Content:
1. A brief history of the FT
2. What does Big Data mean to the FT?
3. The benefits of Big Data & how we use it
4. How we do it
5. What’s next for us?
9. Measuring Reader Engagement
We look at reader behaviour over the last 90 days:
• Recency – when did they last visit?
• Frequency – how often do they visit?
• Volume – how many articles have they read?
Engagement scoreCancellationrate
More engaged readers are less likely
to cancel
14. “The analytics team (with support
from tech, commercial and third
parties) will explore ways of
finding value as a prerequisite to
building in new capability”
The FT’s “Analytics First” approach to Big Data
15. Team and organisational structure
Chief Data Officer
Analytics
Reporting
Data
Intelligence
Data
Science
Vertical
Specialists
Campaign Management
Data Strategy
Technology
Product
Research
3rd parties
Key supporting
functions:
Customers of
Data and Analytics
B2C and B2B
Editorial
Product
Finance
Advertising
Board & Strategy
17. What are we planning for 2018?
• In house attribution modelling
• Linking customers across all of our publications (FT specialist)
• Next level ads – more than just impressions, outcome based values
• More self-service – double usage levels; expose models to users.
• GDPR
• Expand KPIs to be more team focused goals: product specific goals and targets, interaction level engagem
ent.
• Dynamic pricing, revenue optimisation, yield management, better view of lifetime value (LTV).
• More segmentation of our audiences across the board: prospects, subscribers, advertising audiences, unm
asking the anonymous
• Do more cool stuff with data – quarterly innovation month.
Start off with a brief history of the FT – not everyone knows who we are, or might have a limited perception of what we do
Then talk about what big data means to us at the FT – what do we really mean by big data, and which of it is most useful
Next talk about the benefits we have derived from Big Data – real, tangible benefits
Then talk about how we do it – the structure of the team and the importance of cultural change and support
Finally what’s next – what are the challenges and opportunities?
And we’ll have time at the end for questions, or course
Most famous as a leading ‘financial newspaper’ – printed on pink news print ‘the pinkun’
Founded in 1888 – coloured pink to distinguish us from a key competitor
1995 – launched the website
2007 metered access model – view a certain number of articles for free (registration) + then subscription
It was very important we didn’t start giving our journalism away for free
Other benefit of registration is fantastic user level data – we could see usage across multiple devices. KEY!
Also launched our own web app, bypassing the app store? Why? Collecting our own data was key = we’ll talk about this at the end of the preso (distributed content).
Last year we shifted away from metered model to trials model – this decision was 100% data driven (talk about later).
We now have the highest number of paid for readers in our history – 770k across print, digital and B2B.
Data is the core for us – we now have chief data officer
There are 3 key types of data for us:
User demographics; we want to know this for Ads, marketing, billing, etc.. Is key – and if we don’t have it we, use data science to work it out (e.g. Gender data)
Behavioural – real benefit of the new world, so much is based on that and it often delivers the best results – e.g. marketing stuff.
Meta data – how do we explain what it all means? Historically we have not been great at this and it’s a focus. Very important as we take it to the next level – both for models, but also for the website, etc..
And of course, bringing this data together is vital – how to store for usage both for analysts and systems.
Mention Redshift and NGDA (collection as well as storage) – NGDA powers analysts, Chart.io and Lantern
Issue when we don’t have full control – e.g. Ads, takes ages to get stuff, and no historical data
All of this we could call first party data
We did some analysis last year and concluded that 80% of revenues in some way relay on this data.
We think of our data as a key asset – up their with the quality of our journalism and the brand
So here is a very clear example of how important data is to the FT!
Our CEO has set a very simple – but not easy!! – target of reaching 1 million paying subscribers (remember – we are not at c780k.)
We’ve got to where we are now off the back a strong business model, fantastic journalism and a volatile economic environment
But the next 220k will be trickier – print will decline, so digital needs to grow more than 220k
To do this we need to achieve 2 key things: engage existing customers in order to retain them: RFV
And find new audiences to sell to: Reach / QV
All investment now seen through this prism, data model driven by analytics, finance, strategy & the board
Here’s our very successful R / F / V model – stolen from retail, TBH
It works because we correlated it to canx
We have an important threshold –
This stops focus on people it’s easy to win over. After 200 is diminishing returns
Moving from unengaged to engaged is worth £108 in lifetime revenue
RFV also allows us to describe customers a lot more easily
Scale from engaged to unengaged
FT Fans vs. EFF
How to use this?
Alert for people moving from 1 group to another – DS model
Power marketing campaings
Measure next: key metric to track success and migration
RFV etc is very much as quite a high level, but what about more day to day
Data is a big way of powering personalisation on site. We have been a bit behind on this, but myFT is very much about using opted in and derived data to power that. Collecting data is key – preference centre, yes, but also collect data in context (B2B 25%, B2C 43%)
APIs = talk about HUI, taking what we know and feeding it to other systems
Editorial authority v important – majority of our subscriber journesy start on the homepage / 5% of digital traffic is ePaper
Use of lionel on barriers etc – wins on A/B tests, so editoirial authority also help us sel!!
Having the data, it all in one place and team of smart people – means you can innovate
E.g. business travellers segment came from request from Ads
HUI API stuff
Allows us to be at the forefront of discussions with Google, Facebook, Apple… etc..!
Meta data stuff
Risk of letting developers create stuff that isn’t joined up – it can be perfect but expensive
Analytics Frist focus helps us to best opportunities
Having analysts with commercial savvy, comms skills, etc.. Is VERY important.
A good lead data scientist does not sit in the corner just being clever,
Importance of Chief Data Office – highlights how important it is to the board
Talk through team
CDO role is as much about strategy and mind-set as direct reports
Technology is vital to support
Range of stakeholders – EVERY key department
Decent size team – 15 analysts + as many again supporting in tech, research, product etc…. 30 people (1.5% of FT total)
What’s next?
Data democratisation – push culture through org, unified set of data to support everything. Screen around office etc – always in people’s faces
How do we collect data from distributed content – very important
Testing – building in house capability, key driver for trials work, big cultural issue to organise it properly
What else – see list