IAB Europe Report: Using Data Effectively in Programmatic V2.0 (GDPR Update)

WHITE PAPER
Using Data Effectively
in Programmatic
2.0 - GDPR Update
July 2018

IAB Europe Using Data Effectively in Programmatic White Paper
CONTENTS
Page 2
3 1. Introduction
6 2. Understanding Audience Data
6 2.1 Buy-side audience data
13 2.2 Sell-side audience data
16 3. Using Data Effectively and Building a Data Strategy
16 3.1 Combining data into target segments and activating
campaigns
19 3.2 Key considerations for using data effectively
24 4. Cross-Device Opportunities and Challenges
29 5. GDPR Considerations
32 6. Conclusions
35 7. With Thanks

The IAB Europe Using Data Effectively in Programmatic White Paper has been updated to
reflect the changing nature of the European data landscape. This updated version includes
revised content with an additional section covering GDPR considerations and new case
studies.
Consumers are using an ever-increasing variety of channels and touch points across media to
shop and to consume content. The interactions and behaviours across these touch points:
online video; news and information sites; social media; gaming; to name a few create myriad
data points that initially can seem disconnected. Combine this fact with the increasing use of
ad blocking tools and the challenge of delivering relevant advertising becomes even more
real.
As consumers navigate through these channels and touch points they should receive only
relevant and quality advertising messages and the promise of programmatic has been to
reach the right consumer at the right time in the right place with the right message. Indeed,
programmatic advertising has continued strong growth over the past few years and is now an
€8.1bn market in Europe.
This promise cannot be delivered without data about the consumer. Leveraging data
intelligently can both reveal and build connections across these points that make it possible
to understand and target customers as they progress through the buying journey.
Data is a hot topic but also a very generic term, it is important to understand the different
types of data available in marketing (see following page).
1. INTRODUCTION
Page 3

INTRODUCTION
Types of data available
• Financial data can be data about sales, market-share, return on investment.
• Research data can be data from media usage panels, brand studies and consumer
segmentation – this data is a key source for consumer insights in marketing.
• Known consumer data on the other hand, is linked to consumers such as CRM data or
eCommerce sales data.
• Pseudonymous audience data is also linked to individual identifiers (like a cookie or
a cross-device ID), but as all links to personal data are removed. This can be used for
reaching audiences on a large scale.
• Campaign data can be impressions, clicks, engagement tracked by an ad server or
website analytics data. This is usually not linked to pseudonymous audience data –
however it can be linked to make it a source of audience data.
This white paper focusses on the fourth point, pseudonymous audience data that is used
for targeting in programmatic campaigns. It provides guidance on the types of audience
data available and how to use it more effectively, particularly in programmatic trading.
Data is considered to be the “new oil”. Data-poor advertisers without a direct connection
to their consumers, such as FMCG brands that are selling via supermarkets, hope to get
closer to them by understanding their engagement with advertising and owned content.
Building their own segmentation, these advertisers want to become more relevant to the
consumer by targeting the right person with the right message which in turn means
consumers won’t receive advertising that is not of interest to them.
Data-rich advertisers, for example those where a large part of the purchase journey
happens online, want to extend the personalised communication from their CRM
programmes and existing consumers to prospect consumers.
Page 4

INTRODUCTION
Both groups of advertisers can benefit from collecting, analysing and activating audience
data to give them a much better understanding of fast-changing consumer needs, equip
them with better tools to engage with consumers through relevant messages and
ultimately to increase the effectiveness of their communications along the purchase
funnel – from mass communication to the one-to-one level.
Page 5

Advertisers are wise to consider all types of data sources to address the various audiences
and goals they are trying to reach.
Buy-side data refers to data that advertisers own or can access and use. Every time an
interaction happens between consumer and advertiser, data is created. This can include
data about their:
• Interests
• Decisions
• Points of contact
• Activities relative to those moments of engagement,
• Perceived needs
• Key demographic and behavioural background and how that background compares to
other consumers
• Length of time between the current and previous interactions
• Frequency of engagement
• Exposure to advertising
This should serve as a robust platform from which to build insights about consumers, what
motivates them, and how to reach them in the moments that matter.
2.1 Buy-side audience data
2. UNDERSTANDING
AUDIENCE DATA
Page 6

UNDERSTANDING
AUDIENCE DATA
2.1.1 First party data
First party data refers to any data collected by the advertiser. Examples of this are:
• Transactional or registration CRM data, whether online or offline, from people who
have made purchases at physical stores or via an e-commerce site; from sales leads;
from database sign-ups, and from people who have interacted with a call centre.
• Website data and campaign performance data from email, display, video, mobile,
direct mail, TV, print.
Benefits of first party data
• First party data is usually the data from which advertisers can derive the most value as
it is data from consumers that have shown interest in the advertiser’s offer already or
are consumers already and therefore are much more likely to convert or engage.
• This data also enables the advertiser to create segments as broad or narrow as needed
to meet the goals of the campaign.
• First party CRM data tied to an email address for example, can be used to identify
consumers across all browsers and devices.
How to utilise first party data
First party data can be collected via tags on an advertiser’s digital assets or by on-
boarding offline data and matching it to digital identifiers, i.e. cookies
The most common way to build up first party data at scale is to use the technology of a
data management platform (DMP) or the DMP functionalities within buying tools (e.g.
DSP, Search or Social Bid-Management tools or an Ad Server). These can place tags
throughout the website, apps, ads or email newsletters to track the type of content
consumed and the frequency of content consumption.
CRM data can be on-boarded with data-matching services, which have large databases of
personal data such as email addresses, phone numbers or home addresses and cookies or
digital IDs.
Page 7

