Business Intelligence and the Principle of Jidoka: Interview with Toyota's Simon Dorrat
Automotive manufacturer Toyota is widely credited for creating the precursor to Lean manufacturing, an approach which seeks to remove inconsistency and waste from processes. But how does Toyota translate its own approach - the Toyota Production System - from the manufacturing line into other areas of the business?
Simon Dorrat is Manager of Toyota’s Business Intelligence function where he is responsible for defining and delivering all services relating to Business Intelligence and Data Warehousing including BI, ETL, Data Quality, Master Data and OLAP.
In this PEX Network interview, Simon shares his thoughts on how Business Intelligence fits with the Toyota Way, suggests three ways for IT to provide better value to the business and even explains why doing a kitchen renovation helped some illuminate important aspects of software development.
Q: We often think of the Toyota Production System as one of the exemplars of continuous improvement culture. Where does your Business Intelligence program fit in?
Continuous Improvement at an organisational level requires a combination of understanding how we are performing and why, (using the principles of Genchi Genbutsu - going to the source), and an ability to forecast future performance and direction and work that into a strategic vision. Both of these activities require accurate and trusted information that can be analysed easily by the business users through our enterprise analytic and visualisation tools.
The TPS system also contains the principle of Jidoka. This entails process automation with inbuilt safety mechanisms to halt processes when a problem is sensed. One way our Business Intelligence program supports this is by providing visual representations of complex processes, such as our vehicle delivery supply chain, which tracks the progress of vehicles through the various stages from order to delivery to the customer, highlighting the emergence of bottlenecks or gaps which may lead to late delivery or oversupply. These issues can then be dealt with and improvements put in place to ensure the principle of Just in Time is achieved and customer satisfaction is maximised.
Q: Do you incorporate elements of the Toyota Production System - such as visual management, mistake-proofing, going to the "Gemba" and Kaizen, PDCA - into your Business Intelligence programme? If so, how?
The Toyota Production System is a practical expression of The Toyota Way - principles that guide everything we do in Toyota, based on Continuous Improvement and Respect For People. In our Information Systems Division we incorporate those principles in a number of specific ways, including by standardising processes, partnering with suppliers for scalable, flexible resourcing and by practising and encouraging Kaizen (Continuous Improvement).
In the Business Intelligence group specifically Kaizen led to us creating the Rapid Value Delivery methodology to allow us to provide a more timely and "value for money" solution delivery service. This was based on extensive feedback from the business (our customers) around what their perceptions of our service were and what improvements they would like to see, combined with the results of our own reviews and industry experience.
The outcome was a marriage of agile approaches with our standard project framework leading to a leaner and more efficient process. This provides a faster turnaround and a much higher level of confidence that what is being delivered matches the users’ expectations as they are a full partner in the development process. "Mutual Trust and Mutual Responsibility" with internal customers and partners is a key element of the Toyota Way and this new process relies on strengthening and then leveraging that relationship.
Q: One of the challenges of anything that involves information technology is the length of time it takes to deliver them - by the time they get in the hands of users, the business needs have often moved on. I understand that you recently applied some lessons from a kitchen renovation you were doing to the difficult task of developing appropriate business intelligence tools. Can you elaborate?
The input to the Rapid Value Delivery process was business feedback that we solicited combined with our learnings from many years of delivering BI solutions. The kitchen renovation was mainly a real world example used to explain the problem and the solution in a way anyone - especially non-IT people - could relate to. There were many parallels so it works well for this purpose. Also it's true that as IT practitioners we sometimes can get out of touch with the experience from our customers' point of view.
I recently went through a kitchen renovation - a field I knew little about - and I was exposed to a practical demonstration of the difficulties our users often face. For example, one of the challenges is being able to properly articulate their requirements when you do not have a background in the field (wheter it’s BI or Kitchens) nor a common language (Dimensions, Hierarchies, Measures/Vinyl wrap, Laminex, Splashbacks, canopies) to communicate with. It was also a practical example of how being able to look, touch and revise was critical to me being able to get my kitchen "requirements" right. This was provided via a combinations of showroom visits, sample paint patches and 3D printouts that were updated regularly based on my feedback and growing understanding of what I wanted and what was possible.
In the BI world, we provide this by heavy use of prototyping and constant user involvement in the solution validation phase. When this is completed both the business and us have a detailed understanding of what will be delivered and how, and high level of confidence that it will meet their needs. Then we move in to the traditional project lifecycle to do the back end work, the detailed build and the testing, thus ensuring the quality of the final solution.
