Community Forum – Types of R&D Organization

How do you organize R&D?

The CTO from Procter & Gamble, Bruce Brown, gave an excellent presentation on “Why Innovation Matters: Mastering a Reliable, Repeatable Process” as the opening keynote for our Annual Meeting on May 4th. He discussed changes in how R&D was organized at P&G – traditionally as centralized to totally decentralized in 2000, and then, recently, as a “Recentralized Hybrid Global Capability” where the R&D will be a 60/40 Business/Corporate split activity. He cited enhanced innovation, increased R&D productivity, leveraging scale, improved agility and benefit to the employees as reasons for this shift.

My questions are: What do you think the pros and cons are for each type of R&D organization – Centralized, Decentralized and Hybrid? What has your experience been with effective R&D organizations (you may have experienced all three)? What type of R&D organization do you have in your company and what makes it effective? Lastly, if you have central or hybrid R&D, how do you name your groups so that anyone in the company knows “where to come to for what”? – Martha Collins, Director, Global Technology Center, Air Products and Chemicals, Inc.

Community Responses

Ted Ciolkosz, Senior Director, Instrument Development, Waters Corporation​
Here at Waters Corp, I would say we are in the 60/40 or even 80/20 range of Business/Corporate activity. We have four main business areas: Liquid Chromatograph (LC) Instruments, Mass Spectrometry (MS) Instruments, Thermal Analysis (TA) Instruments, and Chemistry Operations. These businesses are scattered around the globe – US, UK, Singapore, etc.

Each of these business units runs their research and development mostly autonomously. The corporate coordination and direction is accomplished in four main ways:
1.  We talk frequently – mostly via teleconferences – but also in person. If there is an interface required to products in another business unit (almost always these days), we will have a representative(s) participate in weekly project meetings. We also encourage more frequent personal conversations and emailing.
2.  We are now using a Corporate Instrument Development Process (referred to as the CIDP). In the past, each business had their own process where many of the phase gates had different definitions and deliverables. Two years ago, we formed a team to unify the independent processes. This was no easy task and, although this effort was promoted and sanctioned at the senior level, the time/resources to accomplish it were not directly planned on. Initially, the motivation was mainly to simplify our position with various regulatory bodies such as ISO, FDA, etc but the advantages of using a common language when defining and reporting on projects was also recognized. We are now fully onto the new process (no remaining ‘grand-fathered’ projects) and have fully trained everyone. We have built a fair amount of flexibility into our CIDP by letting the Marketing and Project managers define the scope up front as needed. For example, a relatively minor hardware improvement to an existing instrument in which the research group has already built a prototype, might elect to go directly from ‘Spec and Planning’ to ‘Beta’ phase – skipping the usual ‘Alpha’ phase. Any deviations from the norm have to be presented and approved by Senior level management at the Feasibility phase gate review so there aren’t any surprises later on and before the ‘Spec and Planning’ phase can begin. Enforcing our CIDP has had some difficulties. Its all too easy to let the process slide when faced with the inevitable technical and scheduling challenges and it is also very, very tempting for upper level management to specify a product delivery date (i.e. for a trade show) before feasibility has been achieved – which means the project team is faced with inventing while on a timeline!!! Suffice it to say, we need better adherence to the process while not losing sight of the business demands.
3.  We have yearly planning meetings for each of the business areas and then a combined meeting where product road maps are worked out. This has a very strong emphasis on required support from other business units. These meetings involve all senior management – including our CEO. This goes a long way towards getting everyone on page.
4.  The entire corporation is now using common engineering tools. This may seem obvious, but it wasn’t true until recently and it isn’t easy to transition a business unit onto different tools because of legacy support and use of existing subassemblies, PCBs, software algorithms, etc. in new products – not to mention training – all while still pumping out new products. Keeping these tools at the same revision level and maintaining common practices within the tools is an ongoing challenge.

Items 1-3 above also apply within each business area where several product lines often compete for precious resources.

