Tesla: Why Elon Musk ditched ERP vendor SAP to build his own from scratch ?

I want you to think about a situation when you are deciding to wear a suite for a special occasion. You have the option to buy a ready-to-wear suite from a store (Boss for example) with the possibility to customize it a little bit at the tailor so it fits better. The solution is cheap and operational.  The other option is buying the material for the tailor so he can tailor your suite from scratch, it will cost you more but will fit you perfectly.

Now think about an analog situation: You are Elon Musk (Tesla’s CEO), the suite is an ERP system, the store is SAP, the tailor is your IT team and the occasion is going to the electrical car market and be a successful and innovative company.

Back in 2012, Elon Musk decision was pretty straightforward, he had his CIO Jay Vijayan ditch ERP vendor SAP and build his own custom ERP from scratch.

The idea came as a solution to deal with one main issue:

Most automakers use vertical ERP systems (for example SAP for car industry). However, this typical solution is an issue for an innovative company like Tesla because their model was different from the typical car makers.

Typically, the industry is working with the dealership model: all sales are required by law in most states to go through franchised dealers. That also includes service, although dealers get reimbursements for warranty-related maintenance and repairs.

Nevertheless, besides the innovative approach of Tesla in making electrical cars, they also wanted to innovate their sales model by maintaining a very direct relationship with its customers since the company has vertically integrated distribution and service.

In this case, coping with a new sales process will affect other processes like procurement and production. For these reasons, the vertical ERP solution was deemed ineffective and constraining for Tesla’s adaptability and ability to move quickly (which is the main competitive advantage that the company has over its larger rivals).

In a 2014 interview with CIO Insight, Vijayan (Tesla’s CIO) commented on the decision to build their own customized ERP:

Elon’s vision is to build a vertically integrated organization where information flow happens seamlessly across departments and where we have a closed feedback loop to our customers. By doing this, we can provide the best possible product, service and overall experience to our customers in the fastest way possible, while also operating efficiently as a business.

Tesla proved that process innovation needs the right tools to support it. The ERP system being a central one, an innovative company should make the strategic and difficult decision of going for a vertical solution with lower customization capabilities or building a laborious highly customized system from scratch. Yes, we are back to the suite dilemma and Elon Musk decision was to wear the tailored suite.

Beyond ERP Systems: Hybrid ERP

New enterprise software companies like Salesforce.com, Taleo, Workday, and others are disrupting the traditional ERP market by offering sophisticated, flexible, and easy-to-use applications. These systems are focused on specific functional areas: finance, procurement, supply chain, HR, sales and marketing.

Fitting in these applications into the legacy ERP system is what we call today “hybrid ERP”. The resulting ecosystem is a “fit-for-purpose hybrid ERP software” and it can have a multitude of combinations customized for the company’s needs.

Therefore, new operations are linked to the legacy ERP to form the Hybrid ERP:

  • Web and mobile storefront
  • Sourcing and vendor contract management.
  • Contingent workforce management.
  • Workforce management and recruiting.

hybrid

Source:  Strategy& analysis

The Saleforces.com-like applications are incorporated into the hybrid ERP and are replacing the classic/traditional view of the system as organizations are becoming more flexible and more mobile. These new capabilities are integrated in the traditional ERP software ecosystem as “hybrid ERP” systems to enable specific functions needed by the evolution of certain processes. For instance, the human resources function because recruiting process is changing, and e-commerce to operate with greater agility and independence from old-style IT companies.

 

Source: PwC Hybrid ERP: New technology, new options.

Top 6 Tips for Irreproachable MPR

After studying about material requirements planning (MRP) and how critical this process is to the overall performance of a company. I spent some time trying to search articles about how to make a reliable MRP. I found an interesting article about how to make MPR process successful. I am formalizing it on 6 key steps which will help companies increase on time shipments, additional satisfied customers, lower inventory carrying costs and less inventory shortages.

Inventory Accuracy  

If you have incorrect quantity on hand, your MRP is going to plan around those inaccurate amounts. The one saving grace is that at a minimum, your MRP should be able to provide you a summary of the material requirements.

Forecasting 

MRP systems will use open sales orders or forecast to calculate the demand. the forecast can be a sales forecast or production forecast. In both way, MRP system will utilize this information to generate requirements. Therefore, we can approximate material needs beyond the time needed to get all sales orders. Also, forecast helps us in scheduling production process and staffing requirements.

Lead Times 

Date accuracy for sales order and purchase orders is a necessity. However, we need to know how long it generally takes for materials to arrive from the moment we order or how long it takes to make our finished product. ERP systems will have a series of fields where we can enter this information and use in to suggest when to place that PO depending on the lead time.