As you stated, the length of time from conception to delivery of an IT solution is critical and even more so in BI, where the goal is to use the information to create innovation, which usually has a use by date from a competitive point of view. Applied to BI, most traditional (waterfall) delivery methodologies often end up delivering something that does not accurately reflect business needs due to the nature of the requirements gathering, and can also take far too long to deliver. The approach we take is to try to separate the deliverables into manageable chunks to ensure faster and more regular delivery into production. This also has the effect of allowing the business to learn more about what their real needs are for BI, about BI in general and for us to learn more about the business. As this relationship builds and trust in our ability is gained, the subsequent projects have a much higher likelihood of success.
None of this is earth shattering from an IT perspective, the principles of Agile have been around for a long time. What we have tried to do is build a process that leverages those principles without sacrificing the rigour and robustness necessary to ensure data quality, and that can be applied to the heavy lifting nature of the data warehouse back end components of BI projects.
Q: Engaging users throughout the process has got to be key to success; BI is only useful if business users utilise the information available to them. But it's also critical that there are no data errors going into the system. What do you do to engage users so that they can realise the benefits' of BI?
Engagement of the users takes two main forms. One is during the solution validation phase where there is heavy user involvement allowing us to ensure we fully understand the business requirements and the solution required to deliver them. This involves a number of techniques including prototyping of dashboards and reports, definition and validation of KPIs and business rules (through discussions and building examples on real data) and doing a detailed data source analysis. This helps highlight potential data quality issues which could affect accuracy or have a negative effect on project schedule and budget if discovered later.
The second form of user engagement is in UAT. A big emphasis is placed on the users' owning and committing properly to this activity. Often this requires education in advance so they know how to approach it and the likely effort involved. In total, the success of the solution correlates very highly to how effective the engagement with the users was. Often it can be a challenge to convince them of the benefit of committing the time required but we often find the process leads to the business better understanding their KPIs or discovering errors in current data, which clearly demonstrates the value of their efforts.
Q: Can you give me a concrete example?
One project last year discovered a number of flaws in both the implementation of business rules written into a semi-manual extraction of data and in the definition of another rule. These were both hidden from the business but were flushed out during the UAT process when reconciled against current reports. The errors were based on subtle distinctions in data that in certain circumstances distorted a key course attendance KPI. These were fixed and apart from the quality improvement it gave the business increased confidence in using the data sourced from the BI system and in the process we used in the project. They had an increased sense of ownership due to their heavy involvement. This pattern has been repeated in many of the projects we have run using this approach.
Q: Business Intelligence has traditionally been an IT-based function. Do you think IT will continue to take the lead in Business Intelligence programmes? Should it?
BI should always be business driven and business aligned, and for success requires a high level of ownership and involvement from the business. However IT provides the necessary expertise in quality and delivery and also a necessary enterprise viewpoint. This is increasingly important in the era of Big Data as performance, reliability and security become more and more challenging. So the ideal is to combine both, by using strategic planning and governance structures that span the business and IT, such as a BI Steering Committee for high level planning and a BICC for practical collaboration. IT then becomes a partner to facilitate delivery of solutions in the most sustainable way.
Q: What would you say are the 3 keys for IT to provide greater business value?
1. Being service oriented - by properly defining and measuring the services offered to the business with a true costed services model, IT can ensure that the costs can be accurately understood, efficiencies implemented and that the efforts of IT are targeted at the areas most critical to the business functioning successfully. This becomes a basis for clear communication and expectation setting between IT and the business.
2. Responsive and scalable delivery capability - in order to provide the necessary value add to the business, IT need the processes and partner relationships in place that can provide a flexible delivery capability that can react to business needs without sacrificing quality. This could be through a combination of internal and external resourcing options, delivering solutions via the cloud, COTS or in-house development. IT's core value is providing this service with the necessary governance and ensuring the best long-term outcomes for the company.
3. Educating and collaborating with the business to ensure strategic alignment - Strategic alignment of our long term plans with the business is obviously critical. A core part of this is our role in educating the business to allow the conversation to take place at a level where a true collaboration is achieved.
This is especially true with the rapid expansion of the reach of technology, such as in mobile platforms, collaboration and social media. But also in the non-technical space with areas such Information Governance and data quality having to be business owned but IT leading the discussion and enabling the outcomes.