Colab 2019 Making Sense of the Data That Matters

SPONSORED BY
MAKING SENSE OF
THE DATA THAT MATTERS
KEY CONSIDERATIONS FOR
DATA DRIVEN ORGANISATIONS
OCTOBER 2019

MAKING SENSE OF
SPONSORED BY
INTRODUCTION
The data organisation is locked in the grip of profound change, and nowhere is
this more true than in the media owner space. The digital era was once thought
to answer all of our data woes. By generating terabytes of behavioural data
points that could be seamlessly integrated with multiple other data sources and
delivered in insightful real time dashboards, business units can become
empowered to optimise performance.
The truth, however, is a little more nuanced.
Data drives organisational change, and it is unlikely you’d find a business in the
FTSE 100 which wouldn’t describe itself as a data driven organisation. Yet at the
same time, the proliferation of data and the ubiquity of the tools used to tease
insight from it, has prompted fragmentation of data ownership across multiple
business units.
The challenge with this fragmentation is that where there is not joined up
thinking at an organisational level on how data should be integrated, analysed
and actioned, at worst the business’s use of data becomes aspirational only. It
becomes fundamentally backward facing, reporting and analysing and justifying
what has been, rather than trying to impact future performance.
Those business which are truly transformative in their application of data are
fundamentally forward facing. They use data to forecast and make the most of
predicative capabilities to optimise performance. This paper will explore the
experiences and opinions of a range of media industry experts in relation to the
business critical challenge of making sense of the data that matters.
1
INTRODUCTION
2
EXECUTIVE SUMMARY
3
KEY ISSUES IN MAKING
SENSE OF THE DATA
THAT MATTERS
4
CONSIDERATIONS
FOR PUBLISHERS
AND VENDORS
5
ACKNOWLEDGMENTS
2
“Analytics is not data-driven if its findings are never seriously considered or
acted upon. If they are unread, ignored and the boss is going to do whatever
he or she wants, then they are ineffectual.”
Carl Anderson
Creating a Data Driven Organisation

MAKING SENSE OF
SPONSORED BY
Source: Revenue Operations Barometer Pro Survey H1 2019, CoLab
REVENUE OPERATIONS
AND THE NEW CUSTODIANS
OF ORGANISATIONAL DATA
Revenue operations teams increasingly find themselves the custodians of digital audience data that was once
owned by business insights and research teams. The ad tech tools which enable revenue operations teams to build
and monetise digital audiences are converging with martech tools to generate strategic audiences insight that has
applications across all areas of the modern data organisation.
CoLab Consulting’s Revenue Operations Barometer, launched in April 2019, sought to gain the perspective of this
often under-analysed group by interviewing rev ops professionals from over fifty leading global publishers. While
business priorities and challenges varied by organisational type, the one consistent finding highlighted was that
improved analytics was seen as the top priority over the coming year.
3
Analytics is a broad field however and covers a range of uses,
from the descriptive and exploratory to the predictive and causal.
To unpick this question further and to shine a light on how leading
publishers are currently making sense of the data that matters, a
roundtable of industry experts was convened in September 2019.
The key themes that emerged are summarised in the following
paper and provide a vital insight for any business in the broader
publishing ecosystem.
How will you prioritise the following areas of the business
over the next 12 months? % High Priority
60%
Improved Analytics
31%
Sales Order Management
23%
Team Size and Structure
15%
DSP
19%
SSP
31%
Consent Management
21%
DMP
10%
Ad Server
James Gilkes
Global Pricing and Inventory Director
BBC
David Evans
Head of Data and Insight
CNBC International
Adjani Nicholson
Sales Manager EMEA
IDG Solutions
JoaoFelizardo
Head Program Manager
IDG UK
Brett Gibson
Client Advisory and Value Consulting Director
Domo
Rupert Staines
Digital Marketing, Data and Technology Expert
Duncan Arthur
Partner
CoLab Consulting
Anne Goodman
Associate
CoLab Consulting
Ian Gibbs
Associate
CoLab Consulting
Rebecca Rangeley
Head of Business Insight
Freewheel
AlbertoSantangelo
Director of Multiplatform Ad Operations
Viacom
Danny Doyle
Head of Ad Operation
Hearst UK
David Hayter
Programmatic, Data and Technology Director
Stylist Group
Roundtable Expert Panel
The roundtable took place on the morning of the 10th September 2019 in London, UK and attendees covered a range
of revenue operations, digital product and business insight roles:
Publisher
Do the key themes and
challenges ring true for
your business and what
are you doing to make
sense of the data that
matters?
Vendors
How can you help
publisher make sense
of the data that matters?

