Banking & Insurance

“Pipple is only truly satisfied with 99% reliability”

Workforce management without prediction models is like a phone without apps  states Jan van den Berg, manager BI & Control GV at Quion. ‘Only when you have proper software, you can use it to do amazing things’. With Pipple as a partner, Quion has started a process to derive more predictive power from their data.  

If there is anyone great with numbers, it is Quion. They take over financier’s back-end processes, related to closing and managing mortgages, and advise them on how to streamline their accounting processes. However, the internal control of data can use a boost.

A wide statistical funhouse revealed itself when, together with Pipple, they created cases in which prediction models could be very valuable. All kinds of great initiatives were brought up. From a warning system for long-term absence and a fraud detection model, to automatic credit risk assessment. It was decided to begin with a model to forecast the workload for the back office, mid office and telephony on a weekly basis. This would help the operation the most.

Buy in
Earlier, Jan already had a student experiment with such a model. ‘As a result, we knew the relevant parameters and we knew that the quality of the required data was in order,’ explains Jan. ‘In addition to being useful, a further developed model would also be a pleasant buy-in for other models on our wish list.’

Quion brought in Pipple consultant Vera van der Lelij, for a longer period of time. Jan: ‘Her start with us was as I had hoped from Pipple: she didn’t need long conversations, we didn’t have to ‘live through’ the problem first and she didn’t went on endlessly about what we had to do. She just got to work. The process for which Vera is responsible is simple: she pairs with a mathematically trained team member of Quion who knows a lot about the data, after which Vera starts creating a mathematical model. A third team member implements the new model in the data warehouse. If the result is insufficient, the whole thing starts all over again. In this way, Vera was soon able to demonstrate that the original model that Quion adopted deviated significantly from reality. Not something that shocked Jan and his people. A model can only be created iteratively. Failure is part of the process.’

Not just a user course
Pipple is driven to get the figures right, notes Jan. ‘They don’t really like a model until it’s 99% reliable.’ That’s their mathematical vision. Jan lowers the bar a bit. ‘Predictions are just predictions. Sometimes they don’t match reality. If we tolerate a 10% deviation in this specific model, at most we plan 2 FTE too much or too little. That’s already an enormous gain for us.’

What is very important to Jan is that Pipple makes his team understand the model. He won’t settle for a user course to get numbers to roll out. When we maintain the models, if necessary, I want us to be able to make adjustments to the model itself.’ Fortunately, Vera constantly makes sure that the necessary knowledge is present.

As far as Jan is concerned, Pipple will stay with Quion for some time to come. Because, like Quion, they are problem solvers, and their value grows as Pipple gets to know the world of mortgages better. Besides: ‘One single app on your phone is not nearly enough.’

Workforce management without prediction models is like a phone without apps  states Jan van den Berg, manager BI & Control GV at Quion. ‘Only when you have proper software, you can use it to do amazing things’. With Pipple as a partner, Quion has started a process to derive more predictive power from their data.  

If there is anyone great with numbers, it is Quion. They take over financier’s back-end processes, related to closing and managing mortgages, and advise them on how to streamline their accounting processes. However, the internal control of data can use a boost.

A wide statistical funhouse revealed itself when, together with Pipple, they created cases in which prediction models could be very valuable. All kinds of great initiatives were brought up. From a warning system for long-term absence and a fraud detection model, to automatic credit risk assessment. It was decided to begin with a model to forecast the workload for the back office, mid office and telephony on a weekly basis. This would help the operation the most.

Buy in
Earlier, Jan already had a student experiment with such a model. ‘As a result, we knew the relevant parameters and we knew that the quality of the required data was in order,’ explains Jan. ‘In addition to being useful, a further developed model would also be a pleasant buy-in for other models on our wish list.’

Quion brought in Pipple consultant Vera van der Lelij, for a longer period of time. Jan: ‘Her start with us was as I had hoped from Pipple: she didn’t need long conversations, we didn’t have to ‘live through’ the problem first and she didn’t went on endlessly about what we had to do. She just got to work. The process for which Vera is responsible is simple: she pairs with a mathematically trained team member of Quion who knows a lot about the data, after which Vera starts creating a mathematical model. A third team member implements the new model in the data warehouse. If the result is insufficient, the whole thing starts all over again. In this way, Vera was soon able to demonstrate that the original model that Quion adopted deviated significantly from reality. Not something that shocked Jan and his people. A model can only be created iteratively. Failure is part of the process.’

Not just a user course
Pipple is driven to get the figures right, notes Jan. ‘They don’t really like a model until it’s 99% reliable.’ That’s their mathematical vision. Jan lowers the bar a bit. ‘Predictions are just predictions. Sometimes they don’t match reality. If we tolerate a 10% deviation in this specific model, at most we plan 2 FTE too much or too little. That’s already an enormous gain for us.’

What is very important to Jan is that Pipple makes his team understand the model. He won’t settle for a user course to get numbers to roll out. When we maintain the models, if necessary, I want us to be able to make adjustments to the model itself.’ Fortunately, Vera constantly makes sure that the necessary knowledge is present.

As far as Jan is concerned, Pipple will stay with Quion for some time to come. Because, like Quion, they are problem solvers, and their value grows as Pipple gets to know the world of mortgages better. Besides: ‘One single app on your phone is not nearly enough.’