–

Dan Abramowicz, Executive VP Technology & Regulatory Affairs, Crown Holdings
I’ve had experience in completely centralized, completely decentralized, and combination R&D organizational models.  In my view, any of these organizations can work, if the ‘concerns’ are proactively managed.  My brief thoughts follow:
1. Centralized
a. Advantages include: administrative simplicity; more control for the R&D group on portfolio makeup / balance; facilitation of cross-business programs; more interaction with corporate leadership; the ability to utilize core skills across different business units; and a better ability to support longer range developments
b. Concerns include budget risk (as the central number becomes a visible target when major cost reductions are necessary and the risk the R&D group will ‘decouple’ from the businesses (working on problems they feel are relevant, such as blue sky research).
2. Decentralized
a. Advantages include controls to ensure the R&D group is working on business-relevant programs; increased business acumen in the R&D staff; more interaction with Business Leaders; and budget stability (assuming funding comes from several business, offering a ‘diversified’ investment portfolio)
b. Concerns include the risk that the portfolio will become too short-term focused; that the R&D group will spend disproportionate efforts attracting money; that core skills are duplicated for different business units; and that cross-business synergies go unrealized
3. Combined model
a. Advantages include the ability to maintain a balanced portfolio; and to still leverage core skills across different business units and cross business programs; and to effectively manage technology transfer.
b. Concerns include responsibility definitions (who does what) that could lead to gaps or duplications.  Does central research do all the longer term work (and only longer term work, leading to concerns stated above)?  Does the business do all the applied work (and applied work only)?

–

Will Goss, R&D Section Manager, Champion Technologies, Inc.​
This question really keys in on the topic of “Organizing a company’s R&D for success.”  I don’t believe there is any simple answer beyond the following truism: R&D needs to be organized to be effective in the current culture of the company it serves. Several key topics are implicit in this statement:

1.  R&D is a service organization and in any culture needs to deliver to its customers (internal & external) the products and/or services needed for that company to be successful against its competition. The specifics of this are plethora depending on the competitive climate each company and its specific industry plays in but the second topic stems from this.

2.  R&D needs to be effective for each company’s competitive environment. P&G realized this I believe when they moved from centralized to decentralized so that their R&D functions became focused on each business customers in which the company wanted to compete and be innovative. But the downside of this is perhaps why P&G has further evolved to the hybrid scheme; too much specific focus can stifle open innovation and the ability to “see the forest for the trees” or the “big picture” and the trends of an industry. Many a good company failed to become great by not paying enough attention to the bigger picture, the competitive threats from other industries or displacement technologies, things that a decentralized R&D organization needs to guard against. Not that it can’t be done but the rewards and incentives are often not applied to these sustain activities. This leads to next topic.

3.  R&D needs to be organized; a subjunctive verb tense is used to show how this hopes to happen and wants to happen in a renewing sense of leadership for an R&D organization to be the best it possibly can. This leadership has to come from the highest levels in an organization so that the message is neither ambiguous nor debatable and reflects the elements of a company’s culture that integrate the R&D functions with the rest of the organization. The culture and its leadership may favor a more centralized, decentralized, or hybrid structure to accomplish this but the particular structure should be thought out to mitigate the weaknesses and leverage the strengths of the company’s available resources through communication throughout the organization.

My personal experiences over my 26 year career in R&D leads me to conclude that the hybrid model, especially is the 60:40 split if allowed some flexibility around it, is the most flexible and robust for most growing organizations of any size (>= $100 million annual turnover). By its very nature it recognizes the short term and the long term as well as the currently employed versus the “could be” employed technology elements. But any organization with a leadership that recognizes these and other conflicting yet complimentary forces as necessary and vital to a company’s future prosperity can probably make any of the aforementioned R&D structures work by employing balanced metrics and incentives to keep things in balance.

