The document traces the evolution of programmatic advertising from the early 2000s focus on clicks and page views to the current data-driven programmatic model. It discusses key developments like the rise of ad networks and exchanges that consolidated inventory and used data to fuel audience segments. This led to the launch of demand-side platforms and agency trading desks, and the growth of programmatic buying using real-time bidding. The document argues that data and RTB have helped reverse declines in digital display advertising and enabled better targeting across channels. It predicts continued growth and consolidation in the programmatic industry.
3. RECESSION OF
2001 SHAPED THE
DIRECTION OF
INTERNET ADVERTISING
WTF Happened? 3
4. Nothing important
A RECESSION THAT
MADE THE INTERNET PIVOT
has ever been built
without irrational
WTF Happened? 4
exuberance
– Fred Wilson | VC
5. SUCCESS DRIVEN BY CLICKS AND PAGE VIEWS
► Created a direct response
WTF Happened? 5
offering from the web
► Every sales pitch was about
delivering clicks and
performance
► Google launched in UK to
further amplify the sentiment
► Branding was ditched in favour
of impressions and clicks
► Adserving was born to count
6. WITH REVENUE
CAME WEBSITES
AND LOTS OF THEM…
WTF Happened? 6
Number of websites 2000-2014
7. THE RISE OF THE AD NETWORK
WTF Happened? 7
Advertisers purchase
majority of inventory
through publishers directly
Networks begin to
purchase and package
premium placements
2001 2003 2005
Ad networks
consolidate longtail
and remnant
inventory
Advanced data
begins to fuel networks’
audience segments
Ad networks procure large volumes of
bulk inventory, apply unique insights
& data, and resell impressions at a
premium rate
► Take on media risk for potential
upside of 80-90% profit margins
► Advertisers have little insight
regarding where placements
run or what is successful
8. BEHIND THE SCENES CAME THE EXCHANGES
WTF Happened? 8
Technology platforms where buyers
and sellers directly transact, using
auction-based pricing to determine
inventory distribution
Ad exchanges emerge
to address marketplace
inefficiencies
2005
Started to power the Ad Network
model – fuelling cheap, non
transparent selling to agencies
9. ►Agencies being pushed harder for sales
at reduced price
► Publishers and Adnets challenged
to deliver more for less
► Turned to adexchanges and use
of advertiser and publisher first
party data
►Quality of impressions less of concern
► Margin first and foremost
WTF Happened? 9
10. A SHIFT FROM SALES
REPRESENTATION TO
AGGREGATOR OF IMPRESSIONS
WTF Happened? 10
11. FRUSTRATION BUILDING ON BUY AND SELL SIDE
WTF Happened? 11
Publisher frustration
► Low cpms achieved
► Valuable data leakage
► Value created in Ad Net
► Commoditisation of sites
Buy side frustration
► Lack of visibility
► Transparency issues
► Advertiser data abuse
► Need to drive efficiency
► Data sophistication
12. BUY SIDE TOOLS UP – 2007 SEES MULTIPLE LAUNCHES
Demand Side Platforms (DSPs)
► Media buying tools that allow active optimisation
and dynamic bidding
across multiple ad exchanges & data inventory
sources
► Enhanced technologies integrate
media, data, and learning
algorithms which allow more
complicated executions
WTF Happened? 12
Demand Side Platforms (DSPs) expose
exchange inventory
in single environment
► Started to power the Ad Network
model – fuelling cheap, non
transparent selling to agencies
2007
Agency Trading Desks
(ATDs) grow into
centers of audience-buying
expertise
14. LAUNCH OF
TRADING DESKS
AND
NEW PUBLISHER
RULES SAW AN INDUSTRY
REVOLUTIONISED
WTF Happened? 14
15. LET’S AGREE SOME TERMINOLOGY
WTF Happened? 15
Programmatic Buying
Leveraging data driven
technology to automate media
buying in real time to maximise
efficiency.
Real Time Bidding (RTB)
Buying and selling digital ad
impressions via real-time
auctions. Ad buyers can
differentially value each
individual opportunity to buy
an ad impression in real time.
Addressable Media
Leveraging data to
deliver messaging based
on known insights about
individuals.
A channel or device is
addressable when there
is specific data available
to target specific users.
