“A wise man learns by the experience of others; a fool, by his own.”

Gadi Solotorevsky — Mon, 23 Aug 2010 08:48:09 +0000

The use of cellular phones for mobile payments takes two main forms: using the cellular phone as an electronic wallet and as a credit card reader terminal.

With the first option, the phone simply replaces the traditional credit card.

The second option enables the phone owner to accept payments from traditional credit cards or electronic wallets. Earlier versions achieved limited popularity, as they required connecting the phone to an external card reader, or entering the information manually. A more recent development integrated a credit card reader directly into the Smartphone by either using image processing or an RFID / NFC Reader, permitting it to read Credit cards with RFID tags.

The vision is that Smartphones transactions will replace transactions that until now were typically done using cash. For example, Bob and Alice go to a restaurant, Bob pays for dinner with his credit card. Alice wants to share the cheque, but is not carrying cash… and so Alice gives to Bob her credit card; Bob scans the card using his Smartphone and charges Alice for half of the dinner cost.

Communications Service Providers (CSPs) hope to benefit from these financial transactions by establishing a presence at this new market. For example according to Bloomberg AT&T, Verizon and T-Mobile USA are planning a venture to displace credit and debit cards with Smartphone’s. However, they are not alone. many companies from different backgrounds, from Apple, through PayPal to Visa, are entering into the mobile payment market.

So what we have is a big market with huge potential, where CSPs, device manufactures, and Credit/Payments companies are all jostling to get a lead position, because as the song says the winner (segment) takes it all.

Can you imagine a world in which all the payments are made using phones and Credit card companies or the CSPs don’t see a penny of it?

All this excitement opens up new “opportunities” for revenue leakage and fraud. However this time the leakage will not be in minutes, but hard cash. The question is how much money will companies lose before they realize the need to adopt and adapt RA and fraud prevention standards from the telecommunication industry in addition with existing anti fraud best practices from the financial world.

Only time will tell…

Football and data mining – The results

Gadi Solotorevsky — Thu, 29 Jul 2010 05:56:47 +0000

Ok, so I wasn’t as accurate as Paul the octopus in predicting the Football world cup matches results, but I believe I have good reasons:
1.       I predicted all the results in advance, while Paul predicted each game at a time knowing the previous results. Certainly an unfair advantage!
2.       Paul has domain expertise, he started his prediction career two years ago predicting the UEFA Euro 2008 results, and he clearly improved since then. I just started now, with this world cup, so I can still improve!
3.      Scientist claims that octopuses are quite intelligent animals that can even use tools, So Paul’s intelligence is clearly an scientific fact , while no ,   scientist ever claimed that I am intelligent (I know a few that claimed quite the opposite)!Now being more serious, my predictions (see below) were quite good (my correct predictions are marked in red)
·         Reach the quarter finals: Spain, England, Argentina, Brazil, Italy, Holland, Germany, and France – 50% success
·         Reach the semi finals: Spain, England, Argentina, and Brazil – 25% success
·         Reach the finals: Spain and Brazil – 50% - success
·         Champion: Spain – 100% - success
·         Top Goal Scorer: Villa of Spain – 100% - success

Well I am quite sure that you aren’t reading my blog to learn about Football, so let’s look at the RA and Fraud angle.What I did (look at my previous post)?I was merely taking some predictive models (betting odds at several e-gambling sites), and combined them using a consensus/majority approach, the same methodology can be used to more important tasks, e.g., RA or fraud detection, in which you can combine predictions or clustering made by several data mining models into one more robust model, and in order to do this right you do not need a deep understanding of the models.

Saying this please note that the combination of Data mining models is not trivial, and should be done with care and understanding, for example you should take care of avoiding bias caused by redundant models.

Disclaimer – despite my good predictions, please do not attempt this at home, I do not recommend betting, the chances of losing, even if you use a highly sophisticated method, are higher than wining (otherwise gambling wouldn’t be such a great industry).

Until the next world cup, viva España !
Gadi

What’s between football and data mining?

Gadi Solotorevsky — Thu, 10 Jun 2010 12:15:44 +0000

By now you are certainly asking yourselves: Who will win the World Cup? Who will be the finalist? Who will be the top goal scorer? Here at cVidya we decided to make a friendly internal bet. I must admit that I enjoy football as the next person, but I don’t know a lot about it. So what should I do? How do I ensure that I get the upper hand?

Data mining is often used to identify or predict a phenomenon; e.g., identify fraud or predict high risk of churn. However, given the same data, different data-mining models will predict different outcomes, even when each of the models was built by an expert. One of the approaches is to try to select your prediction by consensus; basically, you predict that those clients that were identified as being at high risk of churn by the majority of the models are indeed a risk. The beauty of this is that you do not have to understand why and how each of the models made its prediction in order to reach the consensus. (OK, it is not so simple; you should take care not to have bias by duplication of identical or similar models – but experts in data mining also deserve to make a living!)