MAKING SENSE OF
SPONSORED BY
SEVEN KEY THEMES FOR
DATA DRIVEN ORGANISATIONS
EXECUTIVE SUMMARY
1. Data fragmentation is a key challenge for all. Fragmentation of data sets and
analytics professionals creates organisational inefficiencies and works
against data driven decision making.
2. What do we really mean by data integration? Whether it be the centralisation
of analytics resource or the integration of discreet data sets, it must be
a process baked into organisation culture if it is to succeed.
3. Automation increases human resource burden. Perhaps counter-intuitively,
data automation can result in greater business interest in analytics,
increasing demands on analyst time for higher level analytical work.
4. Centralisation is favoured over shared ownership of data. The centralisation
of data after the expansion of cloud business applications aims to govern
and expand access without restricting the decision-making process.
5. Senior management needs help understanding the business case for
data integration. Without senior management buy-in, data integration
and analytics initiatives will have little longevity or may fail due to
underinvestment.
6. Marketers are hooked on over-abundant behavioural data in campaign
reporting. This is useful data for campaign reporting, but it tells us little
of medium- or long-term value for advertisers.
7. The first-party future. Dependable behavioural data and declared profile
data will help publishers maximise yields in the long run.
4

MAKING SENSE OF
SPONSORED BY
1.
DATA FRAGMENTATION IS
A KEY CHALLENGE FOR ALL
The fragmentation of data and the tools used to draw insight from and monetise it, was a key talking point in our
discussion. Fragmentation can take many forms, but primarily was seen to manifest as:
5
“The difficulty is fragmentation… lots of fragmentation of data [across] different data points. Data is dirty: although
there are rich data sets, finding the truth is really quite difficult”
Roundtable Attendee
In the case of point A, the focus of many expert
panellist’s businesses has been to harness multiple data
sets through data warehouses, DMPs and data lakes, yet
this is a process that takes up varying levels of resource
depending on organisational size. In the words of one
panellist the search for perfection in this area (i.e. all
organisational data perfectly linked in one single,
accessible, system) can often result in a poor effort
vs reward trade off when generating business insight.
In the case of fragmented and siloed data ownership,
the downside to the business is organisational
inefficiency. There will be a commonality of metrics
across ad servers, web analytics systems, DMPs and
third party analytics tools which will not only result in
overlap of effort, but will invariably differ from one
another too – creating multiple versions of the truth and
creating confusion in the process of how to action data.
The most valuable data to revenue operations and data teams requires little processing and additional manipulation
and is tangibly linked to ad dollars, both for delivery optimisation and selling the value of ads more broadly. However,
in a world of data fragmentation, the challenges in getting to this point are abundant.
A.
Discrete data sets existing in separate tools
and systems
B
Data and systems ‘owned’ across multiple
business siloes