UNDERSTANDING
AUDIENCE DATA
These services identify the user from the advertiser’s CRM database and match them with
an associated cookie. It is important to make sure that this is done in a way that complies
with the relevant EU privacy legislation.
Keeping control
Knowledge derived from first party data is a highly valuable and portable asset, and like
other similar company assets, there must be a clear understanding of who has access to
the data and the reasons for their access. Failure to think through and implement
sufficient data governance can result in data leakage, the term used when companies lose
control over their data.
It is important that access to first party data is monitored and appropriate encryption is
used. There are also instances in which organisations share data with partners or
suppliers. It is vital to fulfil obligations under data protection legislation, see section 6 on
GDPR considerations.
Cookies (and other identifiers for the collection of information from users, such as mobile
advertising identifiers) are another key component as they are essential for proper
targeting. It is important that advertisers understand, comply with, and keep track of
regulatory obligations affecting the collection of cookie data. Businesses should assess
whether regulatory compliance requires putting in place contractual agreements, for
example to ensure that the consent of users is organised by a publisher on behalf of its
third-party partners for the collection and use of cookie data where such consent is
necessary. IAB Europe’s e-Privacy Directive implementation centreprovides guidance on
compliance with the cookie consent standards of each EU member state.
Regular auditing of tags on the site and documenting of data flows to ensure they are
approved third parties will help to ensure control.
This being said, it is also important to understand that there are varying "risk levels"
associated with different data types.
Page 8

UNDERSTANDING
AUDIENCE DATA
Cookie IDs as such are comparably uncritical pieces of information as they are truly
pseudonymous and without a match table in place it is very hard to make any sense of
them. It's different with "real" personal data such as email addresses. For advertisers
operating with data that qualifies as personal it is recommended that partners are
carefully selected.
Legal and privacy considerations
Advertisers that want to maximise the value of their first party data should involve their
legal and privacy teams early in the process to ensure full compliance with European data
protection and privacy laws (see section 6 on GDPR considerations).
A good way to keep on top of the latest developments regarding regulation and self-
regulation is via IAB Europe’s Policy Committee and its various task forces. Additional
market specific information can be provided by the relevant national IABs and partners.
2.1.2 Third party data
Third party data is usually not sold but licensed and paid for each time it is activated in a
campaign via a CPM cost model (cost per 1,000 impressions bought against that data) or a
percentage of media spend.
Page 9

UNDERSTANDING
AUDIENCE DATA
Types of (third party) data available in a data market place
Data market places offer a wide range of user profiles, based on different sources:
• Interest data: categories of websites visited by user, topics of articles read by user (i.e.
interested in cars, football or house renovation).
• Purchase intent data: user searches or views of a specific product on price
comparison or e-commerce websites (i.e. new SUV, flat in city centre or baby clothes).
• Geo-localisation data: GPS or WiFi device data (using WiFi in specific places such as
restaurants, public transport, airports, etc.).
• Demographic data: data declared during registration process.
• Sociographic data: social status, family and professional status, education, earnings –
data available from different types of social networks.
• E-mail subscription data
• Mobile applications data: registrations and usage of mobile apps.
• Device type
Benefits of third party data
The value of first party data increases dramatically when overlaid with third party data to
provide a much richer picture of audience segments, including lifetime value, attributes,
behaviours and content consumption. This enables:
• A better understanding of an advertiser’s website visitors, which in some cases can
lead to modification of marketing strategy.
• Ability to increase effectiveness of re-targeting campaigns by combining first and third
party data.
• Foundation for look-a-like targeting, where consumers with similar profiles to those
users that converted are targeted, reaching a larger group of potential prospects.
Publishers may use third party data in order to create a better experience for their users –
such as articles and product recommendations and content personalisation.
Page 10

UNDERSTANDING
AUDIENCE DATA
Challenges of third party data
For many data-poor brands that can capture only a small fraction of their target audience
via their owned channels, third party can be a route forward. However, looking into the
depth and breadth of targeting segments available on the data marketplaces is often
overwhelming for an advertiser.
In some smaller markets, the volume and quality of third party data is limited. This may
be due to local privacy regulation and concerns discouraging data holders to sell the data,
the fact they have not yet assessed whether selling audience data is a good idea, or
whether it is better to activate their audiences only via directly sold publisher targeting.
Even if third party data is widely available, challenges exist:
• Often, the data is de-branded, mixed and modelled from different sources and it is very
difficult to understand where the data came from, how accurate and current it is. Some
publishers may classify a person that has just read one article about the new Ferrari as
“in-market for automotive”, but that segment will most likely not perform well for a
compact car brand trying to generate sales.
• With some third party segments, it is unclear whether the data is true declared data,
observed data or modelled.
• Vendors may sell the same data and segments to multiple advertisers with similar
target audiences, so it is not unlikely that brands may bid their programmatic buys
against exactly the same cookie-pool as their competitors.
2.1.3 Second party data
The limited reach and depth of first party data and the sometimes questioned quality of
third party data has turned many advertisers attention to second party data.
Second party data is someone else’s first party data, for example a publisher’s data about
their audience. Second party data can also come from other non-competing brands,
retailers or other holders of data that don’t sell that data on the large third party data
marketplaces.
Page 11

