datalicious media attribution options
DESCRIPTION
The presentation discusses cross-channel media attribution best practice and technology options.TRANSCRIPT
> Media A)ribu-on < Cross-‐channel media a0ribu3on best prac3ce and technology op3ons
> Short but sharp history § Datalicious was founded in late 2007 § Strong Omniture web analy3cs history § 1 of 4 preferred Omniture partners globally § Now 360 data agency with specialist team § Combina3on of analysts and developers § Carefully selected best of breed partners § Driving industry best prac3ce (ADMA) § Turning data into ac3onable insights § Execu3ng smart data driven campaigns January 2012 © Datalicious Pty Ltd 2
> Smart data driven marke-ng
January 2012 © Datalicious Pty Ltd 3
Media A)ribu-on & Modeling
Op-mise channel mix, predict sales
Tes-ng & Op-misa-on Remove barriers, drive sales
Boos-ng ROMI
Targeted Direct Marke-ng Increase relevance, reduce churn
“Using data to widen the funnel”
> Media a)ribu-on
101011010010010010101111010010010101010100001011111001010101010100101011001100010100101001101101001101001010100111001010010010101001001010010100100101001111101010100101001001001010
January 2012 © Datalicious Pty Ltd 4
> The ideal media dashboard
January 2012 © Datalicious Pty Ltd 5
Channel Investment ROMI Return
Brand equity Baseline ($100) n/a $40
Offline TV, print, outdoor, etc $7 330% $30
Direct Direct mail, email, etc $1 400% $5
Online Search, display, social, etc
$2 1150% $25
> Duplica-on across channels
January 2012 © Datalicious Pty Ltd 6
Banner Ads
Email Blast
Paid Search
Organic Search
$ Bid Mgmt
Ad Server
Email PlaNorm
Google Analy-cs
$
$
$
Central Analy-cs PlaNorm
$
$
$
> De-‐duplica-on across channels
January 2012 © Datalicious Pty Ltd 7
Banner Ads
Email Blast
Paid Search
Organic Search
$
Direct mail, email, etc
Facebook Twi)er, etc
> Campaign flows are complex
January 2012 © Datalicious Pty Ltd 8
POS kiosks, loyalty cards, etc
CRM program
Home pages, portals, etc
YouTube, blog, etc
Paid search
Organic search
Landing pages, offers, etc
PR, WOM, events, etc
TV, print, radio, etc
= Paid media
= Viral elements
Call center, retail stores, etc
= Sales channels
Display ads, affiliates, etc
> Success a)ribu-on models
January 2012 © Datalicious Pty Ltd 9
Banner Ad $100
Email Blast
Paid Search $100
Banner Ad $100
Affiliate Referral $100
Success $100
Success $100
Banner Ad
Paid Search
Organic Search $100
Success $100
Last channel gets all credit
First channel gets all credit
All channels get equal credit
Print Ad $33
Social Media $33
Paid Search $33
Success $100
All channels get par-al credit
Paid Search
> First and last click a)ribu-on
January 2012 © Datalicious Pty Ltd 10
Chart shows percentage of channel touch points that lead to a conversion.