–

Dave Kashdan, IRI Emeritus Member
1.  Centralized R&D organization offers a chance at long-term focus, achieving great company-changing innovation, broad utilization of skill sets and relative ease of achieving critical mass with minimal resources.  But centralized R&D organizations often struggle to direct their efforts in a way that adds value to the existing businesses, and may eventually earn the pejorative nickname of “ivory tower” — a sure sign that it has lost contact with the rest of the organization, and is not working on projects seen as valuable by the company’s existing businesses.  It is the poor interactions with business folks and customers that is often the downfall of centralized R&D organizations.  Companies are driven towards centralized organizations when they have few resources or customers, and they do not perceive that a lot of customer interaction is needed in their R&D processes.
2.   Decentralized R&D organization gives the R&D folks excellent contact with the business people, good direction on projects that are needed to help the existing business, and a chance to learn first-hand what customers need.  But while decentralized organizations often work on good projects, they rarely work on great projects that will make a big impact on the company.   They are often driven to fix the last problem that a customer complained about, and only until the next customer registers his/her complaint about something else.   These organizations often use up the company’s “seed corn” — its collection of ideas and novel scientific leads — as the R&D organization pursues “low-hanging fruit,” and may appear extraordinarily productive for a while.  But these organizations are notoriously poor at creating new novel ideas, new directions to pursue, and the work begins to look unexciting and lacking in creativity after a few years.  As the wealth of new ideas are consumed but not replaced, workers will complain they are just tweaking existing products and processes (vs. inventing new ones) and come to believe that the R&D organization has a lot more D than R in it.  After some time has passed, the business folks will begin to wonder where their next big source of growth will come from.  Companies are driven towards decentralized organizations when they have a great many different customers and their many product lines are customized towards each group of customers.
3.  Hybrid — tries to utilize the strengths and overcome the weaknesses of centralized and decentralized organizations.  In its best format, it is able to do both long and short term work effectively, enable good customer and business contracts for the R&D workers, and receive good direction from business and customers while still having a long-term vision that can make a difference for the company.  Companies are often driven towards a hybrid R&D structure when they have grown large and have experienced the limitations of centralized and/or decentralized organizational structures.

Probably the best situation is when R&D workers can move from one environment to another on a 3-4 year cycle.  As they move into a decentralized organization, they can build business and customer relationships while they can still rely on the relationships with the central R&D function that they developed in prior years (so they know where to go for help).  After a while those old relationships begin to fade, and the person needs a stint back in a centralized organization, where his/her fresh business and customer contacts will help him/her focus on relevant work, yet have the long term focus and resources only available in the centralized organization.  Some companies can achieve this back-and-forth mobility by moving people between organizations, while others periodically change the company’s R&D structure.

You might want to look up:  Tom Tirpak, Roger Miller, Larry Schwartz, David S. Kashdan, Globalization and the R&D Organization, Research-Technology Management, 2006, September-October, pp. 19-26

–

Martin Zachau, Vice President Research and Development, OSRAM Sylvania
Thank you for bringing up this question. I am looking forward to the results of this survey. OSRAM Sylvania is the #1 lighting company in North America and a subsidiary of Siemens. We have a hybrid R&D organization, which fits our business needs. The development departments of the business units focus on executing the product development roadmaps. They can flexibly address the specific market needs of the respective BU. We have decentralized the development organization long ago, and we don’t question that his has been the right move. On the other hand, the tasks of the Central R&D organization are technology scouting, evaluation and funneling of ideas, and development of (platform) technologies up to roughly two years before market introduction, when some technologies are handed over to the business units. OSRAM Sylvania feels that there is a need for an internal Central R&D organization that can address high-risk and cross-divisional topics, or those beyond the scope of the business units. Traditionally we have used the terminology of “Research” and “Development” for the two portions of R&D. Naming conventions change frequently in our organization, and the terms Central Technology, Central R&D, or Conceptual Engineering can also be found. 

–

Terry Rosenstiel, Associate Director, USG Corporation
Our R&D effort at USG Corporation is centralized. 