UNDERSTANDING
AUDIENCE DATA
The key difference from third party data is that second party data is traded or shared
directly between two parties. That brings some major advantages but also some
challenges.
Benefits of second party data
One advertiser can leverage the pseudonymised profile, e.g. hashed personal data, of
another advertiser’s consumers in a second party data partnership and target them if they
fit their consumer profile. Additional behaviours from other third party sources can also
be added to the mix, to further supplement the pseudonymised customer profile and even
more precisely target them. In this way, different verticals, such as retail and fast moving
consumer goods (FMCG), can think strategically about using each other’s data to their
benefit, because a much more multi-dimensional understanding of one’s audience is
achieved.
An alternative can be for advertisers to somewhat limit the use of generally available third
party data by combining it with custom segments specifically built for the advertisers
purpose. By combining in-panel surveys to identify highly custom segments and then
running extrapolation analysis on the results advertisers can in theory fully control the
targeting criteria from third party signals. This approach however is in its early days and
not widely practiced yet.
Other benefits include:
• The data is purchased from a known partner, it is very clear how the data was
generated and therefore much easier to evaluate the quality of the data.
• Often second party data is obtained for partnerships with companies that have
complementary data sets to drive campaign performance and share insights.
• Usually the data vendor does not sell the same data to everyone on the market,
therefore limiting the risk of targeting exactly the same consumer as competitors.
• A further benefit for data holders like publishers or retailers selling their data as second
party data is that they have more control over who uses it and for which purpose.
Page 12

UNDERSTANDING
AUDIENCE DATA
Challenges of second party data
However, there are also some challenges in activating second party data at scale:
• As it is based on a direct connection between the data holder and the data user, it is
much harder to activate than third party data, that is readily available in any DSP or
DMP. It is important to consider what commercial arrangements need to be made to
accurately report and bill for data usage by second party data source.
• Some data holders don’t have a feeling for the value of the data that they hold for other
brands, therefore potentially making the commercial agreement complicated.
To leverage second party data at scale, DMP vendors are increasingly enabling data-
cooperatives between their clients, making the sharing and selling of data as easy as for
third party data but with the full control of a direct second party data relationship.
When talking about first and second party data, it is important to be clear who the data
holder is. A publisher’s first party data can be an advertiser’s second party data and vice
versa. The data only becomes third party when it is acquired via an intermediary, i.e. a
data marketplace.
First party data refers to information collected by publishers about their websites’ visitors
and their behaviours, actions or interests demonstrated across their websites. It can also
include CRM data, subscription data, social data, or cross-platform data from mobile web
or apps (similar data types to advertiser first party data).
2.2.1 Value of publisher data
If a publisher has demand that outweighs supply for its inventory and they have the ability
to segment and store audience data, then selling data via audience extension or to the
buy-side stakeholders could be considered. Not only will this enable a new revenue
stream for the publisher, it will also enable advertisers to reach the publisher’s audience
across other media, satisfying the demand and therefore overcoming the lack of supply.
2.2 Sell-side audience data

UNDERSTANDING
AUDIENCE DATA
Advertisers, however, should be made fully aware whether they are buying the publisher’s
audience on the publisher’s owned sites or the same audience on third party sites that
may not have the same quality.
For publishers, data security is paramount; they need to ensure that the data sold is
safeguarded and not used beyond or outside the terms of the sales agreement. There is
benefit in utilising a trusted third party partner, e.g. a data management platform (DMP)
or data exchange, as both parties can securely access and track usage.
Some advertisers prefer to buy media and data separately, as they value the premium
environment from certain publishers but might consider other data partners better suited
to define the audience to be targeted. Publishers with unique audience data might benefit
from this as well, as they can sell all their inventory only once, but maximise the value of
data as on-top revenue.
Page 14

UNDERSTANDING
AUDIENCE DATA
Overview
• Affiperf partnered with a popular travel brand to increase customer
acquisition month over month. By combining the brand’s first party data -
which by itself did not have high enough volume, Affiperf recognised the
need to leverage both first and third party data sets to boost performance
Strategy
1. Affiperf partnered with a leading third party data marketplace to identify
users who expressed the same attributes as the client’s website converters
2. Affiperf leveraged an advanced modelling platform to create two unique
models that balanced reach against data confidence
Results
3-5X higher return on investment
A strong look-alike methodology has given us incremental reach on those
first-party segments with the highest conversion rates. The implementation
process is always smooth from beginning to end.
Pablo Guereñu, Programmatic Trading Manager – Affiperf
Case study
A Winning Equation = first and third party data
“ “
Page 15