Safety Stock vs. Reorder Point 

Reorder point is commonly used as the level of an item signaling the need for action. Safety stock can be treated as a hard number that we can’t go below, or as a soft number, suggesting a quantity we don’t want to go below with a certain buffer or flexibility.

Blanket Orders 

In MPR, blanket order is a great tool to have a purchasing flexibility. For example, we ask our vendor for a certain quantity of raw materials over the next 12 months and proceed to enter that into our system. However, we won’t have a definitive date for each shipment. Therefore, we only enter one date in the system and enjoy a flexible supply chain.

Capacity Limits 

Capacity limits is about knowing the production limitations of our work centers. Most MRP systems will create a suggested production schedule and must take all the capacity requirements into account for planning.

source: http://www.sikich.com/blog/post/Top-Ten-Tips-for-Successful-Material-Requirements-Planning#.WCllXfkrLIU

Inventory Only Really Matter When It’s Converted Into Sales

After studying the inventory last week, we can say that it is the place where we manage our products to potentially make money when we sell them. For most companies, managing the inventory is also a troublesome activity. It seems that in the management of inventory, we are always left with disappointing choices:

 

In fact, in order to make the optimal amount of money, a business needs to satisfy 100% of its customers 100% of the time with 100% of the offered inventory. Therefore, many companies plan to bear the cost of excess inventory. It’s part of the cost of doing business. They are willing to absorb, within certain limits, the cost of holding excessive inventories to make sure they are always reactive to sales opportunities.

A business aims to satisfy 100% of its customers 100% of the time with 100% of the offered inventory

However, operating a business with excess inventory is a huge and painful burden, requiring markdowns, margin-loss, and most of the time, liquidation of the excess inventory.

Despite years of developing new methods to manage inventory and the implementation of sophisticated ERP systems, many challenges still persist because of the dynamic nature of markets. Therefore, to optimize the inventory management and to get closer to the 100% figures stated above, a company needs to align the best technology, the best processes, and the greatest people to operate in this dynamic environment.

Top 5 causes of bad Master Data Management in ERP systems

Master Data Management (MDM) is “a set of disciplines and processes for ensuring the accuracy, completenessand consistency of the important types of reference data in the enterprise, across different applications (ERP, Financial software, Planning software…) and across multiple business processes, functional areas, organizations, geographies and channels.

Therefore, insuring a good management of our master data will lead to a consistence and valuable use of our ERP system. In this post, we will present the top 5 reasons that can lead us to encounter problems dealing with the SAP ERP system within a global enterprise architecture. Meanwhile, please keep in mind that MDM operations are wider than the ERP itself: it goes beyond one application to ensure data consistency across platforms (from different vendors).

  1. MDM capabilities provided by one single ERP vendor is not flexible enough to deliver consistency for other applications.
  2. Master data is critical and can’t be proactively managed within an ERP system but needs independent capabilities.
  3. ERP products were not mainly designed  to support MDM operations,  instead the they present MDM as add-ins features to handle data.
  4. Mastering your data in an ERP is much more difficult with more than one instance (or with different ERPs altogether).
  5. Not using a decentralized, independent MDM hub to have a workspace between the transactional and analytic systems (please check my OLTP OLAP: a small clarification article). The hub will help to perform data management operations in the back office systems and the data warehouse.

 

capture

Figure: Informatica White Papers

 

Conclusion: ERP systems are simply not the right place to master data. Despite all the positives of ERP, flexibility is usually not associated with it. ERP doesn’t offer the tools to manage master data in the ways required by today’s complex organizations. Although it excels at transaction processing, it isn’t designed for active management of critical master data. Therefore, the use of an MDM Hub is mandatory to support complex enterprise applications.

5 Change Management Strategies for Global ERP Implementations

We saw how global companies are implementing ERP systems across their departments and across their international offices where thousands of employees within the company use the same ERP.

But what if the company needs to change ERP vendor or make a major change ?

For such cases, boards have to deal with the tedious phase of change management where the whole company go through a complex re-organization of the culture of work.

Therefore, global organizations implementing ERP software need to include the following five components to succeed the transition:

 

  1. Project champion: While employees usually report to the headquarters,  their allegiance lies where their leadership resides. So the company need to have local project champions to push toward the acceptance of the new ERP in every big office around the world.
  2. Cultural considerations: Language, education and demographic differences have a significant impact on how employees perceive and understand goals, objectives, risks and benefits of an ERP project.
  3. Employee engagement: The adoption of the new ERP system needs to be a clear objective for the whole organization. Employees need to be involved by quantifying the impact of such important decisions.
  4. Communications: Communication within a global ERP project is absolutely vital but incredibly difficult. Most employees would like to hear updates about the project from their manager during team meetings [1].
  5. Training: Change management plans must be devised in order to address workforce transition to the new ERP system and who will be super-users and trainers.  A comprehensive training strategy is needed [1].