The laboratories are organized at the central technology center in Libertyville in alignment with the business units.  For instance, my lab, the Ceilings Laboratory, encompasses tile and grid and is aligned with the Ceilings Division.  In a similar fashion, another lab, the Performance Surfaces Laboratory, is aligned with the Performance Surfaces Division. And so on.

In the product laboratories, the work is a composite of support activities for the business units (current product property and manufacturing issues), development projects for the business units (new products/processes in traditional product lines), and long-range strategic projects.

Our analytical and systems test laboratories are co-located at the site.

–

Stephen Toton, Director, DuPont Science & Technology
Over my 30 years of experience I have seem all different types of R&D structures.  There is no perfect answer.  It depends on the strategy of the Business.

In general there are two negative extremes you want to avoid.  The functional silo or the business unit silo.  The functional silo can easily get disconnected from the market and the business unity strategy and in some ways become self serving.  The business unit silo can easily prioritize R&D on short term needs and create project stops and starts for hitting quarterly results.  Therefore you need to balance R&D in the enterprise.  At DuPont we do both. We have dedicated R&D for each Business Unit with sufficient critical mass and we have a Central Research Unit focused on long term transformational R&D, with projects that are dedicated to the Business Units.  The CTO maintains governance over all R&D in the corporation for people development, R&D processes, competencies, growth & productivity metrics and infrastructure.  We review the R&D portfolio for each business unit at least once a year to maintain the tension.

–

Bill Miller, IRI Emeritus​
I’ve read the seven responses to Martha Collins’s question about how to organize R&D and I think the responses have missed two important points.

First, the question is somewhat of a trap that ignores a broader view of open innovation and implies that R&D is mainly done only inside an organization. Some of my consulting clients have recognized this trap and have tried to solve it by creating new organizations such as strategic joint ventures with younger, smaller, but rapidly growing organizations that do R&D outside the mother ship. Other clients have made minority equity investments in small companies to get visibility to their R&D and also emerging markets.

When this strategic external R&D happens, the second point then becomes apparent, which is that radical innovation is typically rejected in the larger mother ship even though it needs to be absorped. Therefore, the real question should be how to organize R&D for both incremental and radical innovation or what Christensen calls disruptive and sustaining innovation. Any R&D organization that doesn’t work with marketing or do its own marketing on radical innovation is ineffective. AT&T Bell Labs (centralized R&D) invented the cell phone, but AT&T had to acquire McCaw in 1994 when the market was already well established to get religion. Motorola built the first working cell phone and made the first call in 1973. Many large organizations such as Cisco have grown through acquisition. In that sense, the R&D for supplemental growth is external  – neither centralized or decentralized internally. Even though open innovation might be supported in a hybrid organization and be driven more aggressively by acquisitions, the question remains for what do about R&D for radical innovation that will change the mother ship.

My experience over 30 years  – managing R&D in corporations, then teaching innovation at universities and consulting –  shows that radical innovation requires direct support from the CEO and a new generation of theory and practice for innovation that yields better results. To see the need for a new generation of theory and practice, consider this question:  how would any organization effectively manage innovation to create the next generation of capabilities that radically improve R&D, marketing, product development, manufacturing and field maintenance such that innovation is accelerated but disastrous failures are prevented in complex systems such as those manifested in the Toyota recalls and Deep Water Horizon.

In summary, it seems there are two options for organizing R&D. First option: external R&D may be the best way to manage R&D for radical innovation, whereas internal R&D in any organizational model assisted by open innovation may be the best way to manage incremental innovation for existing businesses. The reason is that history has proven the mother ship will likely reject the results of radical R&D, whereas rejection in an external organization is much less likely. After commercial success is demonstrated, the mother ship can then do an acquisition – if it still exists and has the capital required. Second option: fix the problem with radical innovation within the mother ship so that internal R&D assisted by open innovation can be successful for both radical and incremental innovation. 

–

Have a response to add?  Email us!