As outlined in section two, advertisers have access to a range of data, from self-generated to
purchased, actual to modelled/forecast information. The next stage is to understand the
most effective way to use the data.
Advertisers, media agencies and media owners face a double-sided challenge: avoiding
drowning in data but, at the same time, not missing the huge opportunities that data offers
for engagement with their consumers and users in a personalised and effective way. The
user experience can’t be optimised without using data intelligently and insight provided by
data can lead to more effective use of marketing budgets.
Advertisers today have more power than ever before to ask not only about overall campaign
performance in real time, but demand each site be accountable for every impression served.
Through the availability of first, second and third party data advertisers can discover that
the value of one impression doesn’t necessarily equal the value of the next or the former.
Because of this inherent variability in value, advertisers are demanding and actively using
metrics focused analytical solutions to drive better campaign performance and make every
impression count.
The fundamentals of building audiences are very simple and not different from how
audiences are defined in offline media. The planner combines attributes from first party
data with facts from second or third party data to define an audience, using the Boolean
logic of <AND>, <OR> and <NOT>.
3. USING DATA EFFECTIVELY
AND BUILDING A DATA STRATEGY
Page 16
3.1 Combining data into target segments and activating campaigns

UNDERSTANDING
AUDIENCE DATA
Here are some examples:
• In market customer who has engaged already:
<user has been to product page on advertiser’s website twice in the last three
weeks (first party pixel on product page)> AND NOT <user has bought the product
online (first party pixel on “thank you “)> AND <user is in-market for the product
category (third party data from data marketplace)>
• Prospecting segment
<user is age 25-35 (third party data from data vendor A> OR <user is age 35-45 (third
party data from data vendor A> AND <user is interested in fishing (third party data
from data vendor B)>
• Second party data targeting
<user has configured a VW Golf on Car comparison site (second party data
partnership)> AND <user has configured a Ford Focus (second party data
partnership)> AND NOT <user has been to VW Car configurator>
More sophisticated audiences may be built using audience discovery algorithms (finding
data attributes that over-index for those users that have converted to identify new target
segments), look-alike targeting (finding users with similar behaviour and attributes like the
users that have converted) or building audiences in real-time based on all data signals. But
the principle of combining attributes to define an audience stays the same.
When the audience has been built, it can be activated in marketing tools e.g. demand side
platforms (DSP), for website personalisation or custom audiences on social networks or
search.
Matching data to campaign objectives
In order to identify the most effective data sources to reach certain campaign objectives,
the type of data, reach and granularity have to be taken into account. Branding campaigns
to increase the awareness or affinity for a product in a defined audience, require scalable
data assets with high reach. Socio-demographic information such as age, gender and
income are a classic choice for that, but more and more advertisers tend to address
customised brand audience segments.
Page 17

UNDERSTANDING
AUDIENCE DATA
Data modelling techniques and data aggregation are important to achieve the necessary
reach and to deliver campaigns to large scale audiences.
Programmatic teams should not fall into the trap of focussing mainly on demographics for
their targeting, and indeed while broader communication planning teams have
recognised that today’s varied audiences are much better described by their interests and
attitudes than demographics. So behavioural profiles may be a better choice for targeting
as they can be better signals for product interest, while demographic data is also
impossible to observe? and therefore not available at the scale and accuracy many
advertisers expect.
For performance campaigns that lead to concrete actions, the granularity of the data is
most important.
For these objectives, effective data sources are for example first party re-targeting
information from the advertiser itself or factual intent data from a specific product
category by a suitable third party data provider – to be able to generate leads or increase
sales. For programmatic audience buying systems it is important to cover the whole
customer journey with different data sources, for all campaign goals along the so called
purchase funnel.
Most DSPs are connected to most data marketplaces and data management platforms, so
the operator can activate the audience build in one tool. In the background, however,
complex data matching needs to happen.
Each tool has its own cookie pool (or user profiles) and needs to map them to the other
tools, so that user X from the DMP can be recognised as user Y of the DSP. While those
processes constantly happen in the background, some profiles cannot be matched or
found again, so low data-match rates will limit the reach within the defined audience that
can actually be delivered. Speaking of low data match-rates: it's a widely misunderstood
concept to look at the overall match-rates between two cookie pools.
Page 18

UNDERSTANDING
AUDIENCE DATA
The more valuable information is the match-rate between active cookies (e.g. defined as
"30 day actives"). Changing the focus from "all cookies" to "active cookies" can
tremendously change the picture on "compatibility" between independent data pools.
Single customer view
It is often the case that (internal or external) client briefs focus on needs relating to
specific product or service use cases at the expense of the wider context. With separate
client teams briefing these projects and managing different data sets, there is a real
danger of efforts being duplicated and data being siloed.
Having a solid data governance and management strategy, and a data science-led,
consistent approach to data across the whole business is crucial to avoid a fragmented
view of the customer. A customer/user is not just a buyer of a certain product or service.
They are also a human being with different characteristics, behaviours and attitudes. The
customer may or may not have bought other products from the same brand before. Plus,
there are all those activities that take place out of sight of the brand on various channels
and in different environments. It is therefore paramount, in order to develop and end to
end effective data-driven strategy, to have an approach which will start from the user as a
person rather than a set of disconnected or unrelated data points.
Using audience data to enhance creativity
Technology is often used to spark creativity. Asking the right questions of the data can
lead to new insights and new ideas which in turn can lead to engaging advertising
experiences for the consumer. Third and second party data can be used to understand the
wider interests of the audience segment(s) which in turn can be used to inform the
campaign creative, as well as cluster analysis which can reveal distinct subgroups within
audiences. Data can not only play an important role in the sophistication of the creative,
but also link the creative with overall media strategy.
3.2 Key considerations for using data effectively
Page 19

