with the website?
When am I likely to purchase?
Time
How many times did I interact with the website?
Touchpoint
Cold Warm Hot Customers
3,810
2 days
15,608
7 days
90,448
30 days
172
2 days
856
7 days
6,748
30 days
163,792
30 days
2,146
30 days
12,416
365 days
Move from Pools to Ponds
Attribution is a Point-of-View
Sales vs Attribution Windows
0
250
500
750
1000
PC 1 PC 7 PC 30 PV 1 PV 7 PV 30
Sales
Attribution Methods and Windows
Time Lag from First interaction
January
February
March
Error Magnitude
The further you get from the buyers true goal, 

the greater the error (and it’s cost)
Click
ROMI
View
Error
Acquisition
Be Aware of Reporting Magic Tricks
CPM CPC Clicks Conversions
$3.20 $3.84 1570 62
$1.80 $2.16 1874 54
$2.40 $2.88 1953 71
$1.20 $1.44 2044 97
$3.90 $4.68 1453 57
$4.10 $4.92 1653 48
$3.20 $3.84 1754 66
$2.40 $2.88 1955 74
Conversions
0
25
50
75
100
1 2 3 4 5 6 7 8
Conversions
RED FLAGS
• Totals only (missing
columns)
• Graphs only (no
trends)
• Hiding negatives
• No editing!
✔
Time based campaigns only
Always-on = Always-improving
Time-based Performance
CPA
2.5
3
3.5
3.9
4.4
Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Campaign 1
Campaign 2
Campaign 3
?
Establish a baseline
Key Takeaways
1. Optimise towards Advertiser metrics
2. Start with the goals
3. Map your audience using tag management
4. Develop your own attribution model
5. Make sense of your reports
6. Stay in Market – establish a baseline
Atlassian%20talk gil 6lessons

Atlassian%20talk gil 6lessons

  • 1.
    PROGRAMMATIC ADVERTISING How canmachine and human work together to predict your next purchase?
  • 2.
  • 5.
    If an adloads in a browser and no one is there to notice it, does it really matter?
  • 6.
    PLAN BY GOAL,NOT CHANNEL
  • 7.
  • 8.
  • 9.
  • 11.
  • 12.
    Think like acustomer, not like a publisher Awareness Engagement Leads Direct response Acquisition Online sales Call centre Leads nurturing Store sales Mobile video YouTube Facebook video Instagram Premium programmatic ATL Referral I’m becoming aware of the brand Facebook Native Display SEO SEM Direct traffic Non-branded keywords I’m interested in the brand Facebook Native Display I’m considering the brand Facebook Display SEO SEM Direct traffic branded keywords I’m showing intent to buy the brand. Facebook Display I’m purchasing from the brand SEO SEM Direct traffic branded keywords Engaged Visitors Leads Quotes Completed Views Sales
  • 13.
  • 14.
    Listen to AudienceSignals What product was I looking at? What product was I interested in buying? Product What media source did I come from? How well does this source usually convert? Source What stage of the buying process did I get to? Did I add the product to my cart or enter personal details? Stage What topics have I demonstrated interest in? What other pages have I interacted with? Interest How long ago did I interact with the website? When am I likely to purchase? Time How many times did I interact with the website? Touchpoint
  • 15.
    Cold Warm HotCustomers 3,810 2 days 15,608 7 days 90,448 30 days 172 2 days 856 7 days 6,748 30 days 163,792 30 days 2,146 30 days 12,416 365 days Move from Pools to Ponds
  • 16.
    Attribution is aPoint-of-View
  • 17.
    Sales vs AttributionWindows 0 250 500 750 1000 PC 1 PC 7 PC 30 PV 1 PV 7 PV 30 Sales Attribution Methods and Windows
  • 18.
    Time Lag fromFirst interaction January February March
  • 19.
    Error Magnitude The furtheryou get from the buyers true goal, 
 the greater the error (and it’s cost) Click ROMI View Error Acquisition
  • 20.
    Be Aware ofReporting Magic Tricks
  • 21.
    CPM CPC ClicksConversions $3.20 $3.84 1570 62 $1.80 $2.16 1874 54 $2.40 $2.88 1953 71 $1.20 $1.44 2044 97 $3.90 $4.68 1453 57 $4.10 $4.92 1653 48 $3.20 $3.84 1754 66 $2.40 $2.88 1955 74 Conversions 0 25 50 75 100 1 2 3 4 5 6 7 8 Conversions RED FLAGS • Totals only (missing columns) • Graphs only (no trends) • Hiding negatives • No editing! ✔
  • 22.
    Time based campaignsonly Always-on = Always-improving
  • 24.
    Time-based Performance CPA 2.5 3 3.5 3.9 4.4 Nov DecJan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Campaign 1 Campaign 2 Campaign 3 ? Establish a baseline
  • 25.
    Key Takeaways 1. Optimisetowards Advertiser metrics 2. Start with the goals 3. Map your audience using tag management 4. Develop your own attribution model 5. Make sense of your reports 6. Stay in Market – establish a baseline