Neither first nor last-‐click measurement would provide true picture
Paid/Organic Search
Emails/Shopping Engines
> Ad clicks inadequate measure
February 2012 © Datalicious Pty Ltd 11
Only a small minority of people actually click on ads, the majority merely processes them (if at all) like any other adver3sing without an immediate response so adver3sers cannot rely on clicks as the sole success measure but should instead focus on impressions delivered
> Indirect display impact
April 2012 © Datalicious Pty Ltd 12
> Indirect display impact
April 2012 © Datalicious Pty Ltd 13
Closer
Paid search
Display ad views
TV/print responses
> Full purchase path tracking
January 2012 © Datalicious Pty Ltd 14
Influencer Influencer $
Display ad clicks
Online sales
Affiliate clicks
Social buzz
Offline sales
Organic search
Website events
Direct mail, emails
Life-me profit
Social referrals
Retail store visits
Direct site visits
Introducer
> Tracking offline responses online
§ Search calls to ac3on for TV, radio, print – Unique search term only adver3sed in print so all responses from that term must have come from print
§ PURLs (personalised URLs) for direct mail – Brand.com/clientname redirects to new URL that includes tracking parameter iden3fying response as direct mail
§ Website entry survey for direct/branded visits – Survey website visitors that have come to site directly or via branded search about their media habits, etc
§ Combine data sets into media a0ribu3on model – Combine raw data from online purchase path, website entry survey and offline sales with offline media placement informa3on in tradi3onal media a0ribu3on model
January 2012 © Datalicious Pty Ltd 15
> Search call to ac-on for offline
December 2011 © Datalicious Pty Ltd 16
VickyCarroll.myspaday.com > redirect to > myspaday.com?
CampaignID=DM:123& Demographics=F|35& CustomerSegment=A1& CustomerValue=High& CustomerSince=2001& ProductHistory=P1|P2& NextBestOffer=P3& ChurnRisk=Low [...]
> Personalised URLs for direct mail
December 2011 © Datalicious Pty Ltd 17
> Website entry survey
December 2011 © Datalicious Pty Ltd 18
Channel % of Conversions
Straight to Site 27%
SEO Branded 15%
SEM Branded 9%
SEO Generic 7%
SEM Generic 14%
Display Adver3sing 7%
Affiliate Marke3ng 9%
Referrals 5%
Email Marke3ng 7%
De-‐duped Campaign Report
} Channel % of Influence
Word of Mouth 32%
Blogging & Social Media 24%
Newspaper Adver3sing 9%
Display Adver3sing 14%
Email Marke3ng 7%
Retail Promo3ons 14%
Greatest Influencer on Branded Search / STS
Conversions a0ributed to search terms that contain brand keywords and direct website visits are most likely not the origina3ng channel that generated the awareness and as such conversion credits should be re-‐allocated.
> Cross-‐channel impact
April 2012 © Datalicious Pty Ltd 19
> Tracking offline sales online
§ Email click-‐through – Include offline sales flag in URL parameter in welcome email click-‐through URLs (or 1st email newsle0er arer offline sale) to trigger a custom ‘assisted offline sales’ conversion event
§ First login arer purchase – Similar to the above method, however offline sales flag happens via JavaScript parameter defined on login rather than URL parameter
January 2012 © Datalicious Pty Ltd 20
Confirma-on email, 1st login
> Offline sales driven by online
December 2011 © Datalicious Pty Ltd 21
Website research
Phone order
Retail order
Online order
Cookie
Adver-sing campaign
Credit check, fulfilment
Online order confirma-on
Virtual order confirma-on
> Understanding channel mix
January 2012 © Datalicious Pty Ltd 22
April 2012 © Datalicious Pty Ltd 23
> Adjus-ng for offline impact
April 2012 © Datalicious Pty Ltd 24
+15 +5 +10 -‐15 -‐5 -‐10
Closer
25%
> Media a)ribu-on models
January 2012 © Datalicious Pty Ltd 25
Influencer Influencer $100
25% Even A)rib.
Exclusion A)rib.
Custom A)rib.
25% 25%
Introducer
33% 33% 33% 0%
? ? ? ?