UNDERSTANDING
AUDIENCE DATA
For example, rather than showing a consumer the same display ad repeatedly, an advertiser
can use the technology that exists to either re-target with a better offer, different creative or
showcase a complementary product.
Serving sequential or consequential ads to consumers needs to become ever more prevalent
to avoid the problem of ads being seen as ‘stalking’ consumers. With the constant evolving
sophisticated dynamic advertising solutions it can be expected that a more programmatic
approach to creative strategies will soon become a normal practice.
Case study
Using Data to Enhance the Creative
A large automotive manufacturer wanted to know who was buying its SUVs and
saloons. By matching its first party data to third party data in a DMP provider,
the manufacturer gained some new and surprising insights: people who bought
SUVs were more likely to own a pair of skis, meaning the creative should show
the car in a wintry setting. The data also showed that people who buy saloon
cars were three times more likely to be Android users – not ideal when the
creative had been talking about the iPhone capabilities of the car! Using these
audience insights, the reworked creative lead to a 50% reduction in cost per
lead.
Page 20

UNDERSTANDING
AUDIENCE DATA
Using campaign data to optimise campaign performance
With real-time campaign performance reports, advertisers are able to regularly evaluate
impressions, clicks, and actions over any designated period of time since the campaign
start date, answering key questions about the optimal settings and mix across audience
profiles, channels, inventory, and creative.
• CRM data can be used to reach known customers and re-engage them with product
offers. This is a good way to increase frequency and conversions amongst the most
loyal consumer base.
• Demographic targeting: audience insights can provide the advertiser with real-time
information about the age, income, lifestyle, and affinity of the people who responded
best to the advertising.
• Geo-targeting: advertisers are able to utilise data to evaluate campaign metrics by
geographic area (for instance at postcode level), then adjust the campaign to areas that
show the best campaign performance.
• Day part targeting: day part reports show a variety of performance metrics by day of
week and by time of day, enabling the campaign to optimise to the most effective times
to serve ads.
• Velocity targeting: used for direct response campaigns, takes into account a
consumer’s frequency of and recency of visits to the brand’s website to either increase
or decrease the user’s exposure to advertising and adapt the messaging appropriately.
• Devices / channels: reviewing campaign performance by device type report can show
which type of ads perform best on which channel, enabling budget to be optimised
across the best mix of channels.
Page 21

UNDERSTANDING
AUDIENCE DATA
• Publishers: campaign data reports are able to map the number of conversions against
unique users for a specified time period, and this can be tracked back to create a
“whitelist” of top-performing publishers.
• Creative: advertisers can evaluate the results of each creative in the campaign,
including the ad size, enabling optimisation to the best performing creative, whilst
balancing the available inventory at that size (and cost) to give sufficient reach.
Having a holistic view of the campaign across audience profiles, channels, inventory and
creative messaging, rather than viewing each in silo, is critical to using data to fully
optimise the campaign and maximise the return on investment.
Insights and reporting
Data should be used to inform, learn and help make decisions. In order to do that, data
needs to be organised and digestible.
Insights can provide the ability to an accurate and deep understanding of campaign
performance which is essential in decision making. Insights can also provide an
understanding of not only who committed an action (purchase, sign-up, etc.) but also
how that information ties back to what is already known about the consumer and/or the
audience segment; how that audience indexes against the buyer’s entire ecosystem and
what other attributes rank highly for that particular audience.
Closed loop reporting is a tactic that can be used to better understand the most valuable
customers and “close the loop” between marketing and sales efforts.
.
Page 22

UNDERSTANDING
AUDIENCE DATA
Overview
Kraft Heinz, its agency and DSP, wanted to increase consumption of its soup products
among canned-soup buyers and drive trial of a new carton of soup to win back share from
the fresh soup category in the UK.
Strategy
To meet the challenge, Kraft developed Global Purchase-Based Targeting (PBT) strategy, an
innovative method of leveraging real purchase data in the U.S. to inform and power global
audiences—filling the void where international purchase-based data is sparse.
On average, Global PBT proves to be more 3x effective at identifying brand and category
buyers. Ideal buyers were developed from several premium data assets including:
• CPG transaction data, sourced from 74MM U.S. households
• Cookie-based activity and interest
Results
The teams observed strong performance against audience reach (15-20MM online
profiles), viewability (60 percent average viewability higher than the EMEA industry average,
55 percent), and above-average video completion (75-80 percent). Kraft Heinz and team
will continue to leverage Global PBT for additional campaigns across France, Germany,
and Spain.
Global PBT solution has become a competitive advantage for us. It’s enabled us to
connect o international buyers with greater accuracy and outperformed other targeting
segments in market. We’re looking forward to leveraging Global PBT to enhance other
purchase-based targeting efforts across EMEA.
Nick Bartlett, Digital Development Director, Starcom UK
Case study
Accelerating international purchase-based
success
“ “
Page 23