16. SO THAT TAKES US UP TO TODAY AND TOMORROW
Demand Side Platforms
(DSPs) expose
exchange inventory in
single environment
2001 2003 2005 2007 2009 2011 Today
WTF Happened? 16
Advertisers purchase
majority of inventory
through publishers
directly
Networks begin to
purchase and package
premium placements
Ad exchanges emerge
to address marketplace
inefficiencies
Agency Trading Desks
(ATDs) grow into centres
of audience-buying
expertise
Ad networks
consolidate
longtail and
remnant
inventory
Advanced data
begins to fuel
networks’
audience
segments
Technology &
data convergence
leads to Real Time
Bidding/Buying
Data Management Platforms
(DMPs) evolve to manage
data
17. AND TOMORROW BRINGS INCREDIBLE GROWTH
WTF Happened? 17
50-60% of all digital
will be programmatic in 2yrs
19. DATA AND RTB
HAVE SAVED
DIGITAL ADVERTISING
Reversed declines in display
Industry growing
innovation, investment
and moving upscale
WTF Happened? 19
25. INNOVATIVE PRODUCTS
► Triggering Ads on 2nd Screens based on a TV
WTF Happened? 25
Signal
Classical
► Brand Spots
► Single Spots
► Special Spots
► Products Spots
Competitor
► Brand Spots
► Single Spots
► Special Spots
► Products Spots
Contextual
► Verticals
► Individual
27. 5 BIG THEMES TO WATCH
WTF Happened? 27
The rise of GIANTS
One Data Management Platform to rule them all
ENTER the non media company
Consolidation is accelerating
Cross channel expertise
28. PROGRAMMATIC FOR YOU
WTF Happened? 28
► Be open minded
►Be brave – time to revolutionise
► Move fast
► Be inquisitive
Notes de l'éditeur
Created a direct response offering from the web
Every sales pitch was about delivering clicks and performance
Google launched in UK to further amplify the sentiment
Branding was ditched in favour of impressions and clicks
Adserving was born to count
Publishers challenged to sell 100% of inventory direct - initially all impressions were bought directly through publishers, but as impression volumes exploded, this model was too time intensive to scale
Pre-2005
Publishers seek out guaranteed revenue streams to offset volatility
Risk of unsold inventory shifts from publisher to network, providing publisher yield consistency
Buyers are seeking stronger targeting and lower cost options for achieving scaled reach against specifically defined target audiences vs. direct publisher buys
Data Insights: Basic – Geography or Context
Advanced – Behavioral or Retarget
Complex – Optimization Models
2005 – 2010
Ad networks once provided the only reach & efficiency offer; Now ad exchanges provide a larger, safer and far more transparent means to buy audiences
Akin to a stock exchange - Significant infrastructure & overhead costs limit the number of exchanges
Initially ran as “futures” based - each bidder loaded their rules upfront
targeting parameters
Bid Maximum
Pacing & Frequency Caps
Trafficking Instructions
The exchange platform processed every impression per these rules & determined a winner
Updates were made manually or via an API every 60 minutes
In a “real time” exchange, every impression is bidded on individually
Each bidder reviews the impression
Evaluates against proprietary data
Decides optimal bid at that moment
Relays bid to exchange (<100ms)
Exchange determines winner
Private Exchange: Invitation-only marketplace where buyers (advertisers) and sellers (publishers) transact based on a negotiated CPM. The PMP provides preferred access to inventory from Premium Publishers that is not typically made available within the open exchange. In order to connect the buy and sell side, conduits such as AdX (Google’s inventory source), Rubicon Project, Pubmatic, etc., provide a Deal ID associated with a specific publisher or area on a publisher’s page, to allow the buyer and seller to transact outside of the open exchange. A Deal ID enables a publisher to give exclusive access or rates for their inventory to a specified buyer.
PMP pricing can either be fixed or a floor. Fixed pricing is based on a fixed CPM rather than a dynamic CPM. Audience On Demand® (AOD) still purchases the inventory programmatically, but the price per impression will not fluctuate as it does within the dynamic CPM model that AOD typically employs. Alternatively, there may be a floor CPM, meaning the cost a buyer will pay must be higher than the minimum (floor) set by a publisher. AOD recommends a $60,000 minimum spend per month for client programs that require custom sites and client specific pricing.
PMPs are designed to use the same model as the open exchange, layering data to target audiences on a specific site or set of sites. As of May 2014, this is still a future state though, as they are currently operating the same as a direct buy but instead of buying the impressions upfront for a contracted flat CPM, buyers bid on the impressions in real time via a DSP (but only the specific advertiser, has access to the “auction”…hence “Private Marketplace” or “Private Exchange”). Think of it as regular RTB auction but only one buyer and one seller (that’s an oversimplification, but it gives you an idea of how it works). Issues remaining include problems with scale – you limit scale by default when working with a PMP and pricing – prices are higher in a PMP because it is considered “premium” inventory, so adding data costs can be prohibitive.
Agencies being pushed harder for sales at reduced price
Publishers and Adnets challenged to deliver more for less
Turned to adexchanges and use of advertiser and publisher first party data
Quality of impressions less of concern
Margin first and foremost
2006 – 2010
Trading desks launched just before DSPs begin appearing and quickly took over this aspect of exchange buying as a specialty area.
DSPs: Similar to SEM bid management platforms
Benefits:
Impression level transparency
Control over data, optimization decisions, and ability to integrate into agency workflow
Centralization / simplification of access to complex inventory pools
Implement Real Time Buying, by integrating with exchanges’ RTB apparatuses
Audience targeting – incorporating 3rd & 1st party data to define active targets
Predictive modeling – algorithm driven performance engines that drive conversion lift
Cons: Managing DSP relationships requires a team of dedicated resources to: oversee technology relationships with the DSPs, data providers and inventory providers; set up, execute and optimize campaigns; and translate the unique data generated by DSPs into an actionable format
DSP operation & campaign optimization is complicated and requires specialized departments. Trading desks build or contract DSP technology, and serve as expert users to create value for clients
DSP technology requires experienced “traders” to deliver maximum output - The core of the trading desk is its team of experts
Benefits:
Keep clients’ data secure so it doesn’t become a 3rd party data segment for another advertiser
Client investment that historically allowed vendors & competitors to build their businesses, now creates long term value for agency & client simultaneously
Cadreon (IPG); Accuen (Omnicom); Xaxis (WPP)
Addressable Media - Leveraging data to deliver messaging based on known insights about individual users regardless of placement or device. A channel (display, video, social, mobile) or device (smartphone, connected TV) is addressable when there is specific data available to target specific users.
RTB - Dynamic process of buying and selling ad impressions via real-time auctions, in which the highest bidder wins the right to place a display ad. An ad buyer can differentially value each individual opportunity to buy an ad impression in real time.