Vodafone and data warehousing: A translation of that abstract

<cynic mode firmly switched on, with apologies to Rachel>

Via Mauricio, it seems from this, that Vodafone  are using data warehousing techniques to drive customer retention. I’ve had a go at interpreting the abstract of “The Customer Retention Journey at Vodafone New Zealand” below. Speaker Rachel Harrison is the Vodafone Lead Analyst who is giving (in Las Vegas) and gave the presentation – she is a SAS and data warehouse pro. Apparently the presentations were made avaialble to Teradata Universe conference atendees – so if you are resourceful then you may be able to find one.

A translation of the abstract

Vodafone is the leading mobile provider in New Zealand (NZ). NZ currently has 2 mobile competitors and over 100 % market penetration.

We are part of another cosy Kiwi duopoly. It’s nice – secure profits for both companies. Isn’t the iPhone cool?! we have the most expensive plans in he world!!

NZ’ders are early adopters of new technologies and our products constantly change and evolve.

But not as fast as other parts of the world which have real competition – see below.

We now have a shift to focusing on Customer Retention.

Because Telecom is shortly going to “compete” with us using current, rather than obsolete, technology. Our collegues in Europe are working on much crunchier problems, which frankly we would like to work on as well. Here’s a quote from collegue Nebahat Donmez : “In highly saturated telecommunications markets with a lot of competition, post-pay consumers revise their contract decisions easily anytime without giving much importance to the brand”.

With the completion of our new Teradata Enterprise Data Warehouse (EDW) we are well on the way in our Customer Retention journey.

We’ve built a big expensive data warehouse, justified on customer retention numbers (a little change means a lot of dollars), but have not done anything with it yet. It’s nice to be in the news for something other than our lousy billing system transtion though. Oh – and this may have been a bit of a corporate manadate as it seems Vodafone use this world-wide.

The new EDW has richer data available to constantly improve our retention programmes.

There is heaps of data in the warehouse and we can’t wait to get going on it.

New products and behaviours flow into the EDW for use in modelling and analysis.

Even more data is being added to the warehouse every day. It’s pretty overwhelming actually.

KXEN has enabled us to develop models quickly to ensure we are constantly including new customer behaviours.

We outsourced a bunch of our work to KXEN, “The data mining automation company” and are still paying them today. It seems Vodafone use them all over the world as well – nice contract.

We now have an iterative modelling process that allows us to keep up with the changing market.

We are a group inside the monolithic Vodafone that analyses and helps retain clusters of people and fee groups. We focus on keeping the ones that give us lots of money and look at things like the intensity of a cusotmer’s calling circle along with credit history. But we are a really smart bunch, and what we really wish is that we were allowed to craft products that deliver what people really want rather than just focus on stopping them from leaving when our policies and procedures piss people off. We can’t, so we do what we can.

SAS has direct against to the EDW and allows us to push queries back to Teradata to speed up analysis.

We can keep using SAS, thankfully, to do our analysis without having to resort to the Tertadata interface that is new.

The retention journey is ongoing; however with our new EDW and toolset we are confident of success.

We have had no results so far. Sorry.

The Warehouse also gave a talk at the Teradata Universe- Ray Renner is the Stock Systems and Process Manager – His Abstract was simply:

Leveraging Teradata Demand Chain Management in a Discount Department Store environment.

much better.

Published by Lance Wiggs


6 replies on “Vodafone and data warehousing: A translation of that abstract”

  1. data warehousing techniques to drive customer retention.

    Data warehousing is not a technique for analytics. It is a storage for cleansed data to be used in data-mining. Data mining is the technique that drives predictive analytics.

    I have to say that KXEN is perhaps the best predictive analytics vendor today (if not, they’re definitely in the top 3). I know (not personally) the KXEN CTO and founder ERIK MARCADE where we were part of the industry expert group that drafted the JDM 2.0 (Java Data Mining version 2) specifications for the official Java technology. There were 2 members from SAS that also got involved in the expert group.

    I have to say, that data-mining is huge today including its adoption by corporations for business intelligence applications such as customer retention.


  2. I noted that Rachel Harrison is not a data mining analyst. In fact she is a statistician as far as I know. Most people confused between a statisticians and predictive modeling analyst (data-mining, machine learning, etc…) as the same thing. KXEN, SAS tools are more heavier towards predictive modeling / data-mining / machine learning than statistics. Statistics is more of hypothesis testing (deductive) while predictive modeling / data-mining / machine learning is more of discovery (inductive) and tasks as customer retention modeling is inductive, ie, to discover the retention pattern.


  3. Falafulu
    thanks for the as always insightful comments. I had not heard of KXEN – my last really heavy duty experience was in the US at a huge credit card bank – and they were pretty wild users of straight SQL.


  4. I hope your current client is not asking for too much if that post demonstrates your understanding of BI and analytics technolgies…

    “…and now helping out at level or two higher with some BI strategy work…”


  5. BI Strategist that quote refers to very different work in a very different field.

    That work is about turning around businesses. Specifically it was about how businesses can dramatically simplify their approach to improving their results. Not a database in sight.

    It’s a far cry from data mining and the like, but the the lesson of simplification or both focus and approach is true across both fields.

    Even at huge US Credit Card bank where a lot of the leading edge stuff was and is developed (Talk to the Homeland security guys now I suppose) the tools were less important than the thinking.


Comments are closed.