The tendency for today’s consumer to own not just one, but a host of connected devices has
opened up a range of new opportunities for advertisers, while at the same time making their
ability to holistically gain control of the proliferated sales funnel a challenge. As one-to-one
marketing becomes an apex more advertisers strive to reach, the ability to find and target
users across the multiple devices they use is more important than ever. In tandem, advertisers
must also increasingly be able to directly attribute resulting business outcomes to those
marketing actions. Identity management or “cross-device” aims to marry the ability to target
accurately across devices with proper attribution of all the touch points along the path to
conversion.
Successful cross-device measurement yields big benefits for advertisers and consumers alike.
Effective cross-device measurement can:
• Fuel more relevant experiences for the user, delivering the most useful message at the
appropriate time, no matter which device they are connected to;
• Provide a holistic view of behaviour throughout the customer journey, allowing advertisers
to analyse advertising spend across an entire campaign, rather than piecemeal across
different screens and formats;
• Offer an essential step in eliminating silos in digital media planning.
The cookie has crumbled
The proliferation of connected devices means that single-device identifiers, such as cookies,
cannot capture a full view of media touchpoints and consumer actions across devices.
Not too long ago, cookies were an indispensable part of online advertising, helping advertisers
keep track of which ads had been viewed, on which site, in which behavioural segment the
user was part of etc.
4. CROSS-DEVICE
OPPORTUNITIES AND CHALLENGES
Page 24

UNDERSTANDING
AUDIENCE DATA
However, with non-cookie supported devices increasingly playing an integral role in the
lives of consumers, the dependency on cookies in regards to gaining useful insights is
steadily losing strength. The rise of non-cookie supported devices has rendered
traditional targeting and retargeting relatively tame, with one person owning a
smartphone, tablet, laptop, desktop, and smart TV being counted by advertisers as five
different users. This fragmentation of users has meant that the industry has needed to
adapt and rethink its targeting options, in order to enable advertisers to once again
effectively connect with their users.
One answer to the problem is the ID Graph. An ID Graph is a series of interconnected
identities, where all IDs belong to a single consumer. In other words, the ID Graph is a
united group of any number of separate (device/cookie) IDs that are tied together into a
single actionable customer profile. While not eliminating cookies totally (as they still work
on desktop), cross-device audience management works as an effective tool to retarget
and reach users across all of their connected devices. This new feature helps to determine
which devices belong to the same user, enabling advertisers to connect with these users
irrespective of which of their devices they are using.
Deterministic data vs. probabilistic algorithms
Advertisers can utilise deterministic data and probabilistic algorithms in order to
accurately identify which screens belong to users. This process must comply with EU
privacy legislation as outlined in section 2.1.
Deterministic refers to data that can be determined (i.e. verified) as true. For example, a
user logging in to an e-commerce site will usually need to sign up using identifiable
information. Thus, every time the user logs into this site, the publisher (website) will be
able to use this deterministic data recognise this user with an extremely high degree of
certainty.
Page 25

UNDERSTANDING
AUDIENCE DATA
For example:
User A owns a laptop, a tablet and a smartphone. On each of these he is logged on to a
certain website across his devices. His logins thereby create a cross-device connection
(through anonymous hashed login IDs), due to the fact that he is logged onto the same
profile from different devices. This data is one-way encrypted and fed into a DMP, from
where it can be accessed via a DSP as one and the same user.
The probabilistic method of establishing a user ID across multiple devices is a little more
complex than deterministic methods, as it relies on algorithms to analyse thousands of
different pseudonymous data points to create statistical, i.e. probable, matches between
these devices.
For example:
User B owns a laptop and a smartphone. The user works from home in the morning, goes
to work, and stops off on the way home to finish off some work in her favourite coffee
shop. She does this five times a week. A matching identifier here is her laptop and mobile
login using the same IPs and browsing the same content. This, and thousands of other
data points are compared to finally come up with a link that shows that these devices are
owned by the same user.
The deterministic and probabilistic data can be fed into a data management platform
(DMP), where consumers and the link between their devices are converted into actionable
consumer insights. These users are then accessible via a DSP to advertisers, who can re-
target consumers across their devices with relevant advertising.
Advertisers should talk about the opportunities and challenges of cross-device with their
technology partners to understand whether and how their approaches tackle each of
them, ideally selecting partners that have not only thought about these issues, but have
designed and implemented solutions that can demonstrably deliver solutions to this
challenge.
Page 26

UNDERSTANDING
AUDIENCE DATA
Effective attribution
It is important to understand how activity on one device can drive conversions on another.
Without cross-device methods only devices can be tracked, not consumers. This leads to
misleading statistics and campaign performance. Cross-device helps advertisers to gain
control of the attribution process and provides a 360-degree view of the consumer journey
and all associated touch points.
Measurement in a cross-device world
When it comes to finding users along the purchase path and taking stock of what they’re
doing at key touchpoints, the capabilities offered by existing solutions vary dramatically. A
good system of cross-device measurement needs to be accurate, scalable, actionable and
respectful of user privacy, and ideally should operate across a variety of properties.
• Accurate: in order to identify consumers across screens, accuracy is critical. If a solution
has no or few logged in users it’s going to be less accurate.
• Scalable: not only does a solution need to know some people, it needs to know a lot of
people. Fundamentally, cross-device measurement is an attribution challenge – but an
attribution model only truly succeeds when it can be applied at scale. This kind of scale is
achievable through programmatic, thanks to the huge reach it can provide.
• Actionable: in order to be actionable, cross-device solutions must be able to track
consumer interactions across the whole digital ecosystem. The wider the reach, the more
reliable results it will produce.
• Compliant: cross-device solutions should be compliant with EU privacy and data
protection laws and ensure transparency and control at all times (see sections 2.1 and
6).
Page 27