MAKING SENSE OF
SPONSORED BY
2.
WHAT DO WE REALLY
MEAN BY DATA INTEGRATION?
If the solution to data fragmentation is data integration,
then consensus must be reached on what exactly this
means within the organisation.
From a technical data point of view, does data
integration mean the linkage of multiple data sets at
their rawest form, enabling once discreet data points to
be cross referenced and cross analysed from a single
source in a data warehouse? Or is it enough to make
discreet data sets accessible via a common portal,
allowing business areas to view and analyse data in
parallel using a common data vocabulary?
6
“The data warehouse where raw data [is organised]
into buckets, where you just push a button and say
“I would like to know this” is the dream. But the
dream would be that it would not just be one person
pushing the button… everyone is able to do it. You
give people access to the reporting so they become
independent and autonomous.”
Alberto Santangelo
Director of Multiplatform Ad Operations, Viacom
The latter point in particular speaks to the power of
data visualisation managed within business units, with
many panellists discussing their successes and failures
in implementing dynamic and engaging systems for
presenting data to the business. On the one hand such
systems democratise data by opening it up to the entire
organisation, but such shared usage does not suit
everyone, and it is acknowledged that there are some
people who will always have their own unique
requirements and dependence on older systems due to
either loyalty or technical debt. There was some debate
as to whether it was reasonable to expect that a single
source of centrally managed and curated data could be
agile enough to meet the diverse demands of analysis
across the business.
The expert panel was of the view that while many
tech vendors sell the promise of truly integrated
organisational data, many fail to live up to the promise.
This was particularly true of data that is used by more
than just one team and therefore might be depended
upon for multiple purposes, as revenue systems are not
oriented to provide capability to answer marketing or
financial questions despite having some commonality
of the underlying data. A stark warning in terms of
vendor credibility in this space.
3.
AUTOMATION INCREASES
HUMAN RESOURCE BURDEN
A core benefit of data integration, management and
the development of self -serve analytics tools for the
business is that it removes the burden of day-to-day data
reporting and queries from data analysts and, in the case
of our expert panel, revenue operations people. Their time
can instead be freed up to shift their focus towards
analysis and insight generation, answering strategic
rather than tactical business queries.
There was a distinct sense amongst our expert panel,
that processes that automate data delivery may not
ultimately reduce the burden on human resource. In fact
in many instances the burden increased as more in depth
insight generation sparked more in depth follow up
queries from an insight-hungry business. However the
demand was addressed, whether through increased
resource or some level of unmet demand, the resulting
business value was significantly greater from being able
to ask questions rather than battling to tame data
preparation, quality and integration challenges.
Data -driven organisation must be ready for the increased
demands for their time that their success will generate
and in so doing must acknowledge that there are two
distinct skill sets required in terms of human resource
when making the most of data assets: Those who extract
the data vs those who derive insight from it.
“There are two separate skill sets. One is extracting
data and one is extracting value from the data.”
David Evans
Head of Data and Insight, CNBC International

MAKING SENSE OF
SPONSORED BY
4.
CENTRALISATION IS
FAVOURED OVER SHARED OWNERSHIP OF DATA
While the physical and virtual integration of data sources goes some way to overcoming the hurdles associated
with fragmentation, our expert panel was almost unanimous in their belief that centralised data teams are also
a vital component of the contemporary data driven organisation.
Pockets of data, insight and reporting currently exist within multiple business units – from marketing, to product
teams, to revenue operations and business intelligence teams. The resulting inefficiencies are often borne out of an
overlap in resource and datasets, but potentially more damaging to the business are the competing and sometimes
conflicting data narratives that results from different sources and analytical techniques.
7
“There’s data from many sources in any organisation.
The immediate issue is that the data in itself,
regardless of where it comes from, is an organisation
within the organisation. If you ask everyone around
the table who have different operational
responsibilities [for data], even the CEO, you’re going
to be asking for different sets of data.”
Rupert Staines
Digital Marketing, Data and Technology Expert
Our panel highlighted that a centralised data team’s core
purpose should be to define how the organisation needs
to perform in terms of data driven decision making and
enable access to data even if not centrally provisioned.
There is a fine balance between democratizing data
throughout an organisation to empower business units
with relevant insight accessible at their fingertips, and
ensuring a consistency of quality in how that data is
applied strategically. A centralized data team should
not necessarily look to “own” all organisational data,
but rather define best practice in how it is analysed and
deployed. It recognises that different team structures,
skills and the complexity and cadence of business
questions means that a one-size-fits-all approaches
won’t work for all requirements.
Ultimately centralisation should improve data access
and data decision making for the entire organisation.