Closer
Channel 1
Channel 1
Channel 1
> Path across different segments
April 2012 © Datalicious Pty Ltd 26
Influencer Influencer $
Channel 2
Channel 2 Channel 3
Channel 2 Channel 3 Product 4
Channel 3
Channel 4
Channel 4
Introducer
Product A vs. B
New prospects
Exis-ng customers
> Purchase path vs. a)ribu-on
§ Important to make a dis3nc3on between media a0ribu3on and purchase path tracking – Not the same, one is necessary to enable the other
§ Tracking the complete purchase path, i.e. every paid and organic campaign touch point leading up to a conversion is a necessary requirement to be able to actually do media a0ribu3on or the alloca3on or conversion credits back to campaign touch points – Purchase path tracking is the data collec3on and media a0ribu3on is the actual analysis or modelling
January 2012 © Datalicious Pty Ltd 27
> Single source of truth repor-ng
January 2012 © Datalicious Pty Ltd 28
Insights Repor-ng
> Where to track purchase path
January 2012 © Datalicious Pty Ltd 29
Referral visits Social media visits Organic search visits Paid search visits Email visits, etc
Web Analy-cs Banner impressions
Banner clicks +
Paid search clicks
Ad Server
Lacking ad impressions Less granular & complex
Lacking organic visits More granular & complex
> Purchase path data samples Web Analy-cs data sample (AD IMPRESSION >) AFFILIATE > SEARCH > $$$ SEARCH > SOCIAL > DIRECT > $$$
Ad Server data sample 01/01/2012 11:45 AD IMPRESSION 01/01/2012 12:00 AD IMPRESSION 01/01/2012 12:05 SEARCH 07/01/2012 17:00 DIRECT 08/01/2012 15:00 $$$
January 2012 © Datalicious Pty Ltd 30
> Purchase path for each cookie
January 2012 © Datalicious Pty Ltd 31
Mobile Home Work
Tablet Media Etc
> PlaNorm op-ons
101011010010010010101111010010010101010100001011111001010101010100101011001100010100101001101101001101001010100111001010010010101001001010010100100101001111101010100101001001001010
January 2012 © Datalicious Pty Ltd 32
> Purchase path op-ons § Google Analy3cs – Mul3-‐channel funnels
§ Adobe SiteCatalyst – Cross-‐visit par3cipa3on JavaScript plugin – DoubleClick/Mediamind Genesis integra3on – Light Server Calls
§ Atlas, Mediamind, DoubleClick, etc – Campaign touch point report (Mediamind) – Raw cookie level data & manual analysis
§ ClearSaleing
January 2012 © Datalicious Pty Ltd 33
> Google Analy-cs § Mul3-‐channel funnels
– Concatenates and groups campaign codes across visits § Pros
– Free of charge as part of main tool – Out of the box reports that require no addi3onal tags – Does not require data warehousing and custom analy3cs
§ Cons – Raw data cannot be exported for advanced modelling – Cannot combine purchase paths across devices – Doesn’t include ad impressions (maybe DoubleClick in the future) – Channels can be grouped but not re-‐classified in the interface
§ Offline responses via search call to ac3on show up as search, etc – Plauorm cannot accommodate offline data (surveys, CRM, etc) – Purchase path data only, no a0ribu3on modeling or ROMI
January 2012 © Datalicious Pty Ltd 34
January 2012 © Datalicious Pty Ltd 35
> Adobe SiteCatalyst § Cross-‐visit par3cipa3on & light server calls
– Concatenates campaign codes across visits into single string – Light Server Call adds missing ad impressions into campaign code string
§ DoubleClick/Mediamind Genesis – Imports last ad impression, click and ad costs into SiteCatalyst as events
§ Pros – Easy add-‐on for SiteCatalyst clients, requires JavaScript change only – Doesn’t require addi3onal data warehousing as stored in SiteCatalyst – Campaign responses can be re-‐classified arer the fact (SAINT)