UNDERSTANDING
AUDIENCE DATA
• Across a variety of properties: users are often logged in on a number different
platforms at once. This makes it possible to layer user intent on top of other data,
offering a more complete picture than solutions reliant on a single property alone.
• Be aware that a single device doesn’t mean a single user in some instances, for
example, where a desktop PC is used by multiple family members.
Page 28

This section has been added to provide some context on the use of data in programmatic
under the General Data Protection Regulation (GDPR), the new EU-wide data protection
law.
The GDPR is a new regulation with comprehensive privacy and security requirements
intended to strengthen and unify data protection in the European Union. The GDPR was
adopted by the European Parliament in April 2016 and became law on May 25, 2018. It
replaces the EU Data Protection Directive 95/46/EC (Directive) and the national European data
protection laws implementing the Directive. As a single regulation, the GDPR is in part
intended to harmonise the different privacy and security requirements imposed by EU
member states. It also will help streamline compliance efforts that formerly caused significant
operational complications and expense for those doing business in the EU or otherwise
collecting personal data from EU residents.
What happens if an organisation fails to adhere to the regulation?
Organisations in breach of the GDPR can be fined up to 4% of their annual global turnover or
€20 Million (whichever is greater), this is assessed by the local data protection authority, such
as the ICO (Information Commissioner’s Office) in UK.
Expected impact on the advertising industry
The GDPR has implications for any industry that processes data of people within the EU’s
borders, regardless of their nationality or whether they are residents. Seven GDPR
requirements that are of particular relevance to digital advertising are outlined on the
following page.
5. GDPR CONSIDERATIONS
Page 29

UNDERSTANDING
AUDIENCE DATA
1 Defining personal data
The definition of personal data is broader under the GDPR than under the Directive. Under the
GDPR, personal data includes any information relating to an identified and identifiable natural
person, including online identifiers: “Natural persons may be associated with online identifiers
provided by their devices, applications, tools and protocols, such as internet protocol addresses,
cookie identifiers or other identifiers such as radio frequency identification tags.”
2 Lawfulness of processing
The GDPR requires companies to process personal data “lawfully, fairly and in a transparent
manner.” This requirement is multifaceted, but is centered around the need for organisations to
secure a legal basis to process personal data. For digital marketing data, consent is the prime
example of a legal basis to lawfully process data, but it must be freely given, specific to the
purpose for processing, informed, and unambiguous. When relying on consent, the marketer also
must demonstrate the individual consented.
3 Transparency
When collecting personal data from an individual, organisations are required to be transparent
and provide certain information to the individuals (including how their data may be used and
shared) to help them make informed decisions about whether or not they choose to share their
personal data. Together with its members, IAB Europe has developed a Transparency & Consent
Framework that allows first-parties to disclose their third-party vendors to users as well as inform
those vendors whether they have been given consent to process the user’s data.
4 Right to access and rectification
Individuals have the right to access and receive certain personal data upon request and to have
personal data that is inaccurate rectified or completed without undue delay.
5 Right to be forgotten
Individuals possess the right to have their personal data erased in certain situations. This right is
commonly referred to as “the right to be forgotten.”
6 Right to portability
The GDPR provides consumers with the right to have personal data transferred to another
controller when technically feasible.
7 Right to object
Individuals have the right to object to processing of personal data for direct marketing
purposes.
Page 30

UNDERSTANDING
AUDIENCE DATA
What’s next?
In addition to the GDPR, the proposed Regulation on Privacy and Electronic
Communications (ePrivacy Regulation) is expected to impact the digital advertising
industry. However, the level of impact and potential overlap with GDPR (for example in
terms of notice and consent requirements) remains to be seen. The ePrivacy Regulation has
not been finalised yet given the broad-ranging positions held by stakeholders. It is an update
to existing law, called the ePrivacy Directive – which already requires consent for accessing
end user devices (i.e. the placing of or reading of a cookie).
Where to go to learn more?
A great source of additional information is the local data protection authority, such as the
ICO (Information Commissioner Office) in UK and also the national IABs across Europe. IAB
Europe has developed the Transparency and Consent Framework to help all parties in the
digital advertising chain ensure that they comply with the GDPR when processing personal
data or accessing and/or storing information on a user’s device, such as cookies, advertising
identifiers, device identifiers and other tracking technologies. More information on the
Framework can be found here :http://advertisingconsent.eu/
Disclaimer:
The information in this document may not be construed or used as legal advice about the content, interpretation
or application of any law, regulation or regulatory guideline. Customers and prospective customers must seek
their own legal counsel to understand the applicability of any law or regulation on their processing of personal
data, including through the use of any vendor’s products or services.
Page 31