So what does all this have to do with football? Instead of data-mining models, I took the bet ratios from several e-gambling sites, combined them, and here is my bet. Next month I’ll let you know whether it was enough to win cVidya’s internal contest.

Disclaimer: The following predictions were made simply using some math and in no way show my personal preferences. And of course, I cannot guarantee that they are correct; quite the opposite, there is a high probability that they are incorrect.

Reach the quarter finals: Spain, England, Argentina, Brazil, Italy, Holland, Germany, and France

Reach the semi finals: Spain, England, Argentina, and Brazil

Reach the finals: Spain and Brazil

Champion: Spain

Top Goal Scorer: Villa of Spain

3D TV – The Wow Effect

Gadi Solotorevsky — Tue, 18 May 2010 09:28:09 +0000

Last week, when I was in China, I saw in Shenzhen airport a cartoon in a 3D TV without glasses set up by a Chinese company called TCL .
Amazing! The quality of the 3D was affected by my exact distance from and position relative to the TV. While the quality was still not HD, it was a marvelous preview into the future. Think of the opportunities, think of watching a ‘clásico’ between Real and Barça (Spanish football, of course) in 3D. Wow! Now think of having cameras installed on the players and watchng the game in 3D from Messi’s perspective.Does it have anything to do with RA? Well, yes! New technologies lead to new services that eventually lead to new billing schemes, which bring new challenges in terms of RA. For example, you pay for watching the game from the perspective of player A but the player is replaced in the middle of the game.
What happens now? How do you verify correct charging? What about the correct revenue share with the player, the team, etc? Now think of all this with real-time charging…After enjoying the match from Messi’s perspective and as he saw it, we RA practitioners need to develop RA techniques that will cope with these new realities.

Much Ado About Nothing

Gadi Solotorevsky — Sun, 21 Mar 2010 12:36:18 +0000

Since I work for an RA vendor and yet wish to avoid turning my blog into a marketing tool, I have decided long ago to refrain from writing in my blogs about any RA vendor. Instead, I try to put my ideas and thoughts, which anybody is welcome to contradict, in writing. In fact, I enjoy very much a healthy discussion.
However, I do wish for the discussions to remain at the thematic level and not descend into what this company is doing or why this person said this or that. Therefore, I abstain from publishing comments that try to use this blog as a platform to fight or praise RA vendors based on personal or business interests.

I am sorry if this policy frustrates some people, yet I do not plan to change itGadi

What’s in a name?

Gadi Solotorevsky — Wed, 10 Mar 2010 09:54:20 +0000

Juliet:
“What’s in a name? That which we call a rose
By any other name would smell as sweet.
“Shakespeare’s Romeo and Juliet (II, ii, 1-2)

Juliet claims that a name is just a name, but clearly all marketing people will tell you that a name is an important part of the essence. Just think that starting tomorrow we would call Coca-Cola Pepsi and vice versa.

I believe that RA is an evolving profession and not a random occupation. We have an established name for our profession, Revenue assurance. It may not be such a great name, because it also covers costs assurance. However, it is a well known and widely used name and so I am happy with it. My problem is with the way we call ourselves; are we RA experts, RA practitioners? (more…)

How do you say Revenue Assurance in Esperanto?

Gadi Solotorevsky — Tue, 23 Feb 2010 13:55:53 +0000

How do you say RA in different languages? I asked this my colleagues at Revenue Assurance Professionals group in LinkedIn, and got many responses. Regretfully I can’t, or at least I do not know how to, publish the answers in a nice table in LinkedIn so here it is

I believe that in Esperanto you just say “Show me the money”, or perhaps Enspezoj Certigo is more correct).

Additional contributions either here or via LinkedIn discussion will be most welcome

Many thanks to all the people to helped to create this list, Aaron Chou, Ahmed Mahmoud, Alexis Mauricio Serna P., Andriy Derevyanko, Ariana Pantazi, Asiedu Ansu, Awai Li, Dan Andersen, Eli Krakauer, George Karakiriakos, Gesona Ngabo,Guera Romo, Hubert Thomaz, Hugh Roberts, Hüsnü Serkan Alpcan, Irma Retzio, Izabela Zimoch-Konopacka, Jiri Ballada, Jose Aparecido Prado, Kaspar Lebedev, Magatte DIOP, Marc Vloebergh, Marcus Bryant, Miroslav Milanovic, Morisso Taieb, Nasir Iqbal, Olga Volkova, Panayotis Georgiou, Pedro Bravo, Rob Dielemans , Thomas Lüthi, Venera Zuna, and Vitalija Kackanaite

Best regards

Gadi

Happy new Decade

Gadi Solotorevsky — Tue, 05 Jan 2010 15:19:42 +0000

These days, while everybody’s busy summarizing the past decade, I want instead to presents my predictions for next year and the second half of the next decade. Even though my predictions about Astrophysics and modern poetry will likely be as accurate, I will limit myself to 3 predictions in the field of Revenue Assurance.

