• Belgian supplier of energy and
natural gaz to consumers and
profession users
• 4th largest player in Belgium
• 350 000 customers of which 24
000 professional
• Active since 2001
• 150 FET
It all started with…
Enjoy
If we know who is gonna call
us….
We could give the answers
before they give a call
3 A approach
“A”cquire
Data
‘A”nalyze
Data
Make it
“Actionable”
• What is the profile
of the calling
customer
• Which parameters
are important
• OUTPUT: algorithm
made by data
scientist
• Collection of data
• Quality check of
data
• Descriptive,
consumption
behavior data, Call-
data
SEEMS EASY, BUT IT
ISN’T
• Apply the defined
profile to NEW
customers with
highest risk of calling.
• HOW ?
Send a personalized
video via email with
all “relevant” data, for
which they normally
call.
• WHEN ?
Before the customer
recieves the invoice
Example of Email with video link in it
Send to those new
customers with the
highest probability of
calling.
learnings
• Guerrilla approach – no big project
• Mixed team, on top off daily business
• Focused innovation, DQ positioned as side effect-MUST
• “Guerilla” lead to attention for DQ towards right audience
• Engage employees for good DQ output
– Input of employees that generate the output
– Leads to a long term commitment
• More impact than big DQ initiatives – part of daily process
Big Data Marketing Seminar Essent Els Descheemaeker

Big Data Marketing Seminar Essent Els Descheemaeker

  • 2.
    • Belgian supplierof energy and natural gaz to consumers and profession users • 4th largest player in Belgium • 350 000 customers of which 24 000 professional • Active since 2001 • 150 FET
  • 3.
  • 5.
  • 6.
    If we knowwho is gonna call us….
  • 7.
    We could givethe answers before they give a call
  • 10.
    3 A approach “A”cquire Data ‘A”nalyze Data Makeit “Actionable” • What is the profile of the calling customer • Which parameters are important • OUTPUT: algorithm made by data scientist • Collection of data • Quality check of data • Descriptive, consumption behavior data, Call- data SEEMS EASY, BUT IT ISN’T • Apply the defined profile to NEW customers with highest risk of calling. • HOW ? Send a personalized video via email with all “relevant” data, for which they normally call. • WHEN ? Before the customer recieves the invoice
  • 11.
    Example of Emailwith video link in it Send to those new customers with the highest probability of calling.
  • 12.
    learnings • Guerrilla approach– no big project • Mixed team, on top off daily business • Focused innovation, DQ positioned as side effect-MUST • “Guerilla” lead to attention for DQ towards right audience • Engage employees for good DQ output – Input of employees that generate the output – Leads to a long term commitment • More impact than big DQ initiatives – part of daily process

Editor's Notes

  • #11 Acquire data: In the first phase we collected all data in order to make the analysis. This seems an easy job, but it isn’t. As the data is not centralized and not all in the same format. Quality is essential eg we saw that a % of our customers was > 112 years, due to the fact that this field in the activation form was not obligatory to change. Call agents did not ask it, or customers did not change it. Analyze data: With a datascientist using a logistic regression model, we tried to define the profile of “calling customers”: who are they, what parameters define a caller ? , when do they call ?. We made the profile based on historical data. The outcome of the analysis, the algorithm was trained, tested and evaluated. Actionable An algorithm is nice, but you need to apply it to the business The algorithm was applied to new customers Eg 1000 new customers 200 fit the highest probability of calling 100 received a personalised video 100 received no personalised video 100 received a video but where randomly choosen This in order to test if 1, the model was working ? 2, does the video has an impact on the calling behaviour
  • #12 Please add the video link to this email format !!!!