A global advertiser in 50 markets with 20 different brands and units will face different
regulations across markets and unexpected challenges. Therefore establishing an effective
programmatic data strategy takes time, resource and the right technology partners.
For an advertiser, it makes sense to replace the word ‘data’ with the word ‘consumer’.
Essentially, data becomes valuable for marketing if it helps to understand consumers and
their needs, helps to reach them with more relevant messages, and helps to track and
optimise those engagements.
As more and more of the purchase journey moves into digital and trackable channels, more
data about the consumer can be collected and the smart advertiser can use it to improve the
consumer experience and the marketing return on investment. Advertisers may be
questioning where to start and a good process may be as follows.
Step 1 - conduct a data audit
Evaluate which percentage of the target audience visits the advertiser’s digital assets, i.e.
websites, apps, online shops. Observe the intensity of their engagement and whether very
different behaviours can be seen that could be the basis for first party data segmentation.
It is important to understand the scale and depth of the first party data available. Once this is
understood the next stage is to look at the third party data available in the market: how deep
and broad is it and is it available at scale? Consider looking beyond the audience data
marketplaces: can a second party data partnership be established i.e. with retail partners,
media owners or other brands?
The output of this data audit should provide a good understanding of the available data that
could be used for segmentation and targeting.
6. CONCLUSIONS
Page 32

UNDERSTANDING
AUDIENCE DATA
The advertising mechanics of the brand should also be considered: does the advertiser
sell to nearly everyone, relying on reach and availability to drive growth, or would better
targeting of well-defined segments have a higher impact?
Step 2 - start small, learn fast
Many advertisers currently get very excited about the data opportunity and want to move
from mass-marketing into one-to-one personalised communication as fast as possible,
collecting all data they can get their hands on and often invest a lot of money into a data
management platform (DMP).
While ultimately most brands will have a DMP to make all of their customer and prospect
data available for insights and targeting, a lot of data usage can be leveraged without high
upfront investments. For example, most DSPs have data management functionalities
built-in and are connected to most third party data providers. Segments can be built,
combining first and third party data. Running tests on different data sources and
segments will show how the targeting data uplifts campaign performance and can be the
foundation of the business plan to invest more in data infrastructure and strategy.
Step 3 - move to a consumer-centric marketing organisation
Ultimately, consumer data and all the technology around it will only create value if it
helps to reach the right consumer with the right message across all touchpoints along the
consumer journey. That should lead to a better consumer experience and higher
marketing ROI.
Page 33

UNDERSTANDING
AUDIENCE DATA
Data, however, is only one ingredient for success. Breaking up silos, changing the mindset
from a product focus to a consumer focus, moving from a campaign organisation to an
always-on communication when and where the consumer wants it and moving into much
more analytical and data driven decision making are core components that need to be
considered.
This is a journey to the marketing organisation of the future and data is only an enabler.
Move in small steps, test, learn, fail, try again and transform your business as your data
confidence grows.
Page 34

IAB Europe would like to thank the contributors that helped to edit and compile the
update to this white paper:
Publishers
White paper leader: Ben Geach, Global Product Strategy
Director, Oracle Data Cloud
Thomas Park, Senior Product Director – Advertising
Products, Adform
Chris Hartsuiker, Manager, Privacy and Public Policy,
IAB Europe
7. WITH THANKS
Page 35

UNDERSTANDING
AUDIENCE DATA
Dino Bongartz, Vice-Chair, IAB Europe Programmatic
Trading Committee and CEO, The ADEX
Ashu Mathura, Product Director DMP, Adform
David Goddard, Global Head of Programmatic
Trading, BBC Advertising, BBC Worldwide
Martin Bromfield, VP Advertising, comScore
Theng Vogels, CEO and Co-Founder, DMA Institute
Matthias Matthiesen, Public Policy Manager, IAB
Europe
Oliver Gertz, Managing Director Interaction, EMEA
and Programmatic Lead Global Clients, MediaCom
Worldwide
Emma Williams, Senior Director, Regional Marketing
EMEA, APAC & LatAM, MediaMath
Artur Banach, Programmatic Task Force IAB Poland
and CEO, Netsprint
Ed Weatherall, DMP Strategy Lead EMEA, Rocket Fuel
And the white paper contributors that provided the content for the original version of this
white paper:
Page 36

About IAB Europe
@iabeurope
/iab-europe
www.iabeurope.eu
IAB Europe is the leading European-level industry association for the digital
advertising ecosystem. Its mission is to promote the development of this
innovative sector and ensure its sustainability by shaping the regulatory
environment, demonstrating the value digital advertising brings to Europe’s
economy, to consumers and to the market, and developing and facilitating the
uptake of harmonised business practices that take account of changing user
expectations and enable digital brand advertising to scale in Europe.
CONTACT
Marie-Clare Puffett | puffett@iabeurope.eu
Business Programmes Manager, IAB Europe
JOIN IAB EUROPE TO GET INVOLVED
All IAB Europe members have the opportunity to contribute to our white papers
and outputs. To find out more about membership please contact:
Alison Fennah | fennah@iabeurope.eu
Executive Business Advisor, IAB Europe

IAB Europe Report: Using Data Effectively in Programmatic V2.0 (GDPR Update)

More Related Content

Similar to IAB Europe Report: Using Data Effectively in Programmatic V2.0 (GDPR Update) (20)

More from Romain Fonnier (20)

Recently uploaded (20)

IAB Europe Report: Using Data Effectively in Programmatic V2.0 (GDPR Update)