2010

1. RA continues as RA, but expands to additional areas (margin, collection, etc.)

2. Quantifying leakages and Justifying the benefits of proactive and preventive RA become more and more important. As a result SPs look for industry body guidance and best practices

3. No revolutions; just calm evolution in the field of RA discipline

2015-2019

1. Telcos continue their struggle not to become dumb pipes, not always successfully. As part of the struggle, many new business models emerge (and sink), requiring agile RA to cope.

2. Other industries (e.g., utilities) improve their metering capabilities and offer richer plans to their customers, leading to evident leakages and the adoption of RA practices

3. RA becomes an established discipline, with formal education and certification (think of chartered accounting).

Of course you are most welcome to have a good laugh at next year and end of decade and point out how wrong I was. But meanwhile … have a great 2010 and even greater decade in every possible aspect.

Gadi

Let’s do it in Assembly (or Excel)

Gadi Solotorevsky — Wed, 02 Dec 2009 09:26:37 +0000

I remember my second year at the university (Studying Computer Science for those who wonder). I wrote an Intel 8086 Assembly program that monitored the keyboard and interfaced directly with the printer (a 9 pin Epson dot printer, no less) to print a letter whenever a certain keyword was typed. I was very happy with my program as it handled all the interrupts really well and was very efficient. If memory serves, I even got a perfect score for it too.

Next day I started a campaign to rid the world of those superfluous printer drivers, replacing these inefficient pieces of software with custom code, whenever something needed printing. As you can guess, the first paragraph is true, but the second one is not. Using assembly, grants you endless flexibility; by writing your own code you can get exactly what you want. However, in real life, assembly is a language of last resort. It will waste your time and energy figuring out how to print a single character, instead of focusing on the important things, such as writing that amazing novel. In real life, the key is to focus on the essence and not the technicalities.

What does this have to do with RA? As it turns out, I have just read a discussion in LinkedIn in which 3 people were telling how great Excel is for doing RA. Certainly, Excel is a very versatile tool and can even be used as an RA tool. I would even go further and say that in some cases I would recommend it as a good tool for quick and dirty check of limited data. However, automation, tracing, follow up and reusability require more than a generic tool. RA is so much more than comparing two columns of data (and yes, I am aware that Excel is much more than that). The RA practitioner should focus on the investigation and resolution of problems and leave tasks that can be automated to the tool. BTW, do you remember how many leakages started in badly maintained reference tables in Excel? Do you want your RA findings to be so cumbersome?

Before you attack me, I admit and have never hidden the fact that I work for a RA systems vendor. However I also believe that billing should be done by a billing system and not by a clever little macro in Excel.Incidentally, if you hear of an opening for a Telecom billing system programmer in Assembly (or excel), let me know.

Customer complaints solve at first time

Gadi Solotorevsky — Thu, 05 Nov 2009 08:37:26 +0000

TIO is the Australian Telecommunications Industry Ombudsman in charge of resolving complaints made by small businesses and residential subscribers against their telephone or Internet service.

In his blog, Tony Poulos summarizes many interesting facts from TIO’s last report. I highly recommend that RA practitioners everywhere read the full report.

I have always considered customer complaints a goldmine of information for RA. As TIO reports show, customer complaints are not just an information source, but also a cost leakage, that should be prevented.

Allow me to explain.
TIO charges the service providers according to the number and type of complaints made against them. The cost per complaint ranges from $60 to $5200 depending on the severity of the complaint. From July 2008 to July 2009, TIO handled 230,065 complaints, charging Service Providers $ 23,676,085. TIO accepts a complaint only after the concerned service provider has had the opportunity to resolve it. The astonishing fact is that 90% of all complaints, some 208,000, were resolved with a referral back to the service provider. These 90% of complaints represent a sum of approximately $21MIn other words, had the service providers resolved these justified complaints, when they had received them the first time, they would have received up to 90% fewer complaints via TIO, saving 90% of the fees they paid to TIO, plus the cost of internally re-handling the complaint when submitted by TIO.

So correctly handling the complaint the first time is a clear issue of Cost Assurance that can show its direct $ value, not to mention the improvement in customer satisfaction, and other indirect impacts. Yet the complaints amount increased from 2007/8 to 2008/9 by 54%; go figure.