§ Cons – Conflic3ng un-‐integrated solu3ons (Light Server Call vs. Genesis) – Addi3onal costs due to a light server call for each ad impression – Light Server Call tracks ad impression but lacks more granular ad data – Cannot concatenate campaign codes across different top level domains – Data can only be exported as concatenated strings without 3me stamps – Plauorm cannot integrate offline data into modelling (surveys, CRM, etc) – Basic a0ribu3on modelling only that cannot be changed in retrospect – Purchase path data only, no flexible a0ribu3on modeling or ROMI
January 2012 © Datalicious Pty Ltd 36
January 2012 © Datalicious Pty Ltd 37
> Purchase path data samples
Web Analy-cs data sample LAST AD IMPRESSION > SEARCH > $$$| PV $$$ AD IMPRESSION > AD IMPRESSION > SEARCH > $$$
Ad Server data sample 01/01/2012 11:45 AD IMP YAHOO HOME $33 01/01/2012 12:00 AD IMP SMH FINANCE $33 01/01/2012 12:05 SEARCH KEYWORD -‐ 07/01/2012 17:00 DIRECT $33 08/01/2012 15:00 $$$ $100
January 2012 © Datalicious Pty Ltd 38
> Atlas, Mediamind, DoubleClick § Campaign touch points (Mediamind only)
– Last 10 touch points before conversion aggregated across users § Raw cookie level data (all ad servers)
– Full list of all ad touch points for each cookie ID § Pros
– Low to very low cost for raw data – Complete raw data available (cookie level op3on only) – Increased flexibility due to complete and very granular data – Campaign responses can be re-‐classified arer the fact – Solu3on can accommodate offline data (surveys, CRM, etc) – Ad server can track across different domains due to 3rd party cookie – Enables advanced ROMI and a0ribu3on modelling with offline data – A0ribu3on model can be changed in retrospect and recalculated
§ Cons – Requires custom JavaScript tags to capture organic touch points – Requires addi3onal data warehousing and custom analy3cs
January 2012 © Datalicious Pty Ltd 39
January 2012 © Datalicious Pty Ltd 40
> Tag implementa-on: SuperTag
41
SuperTag
Conversion Tracking
Conversion De-‐duping
Media A)ribu-on
Behavioral Targe-ng
A/B Tes-ng Heat Maps
Live Chat
Web Analy-cs
Any JavaScript
Easily implement and update any tag on any websites without or limited IT involvement De-‐duplicate conversions for CPA deals and align repor3ng figures across plauorms Collect accurate mul3-‐channel media a0ribu3on data to provide advanced insights Enable advanced features such as targe3ng, tes3ng and chat to op3mise user experience
2012 © Supertag Pty Ltd
> Purchase path data samples
Ad Server summary data sample AD IMPRESSION > DIRECT > SEARCH > $$$ 10x AD IMPRESSION > AFFILIATE > SEARCH > $$$ 5x
Ad Server data sample UID123 01/01/2012 11:45 AD IMP YAHOO $33 UID123 01/01/2012 12:00 AD IMP SMH $33 UID123 01/01/2012 12:05 SEARCH -‐ UID123 07/01/2012 17:00 DIRECT $33 UID123 08/01/2012 15:00 $$$ $100
January 2012 © Datalicious Pty Ltd 42
January 2012 © Datalicious Pty Ltd 43
2012 © Supertag Pty Ltd 44
> ClearSaleing
§ Pros – Fully managed media a0ribu3on plauorm – Extensive out of the box repor3ng func3onality – Can customise a0ribu3on model (not black box) – Supports ROMI calcula3on and a0ribu3on modeling
§ Cons – Priced on percentage of media spend – Requires website tagging and backend integra3ons – Only pre-‐processed data export, not raw data
January 2012 © Datalicious Pty Ltd 45
January 2012 © Datalicious Pty Ltd 46
> Forrester: ClearSaleing ROI/TEI
§ Improved decision making and media buying § Labor savings and improved produc3vity § Accurate media repor3ng and improved visibility § Improved flexibility and faster market response § Transparency and consistency of metrics § Increase in online adver3sing budget September 2011 © Datalicious Pty Ltd 47
Contact me [email protected]
Learn more
blog.datalicious.com
Follow me twi)er.com/datalicious
January 2012 © Datalicious Pty Ltd 48
Data > Insights > Ac-on
January 2012 © Datalicious Pty Ltd 49