Workforce management without prediction models is like a phone without apps  states Jan van den Berg, manager BI & Control GV at Quion. ‘Only when you have proper software, you can use it to do amazing things’. With Pipple as a partner, Quion has started a process to derive more predictive power from their data.  

If there is anyone great with numbers, it is Quion. They take over financier’s back-end processes, related to closing and managing mortgages, and advise them on how to streamline their accounting processes. However, the internal control of data can use a boost.

A wide statistical funhouse revealed itself when, together with Pipple, they created cases in which prediction models could be very valuable. All kinds of great initiatives were brought up. From a warning system for long-term absence and a fraud detection model, to automatic credit risk assessment. It was decided to begin with a model to forecast the workload for the back office, mid office and telephony on a weekly basis. This would help the operation the most.

Buy in
Earlier, Jan already had a student experiment with such a model. ‘As a result, we knew the relevant parameters and we knew that the quality of the required data was in order,’ explains Jan. ‘In addition to being useful, a further developed model would also be a pleasant buy-in for other models on our wish list.’

Quion brought in Pipple consultant Vera van der Lelij, for a longer period of time. Jan: ‘Her start with us was as I had hoped from Pipple: she didn’t need long conversations, we didn’t have to ‘live through’ the problem first and she didn’t went on endlessly about what we had to do. She just got to work. The process for which Vera is responsible is simple: she pairs with a mathematically trained team member of Quion who knows a lot about the data, after which Vera starts creating a mathematical model. A third team member implements the new model in the data warehouse. If the result is insufficient, the whole thing starts all over again. In this way, Vera was soon able to demonstrate that the original model that Quion adopted deviated significantly from reality. Not something that shocked Jan and his people. A model can only be created iteratively. Failure is part of the process.’

Not just a user course
Pipple is driven to get the figures right, notes Jan. ‘They don’t really like a model until it’s 99% reliable.’ That’s their mathematical vision. Jan lowers the bar a bit. ‘Predictions are just predictions. Sometimes they don’t match reality. If we tolerate a 10% deviation in this specific model, at most we plan 2 FTE too much or too little. That’s already an enormous gain for us.’

What is very important to Jan is that Pipple makes his team understand the model. He won’t settle for a user course to get numbers to roll out. When we maintain the models, if necessary, I want us to be able to make adjustments to the model itself.’ Fortunately, Vera constantly makes sure that the necessary knowledge is present.

As far as Jan is concerned, Pipple will stay with Quion for some time to come. Because, like Quion, they are problem solvers, and their value grows as Pipple gets to know the world of mortgages better. Besides: ‘One single app on your phone is not nearly enough.’

Workforce management without prediction models is like a phone without apps  states Jan van den Berg, manager BI & Control GV at Quion. ‘Only when you have proper software, you can use it to do amazing things’. With Pipple as a partner, Quion has started a process to derive more predictive power from their data.  

If there is anyone great with numbers, it is Quion. They take over financier’s back-end processes, related to closing and managing mortgages, and advise them on how to streamline their accounting processes. However, the internal control of data can use a boost.

A wide statistical funhouse revealed itself when, together with Pipple, they created cases in which prediction models could be very valuable. All kinds of great initiatives were brought up. From a warning system for long-term absence and a fraud detection model, to automatic credit risk assessment. It was decided to begin with a model to forecast the workload for the back office, mid office and telephony on a weekly basis. This would help the operation the most.

Buy in
Earlier, Jan already had a student experiment with such a model. ‘As a result, we knew the relevant parameters and we knew that the quality of the required data was in order,’ explains Jan. ‘In addition to being useful, a further developed model would also be a pleasant buy-in for other models on our wish list.’

Quion brought in Pipple consultant Vera van der Lelij, for a longer period of time. Jan: ‘Her start with us was as I had hoped from Pipple: she didn’t need long conversations, we didn’t have to ‘live through’ the problem first and she didn’t went on endlessly about what we had to do. She just got to work. The process for which Vera is responsible is simple: she pairs with a mathematically trained team member of Quion who knows a lot about the data, after which Vera starts creating a mathematical model. A third team member implements the new model in the data warehouse. If the result is insufficient, the whole thing starts all over again. In this way, Vera was soon able to demonstrate that the original model that Quion adopted deviated significantly from reality. Not something that shocked Jan and his people. A model can only be created iteratively. Failure is part of the process.’

Not just a user course
Pipple is driven to get the figures right, notes Jan. ‘They don’t really like a model until it’s 99% reliable.’ That’s their mathematical vision. Jan lowers the bar a bit. ‘Predictions are just predictions. Sometimes they don’t match reality. If we tolerate a 10% deviation in this specific model, at most we plan 2 FTE too much or too little. That’s already an enormous gain for us.’

What is very important to Jan is that Pipple makes his team understand the model. He won’t settle for a user course to get numbers to roll out. When we maintain the models, if necessary, I want us to be able to make adjustments to the model itself.’ Fortunately, Vera constantly makes sure that the necessary knowledge is present.

As far as Jan is concerned, Pipple will stay with Quion for some time to come. Because, like Quion, they are problem solvers, and their value grows as Pipple gets to know the world of mortgages better. Besides: ‘One single app on your phone is not nearly enough.’


Interested?

Get in touch.