3. Data strategy :
Create 360° consumer insights
and turn it into (real time) action
Unregistered
User
Registered User
Single copy
buyer
subscriber
member
Value(ARPU)
Customer journey’s
Engagement
LTV
4. Our program is build around 5 components
Collect Store Control Analyse Act
5. Collect
Data silos prevent a 360° single consumer view
Collect : Anonymous > login > social > profile
Sign in process is crucial
6. Make data available centrally
A new data architecture is needed
Avoid inconsistency in different applications
Store
7. Control
Data Governance to enhance customer experience :
Quality, Curation, Completeness
Data privacy and -protection to avoid leakage
8. Analyse
Make your data speak
Dashboarding = story telling
Descriptive and diagnostic
Datamodelling = scoring consumers for action
Predictive and prescriptive
9. Act
Without action, data is just a bunch of data
From multi-channel to omni-channel
Start your use cases from ‘act
10. Change is not easy
Organizational challenge
It is time- and cost consuming
Tools and techniques evolve quickly
Inject new skills, train your (data) people
Start small, create evidence : you can’t define
everything upfront
11. 11
We love DATA for it’s BIG return
Not for it’s big complexity ;-)
dirk.milbou @ persgroep.be
@dirkmilbou
Notes de l'éditeur
I aim dirk milbou,
Working for almost 4 yaers for de persgroep and responsible for CRM and data
I will try to summarize for you these 4 years in the next 7 minutes
Just a small word about De persgroep :
Today De persgroep is active in 3 countries with 19 different newsbrands and 15 magazines. Togethher they represent 65% of the turnover
In our data strategy, we’re talking about 20 mio unique active consumers
From these 20 mio unique consumer records, the majority is anonymous at the left under.
We like to move them up to the top right, where tey are engaged to consume our brands, buy merchandising and can be easily and sharp targeted by our advertisers, regardless the platform they are using.
To support this strategy with data, we started 3 years ago an internal program, which at the beginning, this program was embedded in the audience marketing department.
But since the fundamentals are there, we started to develop an integrated data strategy for audience, editorial and advertising.
In this approach we are following 5 building blocks : collecting data, store and control it, and after analysing the data turn our insights into appropriate actions.
I will run to each step in the next slides and give you some thoughts and learnings on each building block
At the beginning, many different data silos prevented us to get an holistic 360 degree view on our consumers and inconsistent data diluted the consumer image
So we started to bring all data, residing in different systems into one single environment, matching and deduplication all records into 1 single consumer view to create a golden record.
We didn’t replace those legacy systems, but prepared them for integration to the central datawarehouse
And then we started to collect new data accordingly. Crucial element in this is of course the application to make digital users to register. We decided their to work with an external partner to revamp our sign on process, and make sure we gather data easily tru social logins and collect social interactions properply.
As soon as the customer golden record exists; it should be made available for all consumer related processes,
whilst at the same time conforming to data protection legislation and other regulations.
Of course we try to avoid duplication, not only for the cost of storage, but more important to to avoid inconsistency of our data in the different applications and create a seamless customer experience.
But after one year, we decided to establish a new data architecture
- storing hard data into a central oracle system and all clikcs and interactions into a hadoop environment, combining those two into a customer analytical record, available to our data scientists.
- We decided to rebuild it into more granular components to enable us to have a daily load instead of a bi-weekly refresh of the customer analytical record
The third step in our approach is to make sure that all tools are in place to guarantee an appropriate data governance.
Data governance is all about data quality and data completeness, and enrichment, those components make the value of your data
Data governance is also more and more about compliance to privacy rules and data protection.
What we learned here is that matching and cleansing first party data and completing with second and thrid party data is a tough jobs which requires specific skills and applications and business rules.
Data governance needs data stewards in every domain, so as well in audience marketing as in the advertsing or the editorial department
On top of our single consumer view and customer analytical record, we builded a specific layer to build dashboards and models for scoring consumers.
Dashboards are different from classical weekly or monthly reporting :
a good dashboards uses dynamic visuals and grapichs to tell a story and to explain complex data insights to non-analytical people, so the data can drive decisions
A good dashboard also should give trust in the data for the different stakeholders and give them a feel of control
In the same compute environment our datascientist build models. We have actively a predictive churn and sales model in place and use it succesfully to differentiate our offers accordingly to these scorings
Profiles for business analysts and consumer analysts are different.
Business analysts make descriptive and diagnostic analytics, describing what happened and why it did happen.
Where consumer analysts describe with predictive models what will happen and with prespcriptive models how we can make it happen.
We learned that it is very important that they all are well connected to collaborate and share learnings and best practices, across countries, domains and silos.
The last and most valuable step is to turn these insights into an action.
Without action, your data is just a bunch of data and has no value at all
Today’s customers use an increasing number of channels to have contact with you, and they expect a seamless transition between offline and online offers and across all touchpoints
This required us to install new tools for audience, advertising and editorial to serve our consumers relevant content and offers, according to their data and the segments to which they belong.
To get these things developed and working, we decided to change our breifings to ICT and digtal departments from ‘please collect all possible information everywhere’ to more dedicated and specific use cases explainging ‘I need to know this from my backend, so I can do that in the consumer facing frontends
Organizational challenge : it is important to involve end users throughout the program and not just during requirements gathering and deployment phases
Be aware that this active involvment comes with a cost, don’t try to save this cost because it will bacfire afterwards
You know that tools and techniques evolve quickly and are often being adopted ahead of the learning expertise. That’s why it is important to inject new skills both in business and ICT and to invest and train your key data personnel.
Last but not least : start small, create evidence tru experiments as you can’t define everything upfront. You need several iterations, change routes and priorities along the way
To summarize :
Data, and especially big data is considered as the new oil for publishers.
So you have to make sure you build a solid raffinary to distract gasoline from it
And you have to make sure your empoyees and even you advertising clients have easy access to petrol stations where they can load and use the gasoline to drive their business.
It is complex indeed, but the return both in preventing costs and get new revenu is certainly there, especially when it enables you to offer a seamless customer experience.
THANK YOU for listening