MAKING SENSE OF
SPONSORED BY
5.
SENIOR MANAGEMENT NEEDS HELP UNDERSTANDING
THE CASE FOR DATA INTEGRATION
This was a theme echoed by our expert panel and one
which ultimately manifested itself in two ways.
1. Business processes relating to data and analytics
projects were often deemed to be sub-standard.
More than one expert panellist cited examples of a
poor brief at project kick off stage resulting in
inadequate analysis that failed to get to the heart of
the business’s burning strategic questions. An
inability to iterate through the process locks in any
early gaps in knowledge.
2. When being asked for budgets to conduct large scale
data infrastructure initiatives such as building a data
warehouse (a vital and powerful tool in enabling
organisations to understand relational data and
driving decision making), CEO’s, CFO’s and CIO’s
often do not have a true understanding of what they
are being asked to sign off.
Scoping data management initiatives up front in terms
of tangible business value is critical to help the C-Suite
understand the business case for enhanced data
projects. Multi-year data lake and warehouse projects
should not be undertaken without clearly defining exactly
what return the organisation will gain from the initiative
(return from both a data perspective and an ROI
perspective) and a timeline of when benefits can be
expected.
8
“We are successfully building diverse revenue streams” 52%
AGREED
“We have an effective suite of technology to support
all advertising operations”
48%
AGREED
“The company invests adequately in ad
operations technology”
46%
AGREED
“Senior management understand this area of the
business well and give it the attention it deserves”
37%
AGREED
“We will require new ad operations technology to maximise revenue” 63%
AGREED
With a range of businesses on our expert panel, it was
clear that the younger digital-era businesses were more
capable of being nimble in this regard, scoping and
executing on data initiatives in weeks rather than months
and providing insight in to the return for the business in a
far shorter time frame than larger organisations.
To truly drive success in data initiatives our panel felt
that the resulting data processes must be baked into the
sales organisation – for example, tied to performance
and bonuses. Furthermore, self-service should be
mandatory for certain types of data queries if the
organisational analytics tools have been built, enabling
the centralised data team to focus on higher level,
specialist, analytics output.
It is this process of linking data initiatives to individual
performance that will drive understanding of data
management at a board level.
To what extent do you agree or disagree with the following statements? % High Priority
The CoLab Revenue Operation’s barometer
reveals only 37% of leading publishers
agree that senior management understands
their area of the business well and gives it
the attention it deserves.

MAKING SENSE OF
SPONSORED BY
6.
MARKETERS ARE HOOKED ON OVER-ABUNDANT
BEHAVIOURAL DATA IN CAMPAIGN REPORTING
Given the professional interest area of our expert panel, the topic of campaign and digital advertising effectiveness
reporting was top of mind when discussing the data that matters most to publishers.
For revenue operations and sales professionals, a large amount of their time is used to generate post campaign
reports for clients, often aggregating campaign data from multiple sources (ad server, web analytics and social
data for example) in what is often and time consuming and cumbersome task.
While there was clear demand therefore, for tools that integrated multiple campaign data sources, it was also very
telling that the many agencies and advertisers and still overly-dependent on behavioural data such as clicks, likes
and shares, rather than data points that enable true ROI reporting (such as sales uplift), or report on longer term
brand building goals. This is a well reported problem that has driven much of the over-commoditization of digital
ad inventory in the past decade.
9
“One thing we still struggle with is that manual
process. Clients have access to all of this data and
all these campaign reports, but are calling us and
saying ‘I’m not getting what I paid for what are you
doing about it?’ … We have this great tool that’s going
to give you loads of insight and is going to tell you a
great story, but I still need someone to tell that story.”
Joao Felizardo
Head Program Manager, IDG UK
The expert panel discussed the notion that rather
than being solely dependent on short term behavioural
metrics, digital marketing effectiveness should be
viewed through a short/medium/long-term lens to truly
understand its value to the organisation and unlock
further revenue potential for publishers.
While this lens will naturally vary by industry (product
purchase cycles vary dramatically for high and low
consideration products for example), data and systems
that explore the effectiveness of digital beyond short
term direct response goals will have profound effects
of the digital ecosystem.
7.
THE FIRST PARTY FUTURE
While third party data still has a degree of value to
publishers while brands and digital buyers place value
in it, the longer term and more exciting opportunity for
publishers comes from leveraging first party data assets
for campaign targeting, optimisation and reporting.
Our expert panel discussed the differing value exchange
that takes place for varying types of first party data –
from behavioural page level data, to potentially more
valuable user profile information. The latter data set is
declared and probabilistic and does not rely on the same
deterministic modelling techniques that third-party
audience segments do.
First party data results in a smaller data set consisting
of higher quality data points and safeguards publishers
from the challenges associated with third party
measurement methodology changes.
“We’ve almost got away from using any third party data
at all… The Safari ITP moment [for example]: we have
no problems with that because we can collect data on
those users. We can collect data from those places
where cookies generally don’t work.”
David Hayter
Programmatic, Data and Technology Director, Stylist Group
With large proportions of publisher audiences now
existing on platforms where first party data plays a far
greater role (such as mobile apps and safari), data driven
organisations must align their systems and processes to
capitalise on this shift - one which is more relevant than
ever in a post-GDPR world.

MAKING SENSE OF
SPONSORED BY
CONSIDERATIONS FOR PUBLISHERS AND
VENDORS
FOR
PUBLISHERS
FOR
VENDORS
1
DATA FRAGMENTATION
Is data fragmentation impacting your
business and do you have strategies in
place to mitigate against it?
How can you help publishers with the
systems and processes required to break
down data siloes?
2
DATA INTEGRATION
How integrated do you want your data
to be?
Discreet data sets in one system or truly
joined up, queryable single-source data?
Can you deliver on the promise of true
data integration for a sometimes-cynical
publisher market?
3
DATA AUTOMATION
Is your data organisation sufficiently
resourced to meet the demand that
improved analytics will result in?
How can you help publishers create seamless
processes and tools to manage data and
insight demand?
4
DATA CENTRALISATION
Define how much “control” you want your
centralised team to have. Defining how
the business interacts with data is critical.
Systems that enable data democratization
across the entire organisation are compatible
with centralised data teams.
5
DATA BUSINESS CASE
Poor briefing and definition of scope
will increase the likelihood of data
initiative failure.
Build the business case for data integration
with your clients and prospects.
6
DATA FOR CAMPAIIGN
REPORTING
Centralised systems and dashboards to
enable campaign reporting are powerful
tools, but metrics that focus on real ad
effectiveness must not be ignored.
Can you help publishers report on campaign
ROI or brand impact – i.e. move beyond simple
behavioural metrics?
7
FIRST PARTY DATA
Are you collecting sufficient data to
realise the potential of the first party
data opportunity?
How can you help publishers collect, manage,
integrate, monetise and draw insight from first
party data?
10

MAKING SENSE OF
SPONSORED BY
ACKNOWLEDGEMENTS
We would like to extend our thanks to all participants
in the “Making sense of the data that matters” industry
roundtable, and to those who have given permission
for attributable quotes to be used.
ABOUT OUR SPONSOR
Domo is the fully mobile, cloud-based operating
system that unifies every component of your business
and delivers it all, right on your phone. Domo brings
together all your people, data, and systems into one
place for a digitally-connected business. For more
information about Domo, please contact
EMEAMarketing@domo.com ; visit www.domo.com or
call +44 203 8791006
11
ABOUT COLAB MEDIA CONSULTING
CoLab is a specialist media consulting and market
research business.
We provide insights and leadership at the intersection of
media and technology. Our consultants are digital media
veterans based around the globe who have held senior
positions at major media organisations including the
BBC, The Guardian and Microsoft.
We work collaboratively with our clients to deliver deep
sector knowledge, an impartial view on the best way
forward, and a plan rooted in practise not theory.

Colab 2019 Making Sense of the Data That Matters

More Related Content

What's hot (19)

Similar to Colab 2019 Making Sense of the Data That Matters (20)

More from Ian Gibbs (20)

Recently uploaded (20)

Colab 2019 Making Sense of the Data That Matters