SAP HAUS party event

Last Thursday, I attended the SAP HAUS party event at Palo Alto.

I wanted to share the experience with you because it was very friendly and informative.

Here is the key observations about the event:

  1. It was an non-formal event by SAP to communicate about the career opportunities for students.
  2. Students came from SCU, Stanford university, Carnegie Mellon university, Berkeley and SJSU.
  3. The SAP stuff was very friendly and the ambiance was very cool (music, food, drinks)
  4. Students can chat with developers, designers, recruiters and project managers.
  5. A lot of activities were set to ease communication between all groups.

 

Here is also the key findings about the event:

  1. SAP is a big corporation with opportunities across all segments of the IT enterprise industry.
  2. They are working on different disruptive technology to shape the enterprise IT process: IoT was the the most important field that SAP wants to develop since they are able to drive change for their legacy customers.
  3. They are really cool people to talk to. In fact, they even spend time detailing what they do and giving advises to future graduates.
  4. For those graduating this spring, definitely apply through their career website.

Finally, we also enjoyed our time there and learned a lot about different positions within the company. The event totally worth the time allocated because of the added value that we got from SAP employees.

 

A 2003 Patent: US 20040162768 A1

During the last session, we discussed the procurement process for a Vendor-Managed Inventory (VMI) by using the Safeway example. The concept was initially initiated by Walmart and it’s supplier Protect & Gamble back in the 80s to provide 4 main advantages for the suppliers:

  1. Improved customer service
  2. Reduced demand uncertainty
  3. Reduced inventory requirements: By knowing exactly how much inventory the customer is carrying, a supplier’s own inventory requirements are reduced since the need for excess stock to buffer against uncertainty is reduced or eliminated.
  4. Reduced costs: To mitigate the up-front costs that VMI demands, Fox suggests that manufacturers reduce costs by re-engineering and merging their order fulfillment and Distribution Center replenishment activities.

    Source: Mary Lou Fox

 

Meanwhile, vendor managed inventory (VMI) has evolved as both technology has advanced and as business processes have improved. Therefore, to meet this evolution the ERP system needed to integrate VMI into the general enterprise architecture.

The US patent 20040162768 A1 was one of the first solutions to elaborate an effective implementation of VDI processes into a shared portion of ERP between the customer and the supplier.

You can find here the Patent document. It’s a little bit long (remember that it’s a legal document, so it has to) but very clear. And let’s face it, we don’t spend time reading patents, so it could be your first time doing it. Happy reading experience.

OLTP & OLAP: A small clarification

This post is about curious thoughts that I had during one of my internship when I had to implement an Oracle’s Hyperion Reporting software. I understood by that time that I was using an OLAP environment.

Wait, what is OLAP again? The concept surfaced  again the last week. During the previous sessions, we explored the accounting transaction concept within an ERP system. We highlighted the 3 different types of data:

1. Master data
2. Organizational data
3. Transaction data

The ERP generates a huge volume of these data sets that need to be stored in two different ways, depending on the freshness: “Transaction” data for the current processes and “historical” data for past processes.

  • OnLine Transactional Processing (OLTP) to process the current data needed to fulfill a current business process.
  • OnLine Analytic Processing environment (OLAP) to analyze a huge amount of historical transactions.

OLAP environments needs a lot of processing power to aggregate all the historical data in order to provide the analysts with different “Views” of the overall data. Data are arranged by dimensions and can be modeled by a “Cube”.

what-is-a-cube

 

Re-learning ERP systems (week 1)

During the first week of ERP system course, I started re-discovering ERPs from a different perspective. It’s different because I had a six-months internship in 2015 working on ERPs implementation. I was a consultant for an IT company working on systems implementation of Oracle Financial Management software (Hyperion) and Financial planning. The experience helped me to shape my architectural knowledge of ERP systems within different big French companies.

While enjoying learning about the infrastructural architecture of ERP systems, I was always curious about how the business process architecture looked like. luckily, the first week of the course partially answered my question. Partially because I got to know about the big picture of a general ERP system from a business perspective (main purpose, different BPs).

We also started digging deeper into interesting aspects of the of the system: The role of interfaces and the Silo effect problem (or not, according to our Professor, the silo effect helped each department to maximize efficiency). This aspect was very important on mapping my previous knowledge about ERP infrastructure and the new learned items and the following weeks will probably help me complete the picture of my architectural business